update sdxl-lightning references to flux

Support multiple variants for each generation

.github/FUNDING.yml

@@ -0,0 +1 @@
+github: [abi]

.github/ISSUE_TEMPLATE/bug_report.md

@@ -0,0 +1,21 @@
+---
+name: Bug report
+about: Create a report to help us improve
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+**Describe the bug**
+A clear and concise description of what the bug is.
+
+**To Reproduce**
+Steps to reproduce the behavior:
+1. Go to '...'
+2. Click on '....'
+3. Scroll down to '....'
+4. See error
+
+**Screenshots of backend AND frontend terminal logs**
+If applicable, add screenshots to help explain your problem.

.github/ISSUE_TEMPLATE/custom.md

@@ -0,0 +1,10 @@
+---
+name: Custom issue template
+about: Describe this issue template's purpose here.
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+

.github/ISSUE_TEMPLATE/feature_request.md

@@ -0,0 +1,20 @@
+---
+name: Feature request
+about: Suggest an idea for this project
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+**Is your feature request related to a problem? Please describe.**
+A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]
+
+**Describe the solution you'd like**
+A clear and concise description of what you want to happen.
+
+**Describe alternatives you've considered**
+A clear and concise description of any alternative solutions or features you've considered.
+
+**Additional context**
+Add any other context or screenshots about the feature request here.

Evaluation.md

@@ -5,8 +5,8 @@ Evaluation dataset consists of 16 screenshots. A Python script for running scree
 ### Running evals
 
 - Input screenshots should be located at `backend/evals_data/inputs` and the outputs will be `backend/evals_data/outputs`. If you want to modify this, modify `EVALS_DIR` in `backend/evals/config.py`. You can download the input screenshot dataset here: TODO.
-- Set a stack (`STACK` var) in `backend/run_evals.py`
-- Run `python backend/run_evals.py` - this runs the screenshot-to-code on the input dataset in parallel but it will still take a few minutes to complete.
+- Set a stack and model (`STACK` var, `MODEL` var) in `backend/run_evals.py`
+- Run `OPENAI_API_KEY=sk-... python run_evals.py` - this runs the screenshot-to-code on the input dataset in parallel but it will still take a few minutes to complete.
 - Once the script is done, you can find the outputs in `backend/evals_data/outputs`.
 
 ### Rating evals

README.md

@@ -1,24 +1,50 @@
 # screenshot-to-code
 
-This simple app converts a screenshot to code (HTML/Tailwind CSS, or React or Bootstrap or Vue). It uses GPT-4 Vision (or Claude 3) to generate the code and DALL-E 3 to generate similar-looking images. You can now also enter a URL to clone a live website.
+A simple tool to convert screenshots, mockups and Figma designs into clean, functional code using AI. **Now supporting Claude Sonnet 3.5 and GPT-4O!**
 
 https://github.com/abi/screenshot-to-code/assets/23818/6cebadae-2fe3-4986-ac6a-8fb9db030045
 
+Supported stacks:
+
+- HTML + Tailwind
+- HTML + CSS
+- React + Tailwind
+- Vue + Tailwind
+- Bootstrap
+- Ionic + Tailwind
+- SVG
+
+Supported AI models:
+
+- Claude Sonnet 3.5 - Best model!
+- GPT-4O - also recommended!
+- GPT-4 Turbo (Apr 2024)
+- GPT-4 Vision (Nov 2023)
+- Claude 3 Sonnet
+- DALL-E 3 for image generation
+
 See the [Examples](#-examples) section below for more demos.
 
+We also just added experimental support for taking a video/screen recording of a website in action and turning that into a functional prototype.
+
+![google in app quick 3](https://github.com/abi/screenshot-to-code/assets/23818/8758ffa4-9483-4b9b-bb66-abd6d1594c33)
+
+[Learn more about video here](https://github.com/abi/screenshot-to-code/wiki/Screen-Recording-to-Code).
+
 [Follow me on Twitter for updates](https://twitter.com/_abi_).
 
-## 🚀 Try It Out!
+## 🚀 Hosted Version
 
-🆕 [Try it here](https://screenshottocode.com) (bring your own OpenAI key - **your key must have access to GPT-4 Vision. See [FAQ](#%EF%B8%8F-faqs) section below for details**). Or see [Getting Started](#-getting-started) below for local install instructions.
-
-## 🌟 Recent Updates
-
-- Mar 5 - Added support for Claude Sonnet 3 (as capable as or better than GPT-4 Vision, and faster!)
+[Try it live on the hosted version (paid)](https://screenshottocode.com).
 
 ## 🛠 Getting Started
 
-The app has a React/Vite frontend and a FastAPI backend. You will need an OpenAI API key with access to the GPT-4 Vision API.
+The app has a React/Vite frontend and a FastAPI backend.
+
+Keys needed:
+
+- [OpenAI API key with access to GPT-4](https://github.com/abi/screenshot-to-code/blob/main/Troubleshooting.md)
+- Anthropic key (optional) - only if you want to use Claude Sonnet, or for experimental video support.
 
 Run the backend (I use Poetry for package management - `pip install poetry` if you don't have it):
 
@@ -30,6 +56,8 @@ poetry shell
 poetry run uvicorn main:app --reload --port 7001
 ```
 
+If you want to use Anthropic, add `ANTHROPIC_API_KEY` to `backend/.env`. You can also set up the keys using the settings dialog on the front-end (click the gear icon after loading the frontend).
+
 Run the frontend:
 
 ```bash
@@ -48,17 +76,6 @@ For debugging purposes, if you don't want to waste GPT4-Vision credits, you can
 MOCK=true poetry run uvicorn main:app --reload --port 7001
 ```
 
-## Configuration
-
-- You can configure the OpenAI base URL if you need to use a proxy: Set OPENAI_BASE_URL in the `backend/.env` or directly in the UI in the settings dialog
-
-## Using Claude 3
-
-We recently added support for Claude 3 Sonnet. It performs well, on par or better than GPT-4 vision for many inputs, and it tends to be faster. 
-
-1. Add an env var `ANTHROPIC_API_KEY` to `backend/.env` with your API key from Anthropic
-2. When using the front-end, select "Claude 3 Sonnet" from the model dropdown
-
 ## Docker
 
 If you have Docker installed on your system, in the root directory, run:
@@ -74,6 +91,9 @@ The app will be up and running at http://localhost:5173. Note that you can't dev
 
 - **I'm running into an error when setting up the backend. How can I fix it?** [Try this](https://github.com/abi/screenshot-to-code/issues/3#issuecomment-1814777959). If that still doesn't work, open an issue.
 - **How do I get an OpenAI API key?** See https://github.com/abi/screenshot-to-code/blob/main/Troubleshooting.md
+- **How can I configure an OpenAI proxy?** - If you're not able to access the OpenAI API directly (due to e.g. country restrictions), you can try a VPN or you can configure the OpenAI base URL to use a proxy: Set OPENAI_BASE_URL in the `backend/.env` or directly in the UI in the settings dialog. Make sure the URL has "v1" in the path so it should look like this: `https://xxx.xxxxx.xxx/v1`
+- **How can I update the backend host that my front-end connects to?** - Configure VITE_HTTP_BACKEND_URL and VITE_WS_BACKEND_URL in front/.env.local For example, set VITE_HTTP_BACKEND_URL=http://124.10.20.1:7001
+- **Seeing UTF-8 errors when running the backend?** - On windows, open the .env file with notepad++, then go to Encoding and select UTF-8.
 - **How can I provide feedback?** For feedback, feature requests and bug reports, open an issue or ping me on [Twitter](https://twitter.com/_abi_).
 
 ## 📚 Examples
@@ -94,6 +114,4 @@ https://github.com/abi/screenshot-to-code/assets/23818/3fec0f77-44e8-4fb3-a769-a
 
 ## 🌍 Hosted Version
 
-🆕 [Try it here](https://screenshottocode.com) (bring your own OpenAI key - **your key must have access to GPT-4 Vision. See [FAQ](#%EF%B8%8F-faqs) section for details**). Or see [Getting Started](#-getting-started) for local install instructions.
-
-[!["Buy Me A Coffee"](https://www.buymeacoffee.com/assets/img/custom_images/orange_img.png)](https://www.buymeacoffee.com/abiraja)
+🆕 [Try it here (paid)](https://screenshottocode.com). Or see [Getting Started](#-getting-started) for local install instructions to use with your own API keys.

Troubleshooting.md

@@ -1,4 +1,4 @@
-### Getting an OpenAI API key with GPT4-Vision model access
+### Getting an OpenAI API key with GPT-4 model access
 
 You don't need a ChatGPT Pro account. Screenshot to code uses API keys from your OpenAI developer account. In order to get access to the GPT4 Vision model, log into your OpenAI account and then, follow these instructions:
 
@@ -11,8 +11,12 @@ You don't need a ChatGPT Pro account. Screenshot to code uses API keys from your
 5. Go to Settings > Limits and check at the bottom of the page, your current tier has to be "Tier 1" to have GPT4 access
 <img width="900" alt="285636973-da38bd4d-8a78-4904-8027-ca67d729b933" src="https://github.com/abi/screenshot-to-code/assets/23818/8d07cd84-0cf9-4f88-bc00-80eba492eadf">
 
-6. Go to Screenshot to code and paste it in the Settings dialog under OpenAI key (gear icon). Your key is only stored in your browser. Never stored on our servers.
+6. Navigate to OpenAI [api keys](https://platform.openai.com/api-keys) page and create and copy a new secret key.
+7. Go to Screenshot to code and paste it in the Settings dialog under OpenAI key (gear icon). Your key is only stored in your browser. Never stored on our servers.
 
-Some users have also reported that it can take upto 30 minutes after your credit purchase for the GPT4 vision model to be activated.
+## Still not working?
 
-If you've followed these steps, and it still doesn't work, feel free to open a Github issue.
+- Some users have also reported that it can take upto 30 minutes after your credit purchase for the GPT4 vision model to be activated.
+- You need to add credits to your account AND set it to renew when credits run out in order to be upgraded to Tier 1. Make sure your "Settings > Limits" page shows that you are at Tier 1.
+
+If you've followed these steps, and it still doesn't work, feel free to open a Github issue. We only provide support for the open source version since we don't have debugging logs on the hosted version. If you're looking to use the hosted version, we recommend getting a paid subscription on screenshottocode.com

backend/.pre-commit-config.yaml

@@ -7,19 +7,19 @@ repos:
       - id: end-of-file-fixer
       - id: check-yaml
       - id: check-added-large-files
-  - repo: local
-    hooks:
-      - id: poetry-pytest
-        name: Run pytest with Poetry
-        entry: poetry run --directory backend pytest
-        language: system
-        pass_filenames: false
-        always_run: true
-        files: ^backend/
-      # - id: poetry-pyright
-      #   name: Run pyright with Poetry
-      #   entry: poetry run --directory backend pyright
-      #   language: system
-      #   pass_filenames: false
-      #   always_run: true
-      #   files: ^backend/
+  # - repo: local
+  #   hooks:
+  #     - id: poetry-pytest
+  #       name: Run pytest with Poetry
+  #       entry: poetry run --directory backend pytest
+  #       language: system
+  #       pass_filenames: false
+  #       always_run: true
+  #       files: ^backend/
+  #     # - id: poetry-pyright
+  #     #   name: Run pyright with Poetry
+  #     #   entry: poetry run --directory backend pyright
+  #     #   language: system
+  #     #   pass_filenames: false
+  #     #   always_run: true
+  #     #   files: ^backend/

backend/Dockerfile

@@ -1,4 +1,4 @@
-FROM python:3.12-slim-bullseye
+FROM python:3.12.3-slim-bullseye
 
 ENV POETRY_VERSION 1.4.1
 

backend/access_token.py

@@ -1,21 +0,0 @@
-import json
-import os
-import httpx
-
-
-async def validate_access_token(access_code: str):
-    async with httpx.AsyncClient() as client:
-        url = (
-            "https://backend.buildpicoapps.com/screenshot_to_code/validate_access_token"
-        )
-        data = json.dumps(
-            {
-                "access_code": access_code,
-                "secret": os.environ.get("PICO_BACKEND_SECRET"),
-            }
-        )
-        headers = {"Content-Type": "application/json"}
-
-        response = await client.post(url, content=data, headers=headers)
-        response_data = response.json()
-        return response_data

backend/codegen/test_utils.py

@@ -0,0 +1,57 @@
+import unittest
+from codegen.utils import extract_html_content
+
+
+class TestUtils(unittest.TestCase):
+
+    def test_extract_html_content_with_html_tags(self):
+        text = "<html><body><p>Hello, World!</p></body></html>"
+        expected = "<html><body><p>Hello, World!</p></body></html>"
+        result = extract_html_content(text)
+        self.assertEqual(result, expected)
+
+    def test_extract_html_content_without_html_tags(self):
+        text = "No HTML content here."
+        expected = "No HTML content here."
+        result = extract_html_content(text)
+        self.assertEqual(result, expected)
+
+    def test_extract_html_content_with_partial_html_tags(self):
+        text = "<html><body><p>Hello, World!</p></body>"
+        expected = "<html><body><p>Hello, World!</p></body>"
+        result = extract_html_content(text)
+        self.assertEqual(result, expected)
+
+    def test_extract_html_content_with_multiple_html_tags(self):
+        text = "<html><body><p>First</p></body></html> Some text <html><body><p>Second</p></body></html>"
+        expected = "<html><body><p>First</p></body></html>"
+        result = extract_html_content(text)
+        self.assertEqual(result, expected)
+
+    ## The following are tests based on actual LLM outputs
+
+    def test_extract_html_content_some_explanation_before(self):
+        text = """Got it! You want the song list to be displayed horizontally. I'll update the code to ensure that the song list is displayed in a horizontal layout.
+
+        Here's the updated code:
+
+        <html lang="en"><head></head><body class="bg-black text-white"></body></html>"""
+        expected = '<html lang="en"><head></head><body class="bg-black text-white"></body></html>'
+        result = extract_html_content(text)
+        self.assertEqual(result, expected)
+
+    def test_markdown_tags(self):
+        text = "```html<head></head>```"
+        expected = "```html<head></head>```"
+        result = extract_html_content(text)
+        self.assertEqual(result, expected)
+
+    def test_doctype_text(self):
+        text = '<!DOCTYPE html><html lang="en"><head></head><body></body></html>'
+        expected = '<html lang="en"><head></head><body></body></html>'
+        result = extract_html_content(text)
+        self.assertEqual(result, expected)
+
+
+if __name__ == "__main__":
+    unittest.main()

backend/codegen/utils.py

@@ -0,0 +1,14 @@
+import re
+
+
+def extract_html_content(text: str):
+    # Use regex to find content within <html> tags and include the tags themselves
+    match = re.search(r"(<html.*?>.*?</html>)", text, re.DOTALL)
+    if match:
+        return match.group(1)
+    else:
+        # Otherwise, we just send the previous HTML over
+        print(
+            "[HTML Extraction] No <html> tags found in the generated content: " + text
+        )
+        return text

backend/config.py

@@ -3,9 +3,21 @@
 # TODO: Should only be set to true when value is 'True', not any abitrary truthy value
 import os
 
+NUM_VARIANTS = 2
+
+# LLM-related
+OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY", None)
 ANTHROPIC_API_KEY = os.environ.get("ANTHROPIC_API_KEY", None)
+OPENAI_BASE_URL = os.environ.get("OPENAI_BASE_URL", None)
+
+# Image generation (optional)
+REPLICATE_API_KEY = os.environ.get("REPLICATE_API_KEY", None)
+
+# Debugging-related
 
 SHOULD_MOCK_AI_RESPONSE = bool(os.environ.get("MOCK", False))
+IS_DEBUG_ENABLED = bool(os.environ.get("IS_DEBUG_ENABLED", False))
+DEBUG_DIR = os.environ.get("DEBUG_DIR", "")
 
 # Set to True when running in production (on the hosted version)
 # Used as a feature flag to enable or disable certain features

backend/debug/DebugFileWriter.py

@@ -0,0 +1,30 @@
+import os
+import logging
+import uuid
+
+from config import DEBUG_DIR, IS_DEBUG_ENABLED
+
+
+class DebugFileWriter:
+    def __init__(self):
+        if not IS_DEBUG_ENABLED:
+            return
+
+        try:
+            self.debug_artifacts_path = os.path.expanduser(
+                f"{DEBUG_DIR}/{str(uuid.uuid4())}"
+            )
+            os.makedirs(self.debug_artifacts_path, exist_ok=True)
+            print(f"Debugging artifacts will be stored in: {self.debug_artifacts_path}")
+        except:
+            logging.error("Failed to create debug directory")
+
+    def write_to_file(self, filename: str, content: str) -> None:
+        try:
+            with open(os.path.join(self.debug_artifacts_path, filename), "w") as file:
+                file.write(content)
+        except Exception as e:
+            logging.error(f"Failed to write to file: {e}")
+
+    def extract_html_content(self, text: str) -> str:
+        return str(text.split("<html>")[-1].rsplit("</html>", 1)[0] + "</html>")

backend/evals/core.py

@@ -1,40 +1,42 @@
-import os
-from config import ANTHROPIC_API_KEY
-
-from llm import stream_claude_response, stream_openai_response
+from config import ANTHROPIC_API_KEY, OPENAI_API_KEY
+from llm import Llm, stream_claude_response, stream_openai_response
 from prompts import assemble_prompt
 from prompts.types import Stack
+from openai.types.chat import ChatCompletionMessageParam
 
 
-async def generate_code_core(image_url: str, stack: Stack) -> str:
-    model = "CLAUDE"
-
+async def generate_code_for_image(image_url: str, stack: Stack, model: Llm) -> str:
     prompt_messages = assemble_prompt(image_url, stack)
-    openai_api_key = os.environ.get("OPENAI_API_KEY")
-    anthropic_api_key = ANTHROPIC_API_KEY
-    openai_base_url = None
+    return await generate_code_core(prompt_messages, model)
 
-    async def process_chunk(content: str):
+
+async def generate_code_core(
+    prompt_messages: list[ChatCompletionMessageParam], model: Llm
+) -> str:
+
+    async def process_chunk(_: str):
         pass
 
-    if model == "CLAUDE":
-        if not anthropic_api_key:
+    if model == Llm.CLAUDE_3_SONNET or model == Llm.CLAUDE_3_5_SONNET_2024_06_20:
+        if not ANTHROPIC_API_KEY:
             raise Exception("Anthropic API key not found")
 
         completion = await stream_claude_response(
             prompt_messages,
-            api_key=anthropic_api_key,
+            api_key=ANTHROPIC_API_KEY,
             callback=lambda x: process_chunk(x),
+            model=model,
         )
     else:
-        if not openai_api_key:
+        if not OPENAI_API_KEY:
             raise Exception("OpenAI API key not found")
 
         completion = await stream_openai_response(
             prompt_messages,
-            api_key=openai_api_key,
-            base_url=openai_base_url,
+            api_key=OPENAI_API_KEY,
+            base_url=None,
             callback=lambda x: process_chunk(x),
+            model=model,
         )
 
     return completion

backend/fs_logging/core.py

@@ -0,0 +1,23 @@
+from datetime import datetime
+import json
+import os
+from openai.types.chat import ChatCompletionMessageParam
+
+
+def write_logs(prompt_messages: list[ChatCompletionMessageParam], completion: str):
+    # Get the logs path from environment, default to the current working directory
+    logs_path = os.environ.get("LOGS_PATH", os.getcwd())
+
+    # Create run_logs directory if it doesn't exist within the specified logs path
+    logs_directory = os.path.join(logs_path, "run_logs")
+    if not os.path.exists(logs_directory):
+        os.makedirs(logs_directory)
+
+    print("Writing to logs directory:", logs_directory)
+
+    # Generate a unique filename using the current timestamp within the logs directory
+    filename = datetime.now().strftime(f"{logs_directory}/messages_%Y%m%d_%H%M%S.json")
+
+    # Write the messages dict into a new file for each run
+    with open(filename, "w") as f:
+        f.write(json.dumps({"prompt": prompt_messages, "completion": completion}))

backend/image_generation/core.py

@@ -1,40 +1,72 @@
 import asyncio
 import re
-from typing import Dict, List, Union
+from typing import Dict, List, Literal, Union
 from openai import AsyncOpenAI
 from bs4 import BeautifulSoup
 
+from image_generation.replicate import call_replicate
 
-async def process_tasks(prompts: List[str], api_key: str, base_url: str):
-    tasks = [generate_image(prompt, api_key, base_url) for prompt in prompts]
+
+async def process_tasks(
+    prompts: List[str],
+    api_key: str,
+    base_url: str | None,
+    model: Literal["dalle3", "flux"],
+):
+    import time
+
+    start_time = time.time()
+    if model == "dalle3":
+        tasks = [generate_image_dalle(prompt, api_key, base_url) for prompt in prompts]
+    else:
+        tasks = [generate_image_replicate(prompt, api_key) for prompt in prompts]
     results = await asyncio.gather(*tasks, return_exceptions=True)
+    end_time = time.time()
+    generation_time = end_time - start_time
+    print(f"Image generation time: {generation_time:.2f} seconds")
 
     processed_results: List[Union[str, None]] = []
     for result in results:
-        if isinstance(result, Exception):
+        if isinstance(result, BaseException):
             print(f"An exception occurred: {result}")
             processed_results.append(None)
         else:
-            processed_results.append(result)  # type: ignore
+            processed_results.append(result)
 
     return processed_results
 
 
-async def generate_image(prompt: str, api_key: str, base_url: str):
+async def generate_image_dalle(
+    prompt: str, api_key: str, base_url: str | None
+) -> Union[str, None]:
     client = AsyncOpenAI(api_key=api_key, base_url=base_url)
-    image_params: Dict[str, Union[str, int]] = {
-        "model": "dall-e-3",
-        "quality": "standard",
-        "style": "natural",
-        "n": 1,
-        "size": "1024x1024",
-        "prompt": prompt,
-    }
-    res = await client.images.generate(**image_params)  # type: ignore
+    res = await client.images.generate(
+        model="dall-e-3",
+        quality="standard",
+        style="natural",
+        n=1,
+        size="1024x1024",
+        prompt=prompt,
+    )
     await client.close()
     return res.data[0].url
 
 
+async def generate_image_replicate(prompt: str, api_key: str) -> str:
+
+    # We use Flux Schnell
+    return await call_replicate(
+        {
+            "prompt": prompt,
+            "num_outputs": 1,
+            "aspect_ratio": "1:1",
+            "output_format": "png",
+            "output_quality": 100,
+        },
+        api_key,
+    )
+
+
 def extract_dimensions(url: str):
     # Regular expression to match numbers in the format '300x200'
     matches = re.findall(r"(\d+)x(\d+)", url)
@@ -62,14 +94,18 @@ def create_alt_url_mapping(code: str) -> Dict[str, str]:
 
 
 async def generate_images(
-    code: str, api_key: str, base_url: Union[str, None], image_cache: Dict[str, str]
-):
+    code: str,
+    api_key: str,
+    base_url: Union[str, None],
+    image_cache: Dict[str, str],
+    model: Literal["dalle3", "flux"] = "dalle3",
+) -> str:
     # Find all images
     soup = BeautifulSoup(code, "html.parser")
     images = soup.find_all("img")
 
     # Extract alt texts as image prompts
-    alts = []
+    alts: List[str | None] = []
     for img in images:
         # Only include URL if the image starts with https://placehold.co
         # and it's not already in the image_cache
@@ -77,26 +113,26 @@ async def generate_images(
             img["src"].startswith("https://placehold.co")
             and image_cache.get(img.get("alt")) is None
         ):
-            alts.append(img.get("alt", None))  # type: ignore
+            alts.append(img.get("alt", None))
 
     # Exclude images with no alt text
-    alts = [alt for alt in alts if alt is not None]  # type: ignore
+    filtered_alts: List[str] = [alt for alt in alts if alt is not None]
 
     # Remove duplicates
-    prompts = list(set(alts))  # type: ignore
+    prompts = list(set(filtered_alts))
 
     # Return early if there are no images to replace
-    if len(prompts) == 0:  # type: ignore
+    if len(prompts) == 0:
         return code
 
     # Generate images
-    results = await process_tasks(prompts, api_key, base_url)  # type: ignore
+    results = await process_tasks(prompts, api_key, base_url, model)
 
     # Create a dict mapping alt text to image URL
-    mapped_image_urls = dict(zip(prompts, results))  # type: ignore
+    mapped_image_urls = dict(zip(prompts, results))
 
     # Merge with image_cache
-    mapped_image_urls = {**mapped_image_urls, **image_cache}  # type: ignore
+    mapped_image_urls = {**mapped_image_urls, **image_cache}
 
     # Replace old image URLs with the generated URLs
     for img in images:

backend/image_generation/replicate.py

@@ -0,0 +1,65 @@
+import asyncio
+import httpx
+
+
+async def call_replicate(input: dict[str, str | int], api_token: str) -> str:
+    headers = {
+        "Authorization": f"Bearer {api_token}",
+        "Content-Type": "application/json",
+    }
+
+    data = {"input": input}
+
+    async with httpx.AsyncClient() as client:
+        try:
+            response = await client.post(
+                "https://api.replicate.com/v1/models/black-forest-labs/flux-schnell/predictions",
+                headers=headers,
+                json=data,
+            )
+            response.raise_for_status()
+            response_json = response.json()
+
+            # Extract the id from the response
+            prediction_id = response_json.get("id")
+            if not prediction_id:
+                raise ValueError("Prediction ID not found in initial response.")
+
+            # Polling every 0.1 seconds until the status is succeeded or error (upto 10s)
+            num_polls = 0
+            max_polls = 100
+            while num_polls < max_polls:
+                num_polls += 1
+
+                await asyncio.sleep(0.1)
+
+                # Check the status
+                status_check_url = (
+                    f"https://api.replicate.com/v1/predictions/{prediction_id}"
+                )
+                status_response = await client.get(status_check_url, headers=headers)
+                status_response.raise_for_status()
+                status_response_json = status_response.json()
+                status = status_response_json.get("status")
+
+                # If status is succeeded or if there's an error, break out of the loop
+                if status == "succeeded":
+                    return status_response_json["output"][0]
+                elif status == "error":
+                    raise ValueError(
+                        f"Inference errored out: {status_response_json.get('error', 'Unknown error')}"
+                    )
+                elif status == "failed":
+                    raise ValueError("Inference failed")
+
+            # If we've reached here, it means we've exceeded the max number of polls
+            raise TimeoutError("Inference timed out")
+
+        except httpx.HTTPStatusError as e:
+            raise ValueError(f"HTTP error occurred: {e}")
+        except httpx.RequestError as e:
+            raise ValueError(f"An error occurred while requesting: {e}")
+        except asyncio.TimeoutError:
+            raise TimeoutError("Request timed out")
+        except Exception as e:
+            raise ValueError(f"An unexpected error occurred: {e}")

backend/image_processing/utils.py

@@ -0,0 +1,80 @@
+import base64
+import io
+import time
+from PIL import Image
+
+CLAUDE_IMAGE_MAX_SIZE = 5 * 1024 * 1024
+CLAUDE_MAX_IMAGE_DIMENSION = 7990
+
+
+# Process image so it meets Claude requirements
+def process_image(image_data_url: str) -> tuple[str, str]:
+
+    # Extract bytes and media type from base64 data URL
+    media_type = image_data_url.split(";")[0].split(":")[1]
+    base64_data = image_data_url.split(",")[1]
+    image_bytes = base64.b64decode(base64_data)
+
+    img = Image.open(io.BytesIO(image_bytes))
+
+    # Check if image is under max dimensions and size
+    is_under_dimension_limit = (
+        img.width < CLAUDE_MAX_IMAGE_DIMENSION
+        and img.height < CLAUDE_MAX_IMAGE_DIMENSION
+    )
+    is_under_size_limit = len(base64_data) <= CLAUDE_IMAGE_MAX_SIZE
+
+    # If image is under both limits, no processing needed
+    if is_under_dimension_limit and is_under_size_limit:
+        print("[CLAUDE IMAGE PROCESSING] no processing needed")
+        return (media_type, base64_data)
+
+    # Time image processing
+    start_time = time.time()
+
+    # Check if either dimension exceeds 7900px (Claude disallows >= 8000px)
+    # Resize image if needed
+    if not is_under_dimension_limit:
+        # Calculate the new dimensions while maintaining aspect ratio
+        if img.width > img.height:
+            new_width = CLAUDE_MAX_IMAGE_DIMENSION
+            new_height = int((CLAUDE_MAX_IMAGE_DIMENSION / img.width) * img.height)
+        else:
+            new_height = CLAUDE_MAX_IMAGE_DIMENSION
+            new_width = int((CLAUDE_MAX_IMAGE_DIMENSION / img.height) * img.width)
+
+        # Resize the image
+        img = img.resize((new_width, new_height), Image.DEFAULT_STRATEGY)
+        print(
+            f"[CLAUDE IMAGE PROCESSING] image resized: width = {new_width}, height = {new_height}"
+        )
+
+    # Convert and compress as JPEG
+    # We always compress as JPEG (95% at the least) even when we resize and the original image
+    # is under the size limit.
+    quality = 95
+    output = io.BytesIO()
+    img = img.convert("RGB")  # Ensure image is in RGB mode for JPEG conversion
+    img.save(output, format="JPEG", quality=quality)
+
+    # Reduce quality until image is under max size
+    while (
+        len(base64.b64encode(output.getvalue())) > CLAUDE_IMAGE_MAX_SIZE
+        and quality > 10
+    ):
+        output = io.BytesIO()
+        img.save(output, format="JPEG", quality=quality)
+        quality -= 5
+
+    # Log so we know it was modified
+    old_size = len(base64_data)
+    new_size = len(base64.b64encode(output.getvalue()))
+    print(
+        f"[CLAUDE IMAGE PROCESSING] image size updated: old size = {old_size} bytes, new size = {new_size} bytes"
+    )
+
+    end_time = time.time()
+    processing_time = end_time - start_time
+    print(f"[CLAUDE IMAGE PROCESSING] processing time: {processing_time:.2f} seconds")
+
+    return ("image/jpeg", base64.b64encode(output.getvalue()).decode("utf-8"))

backend/llm.py

@@ -1,18 +1,35 @@
+import copy
+from enum import Enum
 from typing import Any, Awaitable, Callable, List, cast
 from anthropic import AsyncAnthropic
 from openai import AsyncOpenAI
 from openai.types.chat import ChatCompletionMessageParam, ChatCompletionChunk
+from config import IS_DEBUG_ENABLED
+from debug.DebugFileWriter import DebugFileWriter
+from image_processing.utils import process_image
 
-MODEL_GPT_4_VISION = "gpt-4-vision-preview"
-MODEL_CLAUDE_SONNET = "claude-3-sonnet-20240229"
-MODEL_CLAUDE_OPUS = "claude-3-opus-20240229"
+from utils import pprint_prompt
 
 
-# Keep in sync with frontend (lib/models.ts)
-CODE_GENERATION_MODELS = [
-    "gpt_4_vision",
-    "claude_3_sonnet",
-]
+# Actual model versions that are passed to the LLMs and stored in our logs
+class Llm(Enum):
+    GPT_4_VISION = "gpt-4-vision-preview"
+    GPT_4_TURBO_2024_04_09 = "gpt-4-turbo-2024-04-09"
+    GPT_4O_2024_05_13 = "gpt-4o-2024-05-13"
+    CLAUDE_3_SONNET = "claude-3-sonnet-20240229"
+    CLAUDE_3_OPUS = "claude-3-opus-20240229"
+    CLAUDE_3_HAIKU = "claude-3-haiku-20240307"
+    CLAUDE_3_5_SONNET_2024_06_20 = "claude-3-5-sonnet-20240620"
+
+
+# Will throw errors if you send a garbage string
+def convert_frontend_str_to_llm(frontend_str: str) -> Llm:
+    if frontend_str == "gpt_4_vision":
+        return Llm.GPT_4_VISION
+    elif frontend_str == "claude_3_sonnet":
+        return Llm.CLAUDE_3_SONNET
+    else:
+        return Llm(frontend_str)
 
 
 async def stream_openai_response(
@@ -20,26 +37,40 @@ async def stream_openai_response(
     api_key: str,
     base_url: str | None,
     callback: Callable[[str], Awaitable[None]],
+    model: Llm,
 ) -> str:
     client = AsyncOpenAI(api_key=api_key, base_url=base_url)
 
-    model = MODEL_GPT_4_VISION
-
     # Base parameters
-    params = {"model": model, "messages": messages, "stream": True, "timeout": 600}
+    params = {
+        "model": model.value,
+        "messages": messages,
+        "stream": True,
+        "timeout": 600,
+        "temperature": 0.0,
+    }
 
-    # Add 'max_tokens' only if the model is a GPT4 vision model
-    if model == MODEL_GPT_4_VISION:
+    # Add 'max_tokens' only if the model is a GPT4 vision or Turbo model
+    if (
+        model == Llm.GPT_4_VISION
+        or model == Llm.GPT_4_TURBO_2024_04_09
+        or model == Llm.GPT_4O_2024_05_13
+    ):
         params["max_tokens"] = 4096
-        params["temperature"] = 0
 
     stream = await client.chat.completions.create(**params)  # type: ignore
     full_response = ""
     async for chunk in stream:  # type: ignore
         assert isinstance(chunk, ChatCompletionChunk)
-        content = chunk.choices[0].delta.content or ""
-        full_response += content
-        await callback(content)
+        if (
+            chunk.choices
+            and len(chunk.choices) > 0
+            and chunk.choices[0].delta
+            and chunk.choices[0].delta.content
+        ):
+            content = chunk.choices[0].delta.content or ""
+            full_response += content
+            await callback(content)
 
     await client.close()
 
@@ -51,18 +82,22 @@ async def stream_claude_response(
     messages: List[ChatCompletionMessageParam],
     api_key: str,
     callback: Callable[[str], Awaitable[None]],
+    model: Llm,
 ) -> str:
 
     client = AsyncAnthropic(api_key=api_key)
 
     # Base parameters
-    model = MODEL_CLAUDE_SONNET
-    max_tokens = 4096
+    max_tokens = 8192
     temperature = 0.0
 
     # Translate OpenAI messages to Claude messages
-    system_prompt = cast(str, messages[0]["content"])
-    claude_messages = [dict(message) for message in messages[1:]]
+
+    # Deep copy messages to avoid modifying the original list
+    cloned_messages = copy.deepcopy(messages)
+
+    system_prompt = cast(str, cloned_messages[0].get("content"))
+    claude_messages = [dict(message) for message in cloned_messages[1:]]
     for message in claude_messages:
         if not isinstance(message["content"], list):
             continue
@@ -74,8 +109,10 @@ async def stream_claude_response(
                 # Extract base64 data and media type from data URL
                 # Example base64 data URL: data:image/png;base64,iVBOR...
                 image_data_url = cast(str, content["image_url"]["url"])
-                media_type = image_data_url.split(";")[0].split(":")[1]
-                base64_data = image_data_url.split(",")[1]
+
+                # Process image and split media type and data
+                # so it works with Claude (under 5mb in base64 encoding)
+                (media_type, base64_data) = process_image(image_data_url)
 
                 # Remove OpenAI parameter
                 del content["image_url"]
@@ -88,17 +125,22 @@ async def stream_claude_response(
 
     # Stream Claude response
     async with client.messages.stream(
-        model=model,
+        model=model.value,
         max_tokens=max_tokens,
         temperature=temperature,
         system=system_prompt,
         messages=claude_messages,  # type: ignore
+        extra_headers={"anthropic-beta": "max-tokens-3-5-sonnet-2024-07-15"},
     ) as stream:
         async for text in stream.text_stream:
             await callback(text)
 
     # Return final message
     response = await stream.get_final_message()
+
+    # Close the Anthropic client
+    await client.close()
+
     return response.content[0].text
 
 
@@ -108,39 +150,84 @@ async def stream_claude_response_native(
     api_key: str,
     callback: Callable[[str], Awaitable[None]],
     include_thinking: bool = False,
-    model: str = MODEL_CLAUDE_OPUS,
+    model: Llm = Llm.CLAUDE_3_OPUS,
 ) -> str:
 
     client = AsyncAnthropic(api_key=api_key)
 
-    # Base parameters
+    # Base model parameters
     max_tokens = 4096
     temperature = 0.0
 
-    # Stream Claude response
+    # Multi-pass flow
+    current_pass_num = 1
+    max_passes = 2
 
-    # Set up message depending on whether we have a <thinking> prefix
-    messages = (
-        messages + [{"role": "assistant", "content": "<thinking>"}]
-        if include_thinking
-        else messages
-    )
+    prefix = "<thinking>"
+    response = None
 
-    async with client.messages.stream(
-        model=model,
-        max_tokens=max_tokens,
-        temperature=temperature,
-        system=system_prompt,
-        messages=messages,  # type: ignore
-    ) as stream:
-        async for text in stream.text_stream:
-            await callback(text)
+    # For debugging
+    full_stream = ""
+    debug_file_writer = DebugFileWriter()
 
-    # Return final message
-    response = await stream.get_final_message()
+    while current_pass_num <= max_passes:
+        current_pass_num += 1
 
-    print(
-        f"Token usage: Input Tokens: {response.usage.input_tokens}, Output Tokens: {response.usage.output_tokens}"
-    )
+        # Set up message depending on whether we have a <thinking> prefix
+        messages_to_send = (
+            messages + [{"role": "assistant", "content": prefix}]
+            if include_thinking
+            else messages
+        )
 
-    return response.content[0].text
+        pprint_prompt(messages_to_send)
+
+        async with client.messages.stream(
+            model=model.value,
+            max_tokens=max_tokens,
+            temperature=temperature,
+            system=system_prompt,
+            messages=messages_to_send,  # type: ignore
+        ) as stream:
+            async for text in stream.text_stream:
+                print(text, end="", flush=True)
+                full_stream += text
+                await callback(text)
+
+        response = await stream.get_final_message()
+        response_text = response.content[0].text
+
+        # Write each pass's code to .html file and thinking to .txt file
+        if IS_DEBUG_ENABLED:
+            debug_file_writer.write_to_file(
+                f"pass_{current_pass_num - 1}.html",
+                debug_file_writer.extract_html_content(response_text),
+            )
+            debug_file_writer.write_to_file(
+                f"thinking_pass_{current_pass_num - 1}.txt",
+                response_text.split("</thinking>")[0],
+            )
+
+        # Set up messages array for next pass
+        messages += [
+            {"role": "assistant", "content": str(prefix) + response.content[0].text},
+            {
+                "role": "user",
+                "content": "You've done a good job with a first draft. Improve this further based on the original instructions so that the app is fully functional and looks like the original video of the app we're trying to replicate.",
+            },
+        ]
+
+        print(
+            f"Token usage: Input Tokens: {response.usage.input_tokens}, Output Tokens: {response.usage.output_tokens}"
+        )
+
+    # Close the Anthropic client
+    await client.close()
+
+    if IS_DEBUG_ENABLED:
+        debug_file_writer.write_to_file("full_stream.txt", full_stream)
+
+    if not response:
+        raise Exception("No HTML response found in AI response")
+    else:
+        return response.content[0].text

backend/poetry.lock

@@ -1,5 +1,115 @@
 # This file is automatically @generated by Poetry 1.7.1 and should not be changed by hand.
 
+[[package]]
+name = "aiohttp"
+version = "3.9.5"
+description = "Async http client/server framework (asyncio)"
+optional = false
+python-versions = ">=3.8"
+files = [
+    {file = "aiohttp-3.9.5-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:fcde4c397f673fdec23e6b05ebf8d4751314fa7c24f93334bf1f1364c1c69ac7"},
+    {file = "aiohttp-3.9.5-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:5d6b3f1fabe465e819aed2c421a6743d8debbde79b6a8600739300630a01bf2c"},
+    {file = "aiohttp-3.9.5-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:6ae79c1bc12c34082d92bf9422764f799aee4746fd7a392db46b7fd357d4a17a"},
+    {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4d3ebb9e1316ec74277d19c5f482f98cc65a73ccd5430540d6d11682cd857430"},
+    {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:84dabd95154f43a2ea80deffec9cb44d2e301e38a0c9d331cc4aa0166fe28ae3"},
+    {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:c8a02fbeca6f63cb1f0475c799679057fc9268b77075ab7cf3f1c600e81dd46b"},
+    {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c26959ca7b75ff768e2776d8055bf9582a6267e24556bb7f7bd29e677932be72"},
+    {file = "aiohttp-3.9.5-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:714d4e5231fed4ba2762ed489b4aec07b2b9953cf4ee31e9871caac895a839c0"},
+    {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:e7a6a8354f1b62e15d48e04350f13e726fa08b62c3d7b8401c0a1314f02e3558"},
+    {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:c413016880e03e69d166efb5a1a95d40f83d5a3a648d16486592c49ffb76d0db"},
+    {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:ff84aeb864e0fac81f676be9f4685f0527b660f1efdc40dcede3c251ef1e867f"},
+    {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:ad7f2919d7dac062f24d6f5fe95d401597fbb015a25771f85e692d043c9d7832"},
+    {file = "aiohttp-3.9.5-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:702e2c7c187c1a498a4e2b03155d52658fdd6fda882d3d7fbb891a5cf108bb10"},
+    {file = "aiohttp-3.9.5-cp310-cp310-win32.whl", hash = "sha256:67c3119f5ddc7261d47163ed86d760ddf0e625cd6246b4ed852e82159617b5fb"},
+    {file = "aiohttp-3.9.5-cp310-cp310-win_amd64.whl", hash = "sha256:471f0ef53ccedec9995287f02caf0c068732f026455f07db3f01a46e49d76bbb"},
+    {file = "aiohttp-3.9.5-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:e0ae53e33ee7476dd3d1132f932eeb39bf6125083820049d06edcdca4381f342"},
+    {file = "aiohttp-3.9.5-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:c088c4d70d21f8ca5c0b8b5403fe84a7bc8e024161febdd4ef04575ef35d474d"},
+    {file = "aiohttp-3.9.5-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:639d0042b7670222f33b0028de6b4e2fad6451462ce7df2af8aee37dcac55424"},
+    {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f26383adb94da5e7fb388d441bf09c61e5e35f455a3217bfd790c6b6bc64b2ee"},
+    {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:66331d00fb28dc90aa606d9a54304af76b335ae204d1836f65797d6fe27f1ca2"},
+    {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:4ff550491f5492ab5ed3533e76b8567f4b37bd2995e780a1f46bca2024223233"},
+    {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f22eb3a6c1080d862befa0a89c380b4dafce29dc6cd56083f630073d102eb595"},
+    {file = "aiohttp-3.9.5-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:a81b1143d42b66ffc40a441379387076243ef7b51019204fd3ec36b9f69e77d6"},
+    {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:f64fd07515dad67f24b6ea4a66ae2876c01031de91c93075b8093f07c0a2d93d"},
+    {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:93e22add827447d2e26d67c9ac0161756007f152fdc5210277d00a85f6c92323"},
+    {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:55b39c8684a46e56ef8c8d24faf02de4a2b2ac60d26cee93bc595651ff545de9"},
+    {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:4715a9b778f4293b9f8ae7a0a7cef9829f02ff8d6277a39d7f40565c737d3771"},
+    {file = "aiohttp-3.9.5-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:afc52b8d969eff14e069a710057d15ab9ac17cd4b6753042c407dcea0e40bf75"},
+    {file = "aiohttp-3.9.5-cp311-cp311-win32.whl", hash = "sha256:b3df71da99c98534be076196791adca8819761f0bf6e08e07fd7da25127150d6"},
+    {file = "aiohttp-3.9.5-cp311-cp311-win_amd64.whl", hash = "sha256:88e311d98cc0bf45b62fc46c66753a83445f5ab20038bcc1b8a1cc05666f428a"},
+    {file = "aiohttp-3.9.5-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:c7a4b7a6cf5b6eb11e109a9755fd4fda7d57395f8c575e166d363b9fc3ec4678"},
+    {file = "aiohttp-3.9.5-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:0a158704edf0abcac8ac371fbb54044f3270bdbc93e254a82b6c82be1ef08f3c"},
+    {file = "aiohttp-3.9.5-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:d153f652a687a8e95ad367a86a61e8d53d528b0530ef382ec5aaf533140ed00f"},
+    {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:82a6a97d9771cb48ae16979c3a3a9a18b600a8505b1115cfe354dfb2054468b4"},
+    {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:60cdbd56f4cad9f69c35eaac0fbbdf1f77b0ff9456cebd4902f3dd1cf096464c"},
+    {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:8676e8fd73141ded15ea586de0b7cda1542960a7b9ad89b2b06428e97125d4fa"},
+    {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:da00da442a0e31f1c69d26d224e1efd3a1ca5bcbf210978a2ca7426dfcae9f58"},
+    {file = "aiohttp-3.9.5-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:18f634d540dd099c262e9f887c8bbacc959847cfe5da7a0e2e1cf3f14dbf2daf"},
+    {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:320e8618eda64e19d11bdb3bd04ccc0a816c17eaecb7e4945d01deee2a22f95f"},
+    {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:2faa61a904b83142747fc6a6d7ad8fccff898c849123030f8e75d5d967fd4a81"},
+    {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:8c64a6dc3fe5db7b1b4d2b5cb84c4f677768bdc340611eca673afb7cf416ef5a"},
+    {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:393c7aba2b55559ef7ab791c94b44f7482a07bf7640d17b341b79081f5e5cd1a"},
+    {file = "aiohttp-3.9.5-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:c671dc117c2c21a1ca10c116cfcd6e3e44da7fcde37bf83b2be485ab377b25da"},
+    {file = "aiohttp-3.9.5-cp312-cp312-win32.whl", hash = "sha256:5a7ee16aab26e76add4afc45e8f8206c95d1d75540f1039b84a03c3b3800dd59"},
+    {file = "aiohttp-3.9.5-cp312-cp312-win_amd64.whl", hash = "sha256:5ca51eadbd67045396bc92a4345d1790b7301c14d1848feaac1d6a6c9289e888"},
+    {file = "aiohttp-3.9.5-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:694d828b5c41255e54bc2dddb51a9f5150b4eefa9886e38b52605a05d96566e8"},
+    {file = "aiohttp-3.9.5-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:0605cc2c0088fcaae79f01c913a38611ad09ba68ff482402d3410bf59039bfb8"},
+    {file = "aiohttp-3.9.5-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:4558e5012ee03d2638c681e156461d37b7a113fe13970d438d95d10173d25f78"},
+    {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9dbc053ac75ccc63dc3a3cc547b98c7258ec35a215a92bd9f983e0aac95d3d5b"},
+    {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4109adee842b90671f1b689901b948f347325045c15f46b39797ae1bf17019de"},
+    {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:a6ea1a5b409a85477fd8e5ee6ad8f0e40bf2844c270955e09360418cfd09abac"},
+    {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f3c2890ca8c59ee683fd09adf32321a40fe1cf164e3387799efb2acebf090c11"},
+    {file = "aiohttp-3.9.5-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:3916c8692dbd9d55c523374a3b8213e628424d19116ac4308e434dbf6d95bbdd"},
+    {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:8d1964eb7617907c792ca00b341b5ec3e01ae8c280825deadbbd678447b127e1"},
+    {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:d5ab8e1f6bee051a4bf6195e38a5c13e5e161cb7bad83d8854524798bd9fcd6e"},
+    {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:52c27110f3862a1afbcb2af4281fc9fdc40327fa286c4625dfee247c3ba90156"},
+    {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:7f64cbd44443e80094309875d4f9c71d0401e966d191c3d469cde4642bc2e031"},
+    {file = "aiohttp-3.9.5-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:8b4f72fbb66279624bfe83fd5eb6aea0022dad8eec62b71e7bf63ee1caadeafe"},
+    {file = "aiohttp-3.9.5-cp38-cp38-win32.whl", hash = "sha256:6380c039ec52866c06d69b5c7aad5478b24ed11696f0e72f6b807cfb261453da"},
+    {file = "aiohttp-3.9.5-cp38-cp38-win_amd64.whl", hash = "sha256:da22dab31d7180f8c3ac7c7635f3bcd53808f374f6aa333fe0b0b9e14b01f91a"},
+    {file = "aiohttp-3.9.5-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:1732102949ff6087589408d76cd6dea656b93c896b011ecafff418c9661dc4ed"},
+    {file = "aiohttp-3.9.5-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:c6021d296318cb6f9414b48e6a439a7f5d1f665464da507e8ff640848ee2a58a"},
+    {file = "aiohttp-3.9.5-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:239f975589a944eeb1bad26b8b140a59a3a320067fb3cd10b75c3092405a1372"},
+    {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3b7b30258348082826d274504fbc7c849959f1989d86c29bc355107accec6cfb"},
+    {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:cd2adf5c87ff6d8b277814a28a535b59e20bfea40a101db6b3bdca7e9926bc24"},
+    {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e9a3d838441bebcf5cf442700e3963f58b5c33f015341f9ea86dcd7d503c07e2"},
+    {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9e3a1ae66e3d0c17cf65c08968a5ee3180c5a95920ec2731f53343fac9bad106"},
+    {file = "aiohttp-3.9.5-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:9c69e77370cce2d6df5d12b4e12bdcca60c47ba13d1cbbc8645dd005a20b738b"},
+    {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:0cbf56238f4bbf49dab8c2dc2e6b1b68502b1e88d335bea59b3f5b9f4c001475"},
+    {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:d1469f228cd9ffddd396d9948b8c9cd8022b6d1bf1e40c6f25b0fb90b4f893ed"},
+    {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:45731330e754f5811c314901cebdf19dd776a44b31927fa4b4dbecab9e457b0c"},
+    {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:3fcb4046d2904378e3aeea1df51f697b0467f2aac55d232c87ba162709478c46"},
+    {file = "aiohttp-3.9.5-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:8cf142aa6c1a751fcb364158fd710b8a9be874b81889c2bd13aa8893197455e2"},
+    {file = "aiohttp-3.9.5-cp39-cp39-win32.whl", hash = "sha256:7b179eea70833c8dee51ec42f3b4097bd6370892fa93f510f76762105568cf09"},
+    {file = "aiohttp-3.9.5-cp39-cp39-win_amd64.whl", hash = "sha256:38d80498e2e169bc61418ff36170e0aad0cd268da8b38a17c4cf29d254a8b3f1"},
+    {file = "aiohttp-3.9.5.tar.gz", hash = "sha256:edea7d15772ceeb29db4aff55e482d4bcfb6ae160ce144f2682de02f6d693551"},
+]
+
+[package.dependencies]
+aiosignal = ">=1.1.2"
+async-timeout = {version = ">=4.0,<5.0", markers = "python_version < \"3.11\""}
+attrs = ">=17.3.0"
+frozenlist = ">=1.1.1"
+multidict = ">=4.5,<7.0"
+yarl = ">=1.0,<2.0"
+
+[package.extras]
+speedups = ["Brotli", "aiodns", "brotlicffi"]
+
+[[package]]
+name = "aiosignal"
+version = "1.3.1"
+description = "aiosignal: a list of registered asynchronous callbacks"
+optional = false
+python-versions = ">=3.7"
+files = [
+    {file = "aiosignal-1.3.1-py3-none-any.whl", hash = "sha256:f8376fb07dd1e86a584e4fcdec80b36b7f81aac666ebc724e2c090300dd83b17"},
+    {file = "aiosignal-1.3.1.tar.gz", hash = "sha256:54cd96e15e1649b75d6c87526a6ff0b6c1b0dd3459f43d9ca11d48c339b68cfc"},
+]
+
+[package.dependencies]
+frozenlist = ">=1.1.0"
+
 [[package]]
 name = "anthropic"
 version = "0.18.1"
@@ -46,6 +156,36 @@ doc = ["Sphinx (>=7)", "packaging", "sphinx-autodoc-typehints (>=1.2.0)", "sphin
 test = ["anyio[trio]", "coverage[toml] (>=7)", "exceptiongroup (>=1.2.0)", "hypothesis (>=4.0)", "psutil (>=5.9)", "pytest (>=7.0)", "pytest-mock (>=3.6.1)", "trustme", "uvloop (>=0.17)"]
 trio = ["trio (>=0.23)"]
 
+[[package]]
+name = "async-timeout"
+version = "4.0.3"
+description = "Timeout context manager for asyncio programs"
+optional = false
+python-versions = ">=3.7"
+files = [
+    {file = "async-timeout-4.0.3.tar.gz", hash = "sha256:4640d96be84d82d02ed59ea2b7105a0f7b33abe8703703cd0ab0bf87c427522f"},
+    {file = "async_timeout-4.0.3-py3-none-any.whl", hash = "sha256:7405140ff1230c310e51dc27b3145b9092d659ce68ff733fb0cefe3ee42be028"},
+]
+
+[[package]]
+name = "attrs"
+version = "23.2.0"
+description = "Classes Without Boilerplate"
+optional = false
+python-versions = ">=3.7"
+files = [
+    {file = "attrs-23.2.0-py3-none-any.whl", hash = "sha256:99b87a485a5820b23b879f04c2305b44b951b502fd64be915879d77a7e8fc6f1"},
+    {file = "attrs-23.2.0.tar.gz", hash = "sha256:935dc3b529c262f6cf76e50877d35a4bd3c1de194fd41f47a2b7ae8f19971f30"},
+]
+
+[package.extras]
+cov = ["attrs[tests]", "coverage[toml] (>=5.3)"]
+dev = ["attrs[tests]", "pre-commit"]
+docs = ["furo", "myst-parser", "sphinx", "sphinx-notfound-page", "sphinxcontrib-towncrier", "towncrier", "zope-interface"]
+tests = ["attrs[tests-no-zope]", "zope-interface"]
+tests-mypy = ["mypy (>=1.6)", "pytest-mypy-plugins"]
+tests-no-zope = ["attrs[tests-mypy]", "cloudpickle", "hypothesis", "pympler", "pytest (>=4.3.0)", "pytest-xdist[psutil]"]
+
 [[package]]
 name = "beautifulsoup4"
 version = "4.12.3"
@@ -297,6 +437,92 @@ docs = ["furo (>=2023.9.10)", "sphinx (>=7.2.6)", "sphinx-autodoc-typehints (>=1
 testing = ["covdefaults (>=2.3)", "coverage (>=7.3.2)", "diff-cover (>=8)", "pytest (>=7.4.3)", "pytest-cov (>=4.1)", "pytest-mock (>=3.12)", "pytest-timeout (>=2.2)"]
 typing = ["typing-extensions (>=4.8)"]
 
+[[package]]
+name = "frozenlist"
+version = "1.4.1"
+description = "A list-like structure which implements collections.abc.MutableSequence"
+optional = false
+python-versions = ">=3.8"
+files = [
+    {file = "frozenlist-1.4.1-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:f9aa1878d1083b276b0196f2dfbe00c9b7e752475ed3b682025ff20c1c1f51ac"},
+    {file = "frozenlist-1.4.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:29acab3f66f0f24674b7dc4736477bcd4bc3ad4b896f5f45379a67bce8b96868"},
+    {file = "frozenlist-1.4.1-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:74fb4bee6880b529a0c6560885fce4dc95936920f9f20f53d99a213f7bf66776"},
+    {file = "frozenlist-1.4.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:590344787a90ae57d62511dd7c736ed56b428f04cd8c161fcc5e7232c130c69a"},
+    {file = "frozenlist-1.4.1-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:068b63f23b17df8569b7fdca5517edef76171cf3897eb68beb01341131fbd2ad"},
+    {file = "frozenlist-1.4.1-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:5c849d495bf5154cd8da18a9eb15db127d4dba2968d88831aff6f0331ea9bd4c"},
+    {file = "frozenlist-1.4.1-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:9750cc7fe1ae3b1611bb8cfc3f9ec11d532244235d75901fb6b8e42ce9229dfe"},
+    {file = "frozenlist-1.4.1-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a9b2de4cf0cdd5bd2dee4c4f63a653c61d2408055ab77b151c1957f221cabf2a"},
+    {file = "frozenlist-1.4.1-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:0633c8d5337cb5c77acbccc6357ac49a1770b8c487e5b3505c57b949b4b82e98"},
+    {file = "frozenlist-1.4.1-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:27657df69e8801be6c3638054e202a135c7f299267f1a55ed3a598934f6c0d75"},
+    {file = "frozenlist-1.4.1-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:f9a3ea26252bd92f570600098783d1371354d89d5f6b7dfd87359d669f2109b5"},
+    {file = "frozenlist-1.4.1-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:4f57dab5fe3407b6c0c1cc907ac98e8a189f9e418f3b6e54d65a718aaafe3950"},
+    {file = "frozenlist-1.4.1-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:e02a0e11cf6597299b9f3bbd3f93d79217cb90cfd1411aec33848b13f5c656cc"},
+    {file = "frozenlist-1.4.1-cp310-cp310-win32.whl", hash = "sha256:a828c57f00f729620a442881cc60e57cfcec6842ba38e1b19fd3e47ac0ff8dc1"},
+    {file = "frozenlist-1.4.1-cp310-cp310-win_amd64.whl", hash = "sha256:f56e2333dda1fe0f909e7cc59f021eba0d2307bc6f012a1ccf2beca6ba362439"},
+    {file = "frozenlist-1.4.1-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:a0cb6f11204443f27a1628b0e460f37fb30f624be6051d490fa7d7e26d4af3d0"},
+    {file = "frozenlist-1.4.1-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:b46c8ae3a8f1f41a0d2ef350c0b6e65822d80772fe46b653ab6b6274f61d4a49"},
+    {file = "frozenlist-1.4.1-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:fde5bd59ab5357e3853313127f4d3565fc7dad314a74d7b5d43c22c6a5ed2ced"},
+    {file = "frozenlist-1.4.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:722e1124aec435320ae01ee3ac7bec11a5d47f25d0ed6328f2273d287bc3abb0"},
+    {file = "frozenlist-1.4.1-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:2471c201b70d58a0f0c1f91261542a03d9a5e088ed3dc6c160d614c01649c106"},
+    {file = "frozenlist-1.4.1-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:c757a9dd70d72b076d6f68efdbb9bc943665ae954dad2801b874c8c69e185068"},
+    {file = "frozenlist-1.4.1-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:f146e0911cb2f1da549fc58fc7bcd2b836a44b79ef871980d605ec392ff6b0d2"},
+    {file = "frozenlist-1.4.1-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4f9c515e7914626b2a2e1e311794b4c35720a0be87af52b79ff8e1429fc25f19"},
+    {file = "frozenlist-1.4.1-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:c302220494f5c1ebeb0912ea782bcd5e2f8308037b3c7553fad0e48ebad6ad82"},
+    {file = "frozenlist-1.4.1-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:442acde1e068288a4ba7acfe05f5f343e19fac87bfc96d89eb886b0363e977ec"},
+    {file = "frozenlist-1.4.1-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:1b280e6507ea8a4fa0c0a7150b4e526a8d113989e28eaaef946cc77ffd7efc0a"},
+    {file = "frozenlist-1.4.1-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:fe1a06da377e3a1062ae5fe0926e12b84eceb8a50b350ddca72dc85015873f74"},
+    {file = "frozenlist-1.4.1-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:db9e724bebd621d9beca794f2a4ff1d26eed5965b004a97f1f1685a173b869c2"},
+    {file = "frozenlist-1.4.1-cp311-cp311-win32.whl", hash = "sha256:e774d53b1a477a67838a904131c4b0eef6b3d8a651f8b138b04f748fccfefe17"},
+    {file = "frozenlist-1.4.1-cp311-cp311-win_amd64.whl", hash = "sha256:fb3c2db03683b5767dedb5769b8a40ebb47d6f7f45b1b3e3b4b51ec8ad9d9825"},
+    {file = "frozenlist-1.4.1-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:1979bc0aeb89b33b588c51c54ab0161791149f2461ea7c7c946d95d5f93b56ae"},
+    {file = "frozenlist-1.4.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:cc7b01b3754ea68a62bd77ce6020afaffb44a590c2289089289363472d13aedb"},
+    {file = "frozenlist-1.4.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:c9c92be9fd329ac801cc420e08452b70e7aeab94ea4233a4804f0915c14eba9b"},
+    {file = "frozenlist-1.4.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:5c3894db91f5a489fc8fa6a9991820f368f0b3cbdb9cd8849547ccfab3392d86"},
+    {file = "frozenlist-1.4.1-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:ba60bb19387e13597fb059f32cd4d59445d7b18b69a745b8f8e5db0346f33480"},
+    {file = "frozenlist-1.4.1-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:8aefbba5f69d42246543407ed2461db31006b0f76c4e32dfd6f42215a2c41d09"},
+    {file = "frozenlist-1.4.1-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:780d3a35680ced9ce682fbcf4cb9c2bad3136eeff760ab33707b71db84664e3a"},
+    {file = "frozenlist-1.4.1-cp312-cp312-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9acbb16f06fe7f52f441bb6f413ebae6c37baa6ef9edd49cdd567216da8600cd"},
+    {file = "frozenlist-1.4.1-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:23b701e65c7b36e4bf15546a89279bd4d8675faabc287d06bbcfac7d3c33e1e6"},
+    {file = "frozenlist-1.4.1-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:3e0153a805a98f5ada7e09826255ba99fb4f7524bb81bf6b47fb702666484ae1"},
+    {file = "frozenlist-1.4.1-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:dd9b1baec094d91bf36ec729445f7769d0d0cf6b64d04d86e45baf89e2b9059b"},
+    {file = "frozenlist-1.4.1-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:1a4471094e146b6790f61b98616ab8e44f72661879cc63fa1049d13ef711e71e"},
+    {file = "frozenlist-1.4.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:5667ed53d68d91920defdf4035d1cdaa3c3121dc0b113255124bcfada1cfa1b8"},
+    {file = "frozenlist-1.4.1-cp312-cp312-win32.whl", hash = "sha256:beee944ae828747fd7cb216a70f120767fc9f4f00bacae8543c14a6831673f89"},
+    {file = "frozenlist-1.4.1-cp312-cp312-win_amd64.whl", hash = "sha256:64536573d0a2cb6e625cf309984e2d873979709f2cf22839bf2d61790b448ad5"},
+    {file = "frozenlist-1.4.1-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:20b51fa3f588ff2fe658663db52a41a4f7aa6c04f6201449c6c7c476bd255c0d"},
+    {file = "frozenlist-1.4.1-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:410478a0c562d1a5bcc2f7ea448359fcb050ed48b3c6f6f4f18c313a9bdb1826"},
+    {file = "frozenlist-1.4.1-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:c6321c9efe29975232da3bd0af0ad216800a47e93d763ce64f291917a381b8eb"},
+    {file = "frozenlist-1.4.1-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:48f6a4533887e189dae092f1cf981f2e3885175f7a0f33c91fb5b7b682b6bab6"},
+    {file = "frozenlist-1.4.1-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:6eb73fa5426ea69ee0e012fb59cdc76a15b1283d6e32e4f8dc4482ec67d1194d"},
+    {file = "frozenlist-1.4.1-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:fbeb989b5cc29e8daf7f976b421c220f1b8c731cbf22b9130d8815418ea45887"},
+    {file = "frozenlist-1.4.1-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:32453c1de775c889eb4e22f1197fe3bdfe457d16476ea407472b9442e6295f7a"},
+    {file = "frozenlist-1.4.1-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:693945278a31f2086d9bf3df0fe8254bbeaef1fe71e1351c3bd730aa7d31c41b"},
+    {file = "frozenlist-1.4.1-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:1d0ce09d36d53bbbe566fe296965b23b961764c0bcf3ce2fa45f463745c04701"},
+    {file = "frozenlist-1.4.1-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:3a670dc61eb0d0eb7080890c13de3066790f9049b47b0de04007090807c776b0"},
+    {file = "frozenlist-1.4.1-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:dca69045298ce5c11fd539682cff879cc1e664c245d1c64da929813e54241d11"},
+    {file = "frozenlist-1.4.1-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:a06339f38e9ed3a64e4c4e43aec7f59084033647f908e4259d279a52d3757d09"},
+    {file = "frozenlist-1.4.1-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:b7f2f9f912dca3934c1baec2e4585a674ef16fe00218d833856408c48d5beee7"},
+    {file = "frozenlist-1.4.1-cp38-cp38-win32.whl", hash = "sha256:e7004be74cbb7d9f34553a5ce5fb08be14fb33bc86f332fb71cbe5216362a497"},
+    {file = "frozenlist-1.4.1-cp38-cp38-win_amd64.whl", hash = "sha256:5a7d70357e7cee13f470c7883a063aae5fe209a493c57d86eb7f5a6f910fae09"},
+    {file = "frozenlist-1.4.1-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:bfa4a17e17ce9abf47a74ae02f32d014c5e9404b6d9ac7f729e01562bbee601e"},
+    {file = "frozenlist-1.4.1-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:b7e3ed87d4138356775346e6845cccbe66cd9e207f3cd11d2f0b9fd13681359d"},
+    {file = "frozenlist-1.4.1-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:c99169d4ff810155ca50b4da3b075cbde79752443117d89429595c2e8e37fed8"},
+    {file = "frozenlist-1.4.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:edb678da49d9f72c9f6c609fbe41a5dfb9a9282f9e6a2253d5a91e0fc382d7c0"},
+    {file = "frozenlist-1.4.1-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:6db4667b187a6742b33afbbaf05a7bc551ffcf1ced0000a571aedbb4aa42fc7b"},
+    {file = "frozenlist-1.4.1-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:55fdc093b5a3cb41d420884cdaf37a1e74c3c37a31f46e66286d9145d2063bd0"},
+    {file = "frozenlist-1.4.1-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:82e8211d69a4f4bc360ea22cd6555f8e61a1bd211d1d5d39d3d228b48c83a897"},
+    {file = "frozenlist-1.4.1-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:89aa2c2eeb20957be2d950b85974b30a01a762f3308cd02bb15e1ad632e22dc7"},
+    {file = "frozenlist-1.4.1-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:9d3e0c25a2350080e9319724dede4f31f43a6c9779be48021a7f4ebde8b2d742"},
+    {file = "frozenlist-1.4.1-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:7268252af60904bf52c26173cbadc3a071cece75f873705419c8681f24d3edea"},
+    {file = "frozenlist-1.4.1-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:0c250a29735d4f15321007fb02865f0e6b6a41a6b88f1f523ca1596ab5f50bd5"},
+    {file = "frozenlist-1.4.1-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:96ec70beabbd3b10e8bfe52616a13561e58fe84c0101dd031dc78f250d5128b9"},
+    {file = "frozenlist-1.4.1-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:23b2d7679b73fe0e5a4560b672a39f98dfc6f60df63823b0a9970525325b95f6"},
+    {file = "frozenlist-1.4.1-cp39-cp39-win32.whl", hash = "sha256:a7496bfe1da7fb1a4e1cc23bb67c58fab69311cc7d32b5a99c2007b4b2a0e932"},
+    {file = "frozenlist-1.4.1-cp39-cp39-win_amd64.whl", hash = "sha256:e6a20a581f9ce92d389a8c7d7c3dd47c81fd5d6e655c8dddf341e14aa48659d0"},
+    {file = "frozenlist-1.4.1-py3-none-any.whl", hash = "sha256:04ced3e6a46b4cfffe20f9ae482818e34eba9b5fb0ce4056e4cc9b6e212d09b7"},
+    {file = "frozenlist-1.4.1.tar.gz", hash = "sha256:c037a86e8513059a2613aaba4d817bb90b9d9b6b69aace3ce9c877e8c8ed402b"},
+]
+
 [[package]]
 name = "fsspec"
 version = "2024.2.0"
@@ -531,6 +757,105 @@ doc = ["Sphinx (>=1.5.2,<2.0)", "numpydoc (>=0.6.0,<1.0)", "pygame (>=1.9.3,<2.0
 optional = ["matplotlib (>=2.0.0,<3.0)", "opencv-python (>=3.0,<4.0)", "scikit-image (>=0.13.0,<1.0)", "scikit-learn", "scipy (>=0.19.0,<1.5)", "youtube_dl"]
 test = ["coverage (<5.0)", "coveralls (>=1.1,<2.0)", "pytest (>=3.0.0,<4.0)", "pytest-cov (>=2.5.1,<3.0)", "requests (>=2.8.1,<3.0)"]
 
+[[package]]
+name = "multidict"
+version = "6.0.5"
+description = "multidict implementation"
+optional = false
+python-versions = ">=3.7"
+files = [
+    {file = "multidict-6.0.5-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:228b644ae063c10e7f324ab1ab6b548bdf6f8b47f3ec234fef1093bc2735e5f9"},
+    {file = "multidict-6.0.5-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:896ebdcf62683551312c30e20614305f53125750803b614e9e6ce74a96232604"},
+    {file = "multidict-6.0.5-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:411bf8515f3be9813d06004cac41ccf7d1cd46dfe233705933dd163b60e37600"},
+    {file = "multidict-6.0.5-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1d147090048129ce3c453f0292e7697d333db95e52616b3793922945804a433c"},
+    {file = "multidict-6.0.5-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:215ed703caf15f578dca76ee6f6b21b7603791ae090fbf1ef9d865571039ade5"},
+    {file = "multidict-6.0.5-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:7c6390cf87ff6234643428991b7359b5f59cc15155695deb4eda5c777d2b880f"},
+    {file = "multidict-6.0.5-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:21fd81c4ebdb4f214161be351eb5bcf385426bf023041da2fd9e60681f3cebae"},
+    {file = "multidict-6.0.5-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:3cc2ad10255f903656017363cd59436f2111443a76f996584d1077e43ee51182"},
+    {file = "multidict-6.0.5-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:6939c95381e003f54cd4c5516740faba40cf5ad3eeff460c3ad1d3e0ea2549bf"},
+    {file = "multidict-6.0.5-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:220dd781e3f7af2c2c1053da9fa96d9cf3072ca58f057f4c5adaaa1cab8fc442"},
+    {file = "multidict-6.0.5-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:766c8f7511df26d9f11cd3a8be623e59cca73d44643abab3f8c8c07620524e4a"},
+    {file = "multidict-6.0.5-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:fe5d7785250541f7f5019ab9cba2c71169dc7d74d0f45253f8313f436458a4ef"},
+    {file = "multidict-6.0.5-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:c1c1496e73051918fcd4f58ff2e0f2f3066d1c76a0c6aeffd9b45d53243702cc"},
+    {file = "multidict-6.0.5-cp310-cp310-win32.whl", hash = "sha256:7afcdd1fc07befad18ec4523a782cde4e93e0a2bf71239894b8d61ee578c1319"},
+    {file = "multidict-6.0.5-cp310-cp310-win_amd64.whl", hash = "sha256:99f60d34c048c5c2fabc766108c103612344c46e35d4ed9ae0673d33c8fb26e8"},
+    {file = "multidict-6.0.5-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:f285e862d2f153a70586579c15c44656f888806ed0e5b56b64489afe4a2dbfba"},
+    {file = "multidict-6.0.5-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:53689bb4e102200a4fafa9de9c7c3c212ab40a7ab2c8e474491914d2305f187e"},
+    {file = "multidict-6.0.5-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:612d1156111ae11d14afaf3a0669ebf6c170dbb735e510a7438ffe2369a847fd"},
+    {file = "multidict-6.0.5-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:7be7047bd08accdb7487737631d25735c9a04327911de89ff1b26b81745bd4e3"},
+    {file = "multidict-6.0.5-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:de170c7b4fe6859beb8926e84f7d7d6c693dfe8e27372ce3b76f01c46e489fcf"},
+    {file = "multidict-6.0.5-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:04bde7a7b3de05732a4eb39c94574db1ec99abb56162d6c520ad26f83267de29"},
+    {file = "multidict-6.0.5-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:85f67aed7bb647f93e7520633d8f51d3cbc6ab96957c71272b286b2f30dc70ed"},
+    {file = "multidict-6.0.5-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:425bf820055005bfc8aa9a0b99ccb52cc2f4070153e34b701acc98d201693733"},
+    {file = "multidict-6.0.5-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:d3eb1ceec286eba8220c26f3b0096cf189aea7057b6e7b7a2e60ed36b373b77f"},
+    {file = "multidict-6.0.5-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:7901c05ead4b3fb75113fb1dd33eb1253c6d3ee37ce93305acd9d38e0b5f21a4"},
+    {file = "multidict-6.0.5-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:e0e79d91e71b9867c73323a3444724d496c037e578a0e1755ae159ba14f4f3d1"},
+    {file = "multidict-6.0.5-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:29bfeb0dff5cb5fdab2023a7a9947b3b4af63e9c47cae2a10ad58394b517fddc"},
+    {file = "multidict-6.0.5-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:e030047e85cbcedbfc073f71836d62dd5dadfbe7531cae27789ff66bc551bd5e"},
+    {file = "multidict-6.0.5-cp311-cp311-win32.whl", hash = "sha256:2f4848aa3baa109e6ab81fe2006c77ed4d3cd1e0ac2c1fbddb7b1277c168788c"},
+    {file = "multidict-6.0.5-cp311-cp311-win_amd64.whl", hash = "sha256:2faa5ae9376faba05f630d7e5e6be05be22913782b927b19d12b8145968a85ea"},
+    {file = "multidict-6.0.5-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:51d035609b86722963404f711db441cf7134f1889107fb171a970c9701f92e1e"},
+    {file = "multidict-6.0.5-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:cbebcd5bcaf1eaf302617c114aa67569dd3f090dd0ce8ba9e35e9985b41ac35b"},
+    {file = "multidict-6.0.5-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:2ffc42c922dbfddb4a4c3b438eb056828719f07608af27d163191cb3e3aa6cc5"},
+    {file = "multidict-6.0.5-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:ceb3b7e6a0135e092de86110c5a74e46bda4bd4fbfeeb3a3bcec79c0f861e450"},
+    {file = "multidict-6.0.5-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:79660376075cfd4b2c80f295528aa6beb2058fd289f4c9252f986751a4cd0496"},
+    {file = "multidict-6.0.5-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e4428b29611e989719874670fd152b6625500ad6c686d464e99f5aaeeaca175a"},
+    {file = "multidict-6.0.5-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d84a5c3a5f7ce6db1f999fb9438f686bc2e09d38143f2d93d8406ed2dd6b9226"},
+    {file = "multidict-6.0.5-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:76c0de87358b192de7ea9649beb392f107dcad9ad27276324c24c91774ca5271"},
+    {file = "multidict-6.0.5-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:79a6d2ba910adb2cbafc95dad936f8b9386e77c84c35bc0add315b856d7c3abb"},
+    {file = "multidict-6.0.5-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:92d16a3e275e38293623ebf639c471d3e03bb20b8ebb845237e0d3664914caef"},
+    {file = "multidict-6.0.5-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:fb616be3538599e797a2017cccca78e354c767165e8858ab5116813146041a24"},
+    {file = "multidict-6.0.5-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:14c2976aa9038c2629efa2c148022ed5eb4cb939e15ec7aace7ca932f48f9ba6"},
+    {file = "multidict-6.0.5-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:435a0984199d81ca178b9ae2c26ec3d49692d20ee29bc4c11a2a8d4514c67eda"},
+    {file = "multidict-6.0.5-cp312-cp312-win32.whl", hash = "sha256:9fe7b0653ba3d9d65cbe7698cca585bf0f8c83dbbcc710db9c90f478e175f2d5"},
+    {file = "multidict-6.0.5-cp312-cp312-win_amd64.whl", hash = "sha256:01265f5e40f5a17f8241d52656ed27192be03bfa8764d88e8220141d1e4b3556"},
+    {file = "multidict-6.0.5-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:19fe01cea168585ba0f678cad6f58133db2aa14eccaf22f88e4a6dccadfad8b3"},
+    {file = "multidict-6.0.5-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6bf7a982604375a8d49b6cc1b781c1747f243d91b81035a9b43a2126c04766f5"},
+    {file = "multidict-6.0.5-cp37-cp37m-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:107c0cdefe028703fb5dafe640a409cb146d44a6ae201e55b35a4af8e95457dd"},
+    {file = "multidict-6.0.5-cp37-cp37m-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:403c0911cd5d5791605808b942c88a8155c2592e05332d2bf78f18697a5fa15e"},
+    {file = "multidict-6.0.5-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:aeaf541ddbad8311a87dd695ed9642401131ea39ad7bc8cf3ef3967fd093b626"},
+    {file = "multidict-6.0.5-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:e4972624066095e52b569e02b5ca97dbd7a7ddd4294bf4e7247d52635630dd83"},
+    {file = "multidict-6.0.5-cp37-cp37m-musllinux_1_1_aarch64.whl", hash = "sha256:d946b0a9eb8aaa590df1fe082cee553ceab173e6cb5b03239716338629c50c7a"},
+    {file = "multidict-6.0.5-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:b55358304d7a73d7bdf5de62494aaf70bd33015831ffd98bc498b433dfe5b10c"},
+    {file = "multidict-6.0.5-cp37-cp37m-musllinux_1_1_ppc64le.whl", hash = "sha256:a3145cb08d8625b2d3fee1b2d596a8766352979c9bffe5d7833e0503d0f0b5e5"},
+    {file = "multidict-6.0.5-cp37-cp37m-musllinux_1_1_s390x.whl", hash = "sha256:d65f25da8e248202bd47445cec78e0025c0fe7582b23ec69c3b27a640dd7a8e3"},
+    {file = "multidict-6.0.5-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:c9bf56195c6bbd293340ea82eafd0071cb3d450c703d2c93afb89f93b8386ccc"},
+    {file = "multidict-6.0.5-cp37-cp37m-win32.whl", hash = "sha256:69db76c09796b313331bb7048229e3bee7928eb62bab5e071e9f7fcc4879caee"},
+    {file = "multidict-6.0.5-cp37-cp37m-win_amd64.whl", hash = "sha256:fce28b3c8a81b6b36dfac9feb1de115bab619b3c13905b419ec71d03a3fc1423"},
+    {file = "multidict-6.0.5-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:76f067f5121dcecf0d63a67f29080b26c43c71a98b10c701b0677e4a065fbd54"},
+    {file = "multidict-6.0.5-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:b82cc8ace10ab5bd93235dfaab2021c70637005e1ac787031f4d1da63d493c1d"},
+    {file = "multidict-6.0.5-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:5cb241881eefd96b46f89b1a056187ea8e9ba14ab88ba632e68d7a2ecb7aadf7"},
+    {file = "multidict-6.0.5-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e8e94e6912639a02ce173341ff62cc1201232ab86b8a8fcc05572741a5dc7d93"},
+    {file = "multidict-6.0.5-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:09a892e4a9fb47331da06948690ae38eaa2426de97b4ccbfafbdcbe5c8f37ff8"},
+    {file = "multidict-6.0.5-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:55205d03e8a598cfc688c71ca8ea5f66447164efff8869517f175ea632c7cb7b"},
+    {file = "multidict-6.0.5-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:37b15024f864916b4951adb95d3a80c9431299080341ab9544ed148091b53f50"},
+    {file = "multidict-6.0.5-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:f2a1dee728b52b33eebff5072817176c172050d44d67befd681609b4746e1c2e"},
+    {file = "multidict-6.0.5-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:edd08e6f2f1a390bf137080507e44ccc086353c8e98c657e666c017718561b89"},
+    {file = "multidict-6.0.5-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:60d698e8179a42ec85172d12f50b1668254628425a6bd611aba022257cac1386"},
+    {file = "multidict-6.0.5-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:3d25f19500588cbc47dc19081d78131c32637c25804df8414463ec908631e453"},
+    {file = "multidict-6.0.5-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:4cc0ef8b962ac7a5e62b9e826bd0cd5040e7d401bc45a6835910ed699037a461"},
+    {file = "multidict-6.0.5-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:eca2e9d0cc5a889850e9bbd68e98314ada174ff6ccd1129500103df7a94a7a44"},
+    {file = "multidict-6.0.5-cp38-cp38-win32.whl", hash = "sha256:4a6a4f196f08c58c59e0b8ef8ec441d12aee4125a7d4f4fef000ccb22f8d7241"},
+    {file = "multidict-6.0.5-cp38-cp38-win_amd64.whl", hash = "sha256:0275e35209c27a3f7951e1ce7aaf93ce0d163b28948444bec61dd7badc6d3f8c"},
+    {file = "multidict-6.0.5-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:e7be68734bd8c9a513f2b0cfd508802d6609da068f40dc57d4e3494cefc92929"},
+    {file = "multidict-6.0.5-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:1d9ea7a7e779d7a3561aade7d596649fbecfa5c08a7674b11b423783217933f9"},
+    {file = "multidict-6.0.5-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:ea1456df2a27c73ce51120fa2f519f1bea2f4a03a917f4a43c8707cf4cbbae1a"},
+    {file = "multidict-6.0.5-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:cf590b134eb70629e350691ecca88eac3e3b8b3c86992042fb82e3cb1830d5e1"},
+    {file = "multidict-6.0.5-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:5c0631926c4f58e9a5ccce555ad7747d9a9f8b10619621f22f9635f069f6233e"},
+    {file = "multidict-6.0.5-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:dce1c6912ab9ff5f179eaf6efe7365c1f425ed690b03341911bf4939ef2f3046"},
+    {file = "multidict-6.0.5-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c0868d64af83169e4d4152ec612637a543f7a336e4a307b119e98042e852ad9c"},
+    {file = "multidict-6.0.5-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:141b43360bfd3bdd75f15ed811850763555a251e38b2405967f8e25fb43f7d40"},
+    {file = "multidict-6.0.5-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:7df704ca8cf4a073334e0427ae2345323613e4df18cc224f647f251e5e75a527"},
+    {file = "multidict-6.0.5-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:6214c5a5571802c33f80e6c84713b2c79e024995b9c5897f794b43e714daeec9"},
+    {file = "multidict-6.0.5-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:cd6c8fca38178e12c00418de737aef1261576bd1b6e8c6134d3e729a4e858b38"},
+    {file = "multidict-6.0.5-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:e02021f87a5b6932fa6ce916ca004c4d441509d33bbdbeca70d05dff5e9d2479"},
+    {file = "multidict-6.0.5-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:ebd8d160f91a764652d3e51ce0d2956b38efe37c9231cd82cfc0bed2e40b581c"},
+    {file = "multidict-6.0.5-cp39-cp39-win32.whl", hash = "sha256:04da1bb8c8dbadf2a18a452639771951c662c5ad03aefe4884775454be322c9b"},
+    {file = "multidict-6.0.5-cp39-cp39-win_amd64.whl", hash = "sha256:d6f6d4f185481c9669b9447bf9d9cf3b95a0e9df9d169bbc17e363b7d5487755"},
+    {file = "multidict-6.0.5-py3-none-any.whl", hash = "sha256:0d63c74e3d7ab26de115c49bffc92cc77ed23395303d496eae515d4204a625e7"},
+    {file = "multidict-6.0.5.tar.gz", hash = "sha256:f7e301075edaf50500f0b341543c41194d8df3ae5caf4702f2095f3ca73dd8da"},
+]
+
 [[package]]
 name = "nodeenv"
 version = "1.8.0"
@@ -592,13 +917,13 @@ files = [
 
 [[package]]
 name = "openai"
-version = "1.13.3"
+version = "1.14.0"
 description = "The official Python library for the openai API"
 optional = false
 python-versions = ">=3.7.1"
 files = [
-    {file = "openai-1.13.3-py3-none-any.whl", hash = "sha256:5769b62abd02f350a8dd1a3a242d8972c947860654466171d60fb0972ae0a41c"},
-    {file = "openai-1.13.3.tar.gz", hash = "sha256:ff6c6b3bc7327e715e4b3592a923a5a1c7519ff5dd764a83d69f633d49e77a7b"},
+    {file = "openai-1.14.0-py3-none-any.whl", hash = "sha256:5c9fd3a59f5cbdb4020733ddf79a22f6b7a36d561968cb3f3dd255cdd263d9fe"},
+    {file = "openai-1.14.0.tar.gz", hash = "sha256:e287057adf0ec3315abc32ddcc968d095879abd9b68bf51c0402dab13ab5ae9b"},
 ]
 
 [package.dependencies]
@@ -615,90 +940,91 @@ datalib = ["numpy (>=1)", "pandas (>=1.2.3)", "pandas-stubs (>=1.1.0.11)"]
 
 [[package]]
 name = "packaging"
-version = "23.2"
+version = "24.0"
 description = "Core utilities for Python packages"
 optional = false
 python-versions = ">=3.7"
 files = [
-    {file = "packaging-23.2-py3-none-any.whl", hash = "sha256:8c491190033a9af7e1d931d0b5dacc2ef47509b34dd0de67ed209b5203fc88c7"},
-    {file = "packaging-23.2.tar.gz", hash = "sha256:048fb0e9405036518eaaf48a55953c750c11e1a1b68e0dd1a9d62ed0c092cfc5"},
+    {file = "packaging-24.0-py3-none-any.whl", hash = "sha256:2ddfb553fdf02fb784c234c7ba6ccc288296ceabec964ad2eae3777778130bc5"},
+    {file = "packaging-24.0.tar.gz", hash = "sha256:eb82c5e3e56209074766e6885bb04b8c38a0c015d0a30036ebe7ece34c9989e9"},
 ]
 
 [[package]]
 name = "pillow"
-version = "10.2.0"
+version = "10.3.0"
 description = "Python Imaging Library (Fork)"
 optional = false
 python-versions = ">=3.8"
 files = [
-    {file = "pillow-10.2.0-cp310-cp310-macosx_10_10_x86_64.whl", hash = "sha256:7823bdd049099efa16e4246bdf15e5a13dbb18a51b68fa06d6c1d4d8b99a796e"},
-    {file = "pillow-10.2.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:83b2021f2ade7d1ed556bc50a399127d7fb245e725aa0113ebd05cfe88aaf588"},
-    {file = "pillow-10.2.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6fad5ff2f13d69b7e74ce5b4ecd12cc0ec530fcee76356cac6742785ff71c452"},
-    {file = "pillow-10.2.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:da2b52b37dad6d9ec64e653637a096905b258d2fc2b984c41ae7d08b938a67e4"},
-    {file = "pillow-10.2.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:47c0995fc4e7f79b5cfcab1fc437ff2890b770440f7696a3ba065ee0fd496563"},
-    {file = "pillow-10.2.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:322bdf3c9b556e9ffb18f93462e5f749d3444ce081290352c6070d014c93feb2"},
-    {file = "pillow-10.2.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:51f1a1bffc50e2e9492e87d8e09a17c5eea8409cda8d3f277eb6edc82813c17c"},
-    {file = "pillow-10.2.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:69ffdd6120a4737710a9eee73e1d2e37db89b620f702754b8f6e62594471dee0"},
-    {file = "pillow-10.2.0-cp310-cp310-win32.whl", hash = "sha256:c6dafac9e0f2b3c78df97e79af707cdc5ef8e88208d686a4847bab8266870023"},
-    {file = "pillow-10.2.0-cp310-cp310-win_amd64.whl", hash = "sha256:aebb6044806f2e16ecc07b2a2637ee1ef67a11840a66752751714a0d924adf72"},
-    {file = "pillow-10.2.0-cp310-cp310-win_arm64.whl", hash = "sha256:7049e301399273a0136ff39b84c3678e314f2158f50f517bc50285fb5ec847ad"},
-    {file = "pillow-10.2.0-cp311-cp311-macosx_10_10_x86_64.whl", hash = "sha256:35bb52c37f256f662abdfa49d2dfa6ce5d93281d323a9af377a120e89a9eafb5"},
-    {file = "pillow-10.2.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:9c23f307202661071d94b5e384e1e1dc7dfb972a28a2310e4ee16103e66ddb67"},
-    {file = "pillow-10.2.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:773efe0603db30c281521a7c0214cad7836c03b8ccff897beae9b47c0b657d61"},
-    {file = "pillow-10.2.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:11fa2e5984b949b0dd6d7a94d967743d87c577ff0b83392f17cb3990d0d2fd6e"},
-    {file = "pillow-10.2.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:716d30ed977be8b37d3ef185fecb9e5a1d62d110dfbdcd1e2a122ab46fddb03f"},
-    {file = "pillow-10.2.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:a086c2af425c5f62a65e12fbf385f7c9fcb8f107d0849dba5839461a129cf311"},
-    {file = "pillow-10.2.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:c8de2789052ed501dd829e9cae8d3dcce7acb4777ea4a479c14521c942d395b1"},
-    {file = "pillow-10.2.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:609448742444d9290fd687940ac0b57fb35e6fd92bdb65386e08e99af60bf757"},
-    {file = "pillow-10.2.0-cp311-cp311-win32.whl", hash = "sha256:823ef7a27cf86df6597fa0671066c1b596f69eba53efa3d1e1cb8b30f3533068"},
-    {file = "pillow-10.2.0-cp311-cp311-win_amd64.whl", hash = "sha256:1da3b2703afd040cf65ec97efea81cfba59cdbed9c11d8efc5ab09df9509fc56"},
-    {file = "pillow-10.2.0-cp311-cp311-win_arm64.whl", hash = "sha256:edca80cbfb2b68d7b56930b84a0e45ae1694aeba0541f798e908a49d66b837f1"},
-    {file = "pillow-10.2.0-cp312-cp312-macosx_10_10_x86_64.whl", hash = "sha256:1b5e1b74d1bd1b78bc3477528919414874748dd363e6272efd5abf7654e68bef"},
-    {file = "pillow-10.2.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:0eae2073305f451d8ecacb5474997c08569fb4eb4ac231ffa4ad7d342fdc25ac"},
-    {file = "pillow-10.2.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b7c2286c23cd350b80d2fc9d424fc797575fb16f854b831d16fd47ceec078f2c"},
-    {file = "pillow-10.2.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:1e23412b5c41e58cec602f1135c57dfcf15482013ce6e5f093a86db69646a5aa"},
-    {file = "pillow-10.2.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:52a50aa3fb3acb9cf7213573ef55d31d6eca37f5709c69e6858fe3bc04a5c2a2"},
-    {file = "pillow-10.2.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:127cee571038f252a552760076407f9cff79761c3d436a12af6000cd182a9d04"},
-    {file = "pillow-10.2.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:8d12251f02d69d8310b046e82572ed486685c38f02176bd08baf216746eb947f"},
-    {file = "pillow-10.2.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:54f1852cd531aa981bc0965b7d609f5f6cc8ce8c41b1139f6ed6b3c54ab82bfb"},
-    {file = "pillow-10.2.0-cp312-cp312-win32.whl", hash = "sha256:257d8788df5ca62c980314053197f4d46eefedf4e6175bc9412f14412ec4ea2f"},
-    {file = "pillow-10.2.0-cp312-cp312-win_amd64.whl", hash = "sha256:154e939c5f0053a383de4fd3d3da48d9427a7e985f58af8e94d0b3c9fcfcf4f9"},
-    {file = "pillow-10.2.0-cp312-cp312-win_arm64.whl", hash = "sha256:f379abd2f1e3dddb2b61bc67977a6b5a0a3f7485538bcc6f39ec76163891ee48"},
-    {file = "pillow-10.2.0-cp38-cp38-macosx_10_10_x86_64.whl", hash = "sha256:8373c6c251f7ef8bda6675dd6d2b3a0fcc31edf1201266b5cf608b62a37407f9"},
-    {file = "pillow-10.2.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:870ea1ada0899fd0b79643990809323b389d4d1d46c192f97342eeb6ee0b8483"},
-    {file = "pillow-10.2.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b4b6b1e20608493548b1f32bce8cca185bf0480983890403d3b8753e44077129"},
-    {file = "pillow-10.2.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3031709084b6e7852d00479fd1d310b07d0ba82765f973b543c8af5061cf990e"},
-    {file = "pillow-10.2.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:3ff074fc97dd4e80543a3e91f69d58889baf2002b6be64347ea8cf5533188213"},
-    {file = "pillow-10.2.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:cb4c38abeef13c61d6916f264d4845fab99d7b711be96c326b84df9e3e0ff62d"},
-    {file = "pillow-10.2.0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:b1b3020d90c2d8e1dae29cf3ce54f8094f7938460fb5ce8bc5c01450b01fbaf6"},
-    {file = "pillow-10.2.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:170aeb00224ab3dc54230c797f8404507240dd868cf52066f66a41b33169bdbe"},
-    {file = "pillow-10.2.0-cp38-cp38-win32.whl", hash = "sha256:c4225f5220f46b2fde568c74fca27ae9771536c2e29d7c04f4fb62c83275ac4e"},
-    {file = "pillow-10.2.0-cp38-cp38-win_amd64.whl", hash = "sha256:0689b5a8c5288bc0504d9fcee48f61a6a586b9b98514d7d29b840143d6734f39"},
-    {file = "pillow-10.2.0-cp39-cp39-macosx_10_10_x86_64.whl", hash = "sha256:b792a349405fbc0163190fde0dc7b3fef3c9268292586cf5645598b48e63dc67"},
-    {file = "pillow-10.2.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:c570f24be1e468e3f0ce7ef56a89a60f0e05b30a3669a459e419c6eac2c35364"},
-    {file = "pillow-10.2.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d8ecd059fdaf60c1963c58ceb8997b32e9dc1b911f5da5307aab614f1ce5c2fb"},
-    {file = "pillow-10.2.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c365fd1703040de1ec284b176d6af5abe21b427cb3a5ff68e0759e1e313a5e7e"},
-    {file = "pillow-10.2.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:70c61d4c475835a19b3a5aa42492409878bbca7438554a1f89d20d58a7c75c01"},
-    {file = "pillow-10.2.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:b6f491cdf80ae540738859d9766783e3b3c8e5bd37f5dfa0b76abdecc5081f13"},
-    {file = "pillow-10.2.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:9d189550615b4948f45252d7f005e53c2040cea1af5b60d6f79491a6e147eef7"},
-    {file = "pillow-10.2.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:49d9ba1ed0ef3e061088cd1e7538a0759aab559e2e0a80a36f9fd9d8c0c21591"},
-    {file = "pillow-10.2.0-cp39-cp39-win32.whl", hash = "sha256:babf5acfede515f176833ed6028754cbcd0d206f7f614ea3447d67c33be12516"},
-    {file = "pillow-10.2.0-cp39-cp39-win_amd64.whl", hash = "sha256:0304004f8067386b477d20a518b50f3fa658a28d44e4116970abfcd94fac34a8"},
-    {file = "pillow-10.2.0-cp39-cp39-win_arm64.whl", hash = "sha256:0fb3e7fc88a14eacd303e90481ad983fd5b69c761e9e6ef94c983f91025da869"},
-    {file = "pillow-10.2.0-pp310-pypy310_pp73-macosx_10_10_x86_64.whl", hash = "sha256:322209c642aabdd6207517e9739c704dc9f9db943015535783239022002f054a"},
-    {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3eedd52442c0a5ff4f887fab0c1c0bb164d8635b32c894bc1faf4c618dd89df2"},
-    {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:cb28c753fd5eb3dd859b4ee95de66cc62af91bcff5db5f2571d32a520baf1f04"},
-    {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:33870dc4653c5017bf4c8873e5488d8f8d5f8935e2f1fb9a2208c47cdd66efd2"},
-    {file = "pillow-10.2.0-pp310-pypy310_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:3c31822339516fb3c82d03f30e22b1d038da87ef27b6a78c9549888f8ceda39a"},
-    {file = "pillow-10.2.0-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:a2b56ba36e05f973d450582fb015594aaa78834fefe8dfb8fcd79b93e64ba4c6"},
-    {file = "pillow-10.2.0-pp38-pypy38_pp73-win_amd64.whl", hash = "sha256:d8e6aeb9201e655354b3ad049cb77d19813ad4ece0df1249d3c793de3774f8c7"},
-    {file = "pillow-10.2.0-pp39-pypy39_pp73-macosx_10_10_x86_64.whl", hash = "sha256:2247178effb34a77c11c0e8ac355c7a741ceca0a732b27bf11e747bbc950722f"},
-    {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:15587643b9e5eb26c48e49a7b33659790d28f190fc514a322d55da2fb5c2950e"},
-    {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:753cd8f2086b2b80180d9b3010dd4ed147efc167c90d3bf593fe2af21265e5a5"},
-    {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:7c8f97e8e7a9009bcacbe3766a36175056c12f9a44e6e6f2d5caad06dcfbf03b"},
-    {file = "pillow-10.2.0-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:d1b35bcd6c5543b9cb547dee3150c93008f8dd0f1fef78fc0cd2b141c5baf58a"},
-    {file = "pillow-10.2.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:fe4c15f6c9285dc54ce6553a3ce908ed37c8f3825b5a51a15c91442bb955b868"},
-    {file = "pillow-10.2.0.tar.gz", hash = "sha256:e87f0b2c78157e12d7686b27d63c070fd65d994e8ddae6f328e0dcf4a0cd007e"},
+    {file = "pillow-10.3.0-cp310-cp310-macosx_10_10_x86_64.whl", hash = "sha256:90b9e29824800e90c84e4022dd5cc16eb2d9605ee13f05d47641eb183cd73d45"},
+    {file = "pillow-10.3.0-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:a2c405445c79c3f5a124573a051062300936b0281fee57637e706453e452746c"},
+    {file = "pillow-10.3.0-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:78618cdbccaa74d3f88d0ad6cb8ac3007f1a6fa5c6f19af64b55ca170bfa1edf"},
+    {file = "pillow-10.3.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:261ddb7ca91fcf71757979534fb4c128448b5b4c55cb6152d280312062f69599"},
+    {file = "pillow-10.3.0-cp310-cp310-manylinux_2_28_aarch64.whl", hash = "sha256:ce49c67f4ea0609933d01c0731b34b8695a7a748d6c8d186f95e7d085d2fe475"},
+    {file = "pillow-10.3.0-cp310-cp310-manylinux_2_28_x86_64.whl", hash = "sha256:b14f16f94cbc61215115b9b1236f9c18403c15dd3c52cf629072afa9d54c1cbf"},
+    {file = "pillow-10.3.0-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:d33891be6df59d93df4d846640f0e46f1a807339f09e79a8040bc887bdcd7ed3"},
+    {file = "pillow-10.3.0-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:b50811d664d392f02f7761621303eba9d1b056fb1868c8cdf4231279645c25f5"},
+    {file = "pillow-10.3.0-cp310-cp310-win32.whl", hash = "sha256:ca2870d5d10d8726a27396d3ca4cf7976cec0f3cb706debe88e3a5bd4610f7d2"},
+    {file = "pillow-10.3.0-cp310-cp310-win_amd64.whl", hash = "sha256:f0d0591a0aeaefdaf9a5e545e7485f89910c977087e7de2b6c388aec32011e9f"},
+    {file = "pillow-10.3.0-cp310-cp310-win_arm64.whl", hash = "sha256:ccce24b7ad89adb5a1e34a6ba96ac2530046763912806ad4c247356a8f33a67b"},
+    {file = "pillow-10.3.0-cp311-cp311-macosx_10_10_x86_64.whl", hash = "sha256:5f77cf66e96ae734717d341c145c5949c63180842a545c47a0ce7ae52ca83795"},
+    {file = "pillow-10.3.0-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:e4b878386c4bf293578b48fc570b84ecfe477d3b77ba39a6e87150af77f40c57"},
+    {file = "pillow-10.3.0-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:fdcbb4068117dfd9ce0138d068ac512843c52295ed996ae6dd1faf537b6dbc27"},
+    {file = "pillow-10.3.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9797a6c8fe16f25749b371c02e2ade0efb51155e767a971c61734b1bf6293994"},
+    {file = "pillow-10.3.0-cp311-cp311-manylinux_2_28_aarch64.whl", hash = "sha256:9e91179a242bbc99be65e139e30690e081fe6cb91a8e77faf4c409653de39451"},
+    {file = "pillow-10.3.0-cp311-cp311-manylinux_2_28_x86_64.whl", hash = "sha256:1b87bd9d81d179bd8ab871603bd80d8645729939f90b71e62914e816a76fc6bd"},
+    {file = "pillow-10.3.0-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:81d09caa7b27ef4e61cb7d8fbf1714f5aec1c6b6c5270ee53504981e6e9121ad"},
+    {file = "pillow-10.3.0-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:048ad577748b9fa4a99a0548c64f2cb8d672d5bf2e643a739ac8faff1164238c"},
+    {file = "pillow-10.3.0-cp311-cp311-win32.whl", hash = "sha256:7161ec49ef0800947dc5570f86568a7bb36fa97dd09e9827dc02b718c5643f09"},
+    {file = "pillow-10.3.0-cp311-cp311-win_amd64.whl", hash = "sha256:8eb0908e954d093b02a543dc963984d6e99ad2b5e36503d8a0aaf040505f747d"},
+    {file = "pillow-10.3.0-cp311-cp311-win_arm64.whl", hash = "sha256:4e6f7d1c414191c1199f8996d3f2282b9ebea0945693fb67392c75a3a320941f"},
+    {file = "pillow-10.3.0-cp312-cp312-macosx_10_10_x86_64.whl", hash = "sha256:e46f38133e5a060d46bd630faa4d9fa0202377495df1f068a8299fd78c84de84"},
+    {file = "pillow-10.3.0-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:50b8eae8f7334ec826d6eeffaeeb00e36b5e24aa0b9df322c247539714c6df19"},
+    {file = "pillow-10.3.0-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:9d3bea1c75f8c53ee4d505c3e67d8c158ad4df0d83170605b50b64025917f338"},
+    {file = "pillow-10.3.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:19aeb96d43902f0a783946a0a87dbdad5c84c936025b8419da0a0cd7724356b1"},
+    {file = "pillow-10.3.0-cp312-cp312-manylinux_2_28_aarch64.whl", hash = "sha256:74d28c17412d9caa1066f7a31df8403ec23d5268ba46cd0ad2c50fb82ae40462"},
+    {file = "pillow-10.3.0-cp312-cp312-manylinux_2_28_x86_64.whl", hash = "sha256:ff61bfd9253c3915e6d41c651d5f962da23eda633cf02262990094a18a55371a"},
+    {file = "pillow-10.3.0-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:d886f5d353333b4771d21267c7ecc75b710f1a73d72d03ca06df49b09015a9ef"},
+    {file = "pillow-10.3.0-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:4b5ec25d8b17217d635f8935dbc1b9aa5907962fae29dff220f2659487891cd3"},
+    {file = "pillow-10.3.0-cp312-cp312-win32.whl", hash = "sha256:51243f1ed5161b9945011a7360e997729776f6e5d7005ba0c6879267d4c5139d"},
+    {file = "pillow-10.3.0-cp312-cp312-win_amd64.whl", hash = "sha256:412444afb8c4c7a6cc11a47dade32982439925537e483be7c0ae0cf96c4f6a0b"},
+    {file = "pillow-10.3.0-cp312-cp312-win_arm64.whl", hash = "sha256:798232c92e7665fe82ac085f9d8e8ca98826f8e27859d9a96b41d519ecd2e49a"},
+    {file = "pillow-10.3.0-cp38-cp38-macosx_10_10_x86_64.whl", hash = "sha256:4eaa22f0d22b1a7e93ff0a596d57fdede2e550aecffb5a1ef1106aaece48e96b"},
+    {file = "pillow-10.3.0-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:cd5e14fbf22a87321b24c88669aad3a51ec052eb145315b3da3b7e3cc105b9a2"},
+    {file = "pillow-10.3.0-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1530e8f3a4b965eb6a7785cf17a426c779333eb62c9a7d1bbcf3ffd5bf77a4aa"},
+    {file = "pillow-10.3.0-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5d512aafa1d32efa014fa041d38868fda85028e3f930a96f85d49c7d8ddc0383"},
+    {file = "pillow-10.3.0-cp38-cp38-manylinux_2_28_aarch64.whl", hash = "sha256:339894035d0ede518b16073bdc2feef4c991ee991a29774b33e515f1d308e08d"},
+    {file = "pillow-10.3.0-cp38-cp38-manylinux_2_28_x86_64.whl", hash = "sha256:aa7e402ce11f0885305bfb6afb3434b3cd8f53b563ac065452d9d5654c7b86fd"},
+    {file = "pillow-10.3.0-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:0ea2a783a2bdf2a561808fe4a7a12e9aa3799b701ba305de596bc48b8bdfce9d"},
+    {file = "pillow-10.3.0-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:c78e1b00a87ce43bb37642c0812315b411e856a905d58d597750eb79802aaaa3"},
+    {file = "pillow-10.3.0-cp38-cp38-win32.whl", hash = "sha256:72d622d262e463dfb7595202d229f5f3ab4b852289a1cd09650362db23b9eb0b"},
+    {file = "pillow-10.3.0-cp38-cp38-win_amd64.whl", hash = "sha256:2034f6759a722da3a3dbd91a81148cf884e91d1b747992ca288ab88c1de15999"},
+    {file = "pillow-10.3.0-cp39-cp39-macosx_10_10_x86_64.whl", hash = "sha256:2ed854e716a89b1afcedea551cd85f2eb2a807613752ab997b9974aaa0d56936"},
+    {file = "pillow-10.3.0-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:dc1a390a82755a8c26c9964d457d4c9cbec5405896cba94cf51f36ea0d855002"},
+    {file = "pillow-10.3.0-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4203efca580f0dd6f882ca211f923168548f7ba334c189e9eab1178ab840bf60"},
+    {file = "pillow-10.3.0-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3102045a10945173d38336f6e71a8dc71bcaeed55c3123ad4af82c52807b9375"},
+    {file = "pillow-10.3.0-cp39-cp39-manylinux_2_28_aarch64.whl", hash = "sha256:6fb1b30043271ec92dc65f6d9f0b7a830c210b8a96423074b15c7bc999975f57"},
+    {file = "pillow-10.3.0-cp39-cp39-manylinux_2_28_x86_64.whl", hash = "sha256:1dfc94946bc60ea375cc39cff0b8da6c7e5f8fcdc1d946beb8da5c216156ddd8"},
+    {file = "pillow-10.3.0-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:b09b86b27a064c9624d0a6c54da01c1beaf5b6cadfa609cf63789b1d08a797b9"},
+    {file = "pillow-10.3.0-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:d3b2348a78bc939b4fed6552abfd2e7988e0f81443ef3911a4b8498ca084f6eb"},
+    {file = "pillow-10.3.0-cp39-cp39-win32.whl", hash = "sha256:45ebc7b45406febf07fef35d856f0293a92e7417ae7933207e90bf9090b70572"},
+    {file = "pillow-10.3.0-cp39-cp39-win_amd64.whl", hash = "sha256:0ba26351b137ca4e0db0342d5d00d2e355eb29372c05afd544ebf47c0956ffeb"},
+    {file = "pillow-10.3.0-cp39-cp39-win_arm64.whl", hash = "sha256:50fd3f6b26e3441ae07b7c979309638b72abc1a25da31a81a7fbd9495713ef4f"},
+    {file = "pillow-10.3.0-pp310-pypy310_pp73-macosx_10_10_x86_64.whl", hash = "sha256:6b02471b72526ab8a18c39cb7967b72d194ec53c1fd0a70b050565a0f366d355"},
+    {file = "pillow-10.3.0-pp310-pypy310_pp73-macosx_11_0_arm64.whl", hash = "sha256:8ab74c06ffdab957d7670c2a5a6e1a70181cd10b727cd788c4dd9005b6a8acd9"},
+    {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:048eeade4c33fdf7e08da40ef402e748df113fd0b4584e32c4af74fe78baaeb2"},
+    {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:9e2ec1e921fd07c7cda7962bad283acc2f2a9ccc1b971ee4b216b75fad6f0463"},
+    {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:4c8e73e99da7db1b4cad7f8d682cf6abad7844da39834c288fbfa394a47bbced"},
+    {file = "pillow-10.3.0-pp310-pypy310_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:16563993329b79513f59142a6b02055e10514c1a8e86dca8b48a893e33cf91e3"},
+    {file = "pillow-10.3.0-pp310-pypy310_pp73-win_amd64.whl", hash = "sha256:dd78700f5788ae180b5ee8902c6aea5a5726bac7c364b202b4b3e3ba2d293170"},
+    {file = "pillow-10.3.0-pp39-pypy39_pp73-macosx_10_10_x86_64.whl", hash = "sha256:aff76a55a8aa8364d25400a210a65ff59d0168e0b4285ba6bf2bd83cf675ba32"},
+    {file = "pillow-10.3.0-pp39-pypy39_pp73-macosx_11_0_arm64.whl", hash = "sha256:b7bc2176354defba3edc2b9a777744462da2f8e921fbaf61e52acb95bafa9828"},
+    {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:793b4e24db2e8742ca6423d3fde8396db336698c55cd34b660663ee9e45ed37f"},
+    {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d93480005693d247f8346bc8ee28c72a2191bdf1f6b5db469c096c0c867ac015"},
+    {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_28_aarch64.whl", hash = "sha256:c83341b89884e2b2e55886e8fbbf37c3fa5efd6c8907124aeb72f285ae5696e5"},
+    {file = "pillow-10.3.0-pp39-pypy39_pp73-manylinux_2_28_x86_64.whl", hash = "sha256:1a1d1915db1a4fdb2754b9de292642a39a7fb28f1736699527bb649484fb966a"},
+    {file = "pillow-10.3.0-pp39-pypy39_pp73-win_amd64.whl", hash = "sha256:a0eaa93d054751ee9964afa21c06247779b90440ca41d184aeb5d410f20ff591"},
+    {file = "pillow-10.3.0.tar.gz", hash = "sha256:9d2455fbf44c914840c793e89aa82d0e1763a14253a000743719ae5946814b2d"},
 ]
 
 [package.extras]
@@ -825,13 +1151,13 @@ email = ["email-validator (>=1.0.3)"]
 
 [[package]]
 name = "pyright"
-version = "1.1.352"
+version = "1.1.354"
 description = "Command line wrapper for pyright"
 optional = false
 python-versions = ">=3.7"
 files = [
-    {file = "pyright-1.1.352-py3-none-any.whl", hash = "sha256:0040cf173c6a60704e553bfd129dfe54de59cc76d0b2b80f77cfab4f50701d64"},
-    {file = "pyright-1.1.352.tar.gz", hash = "sha256:a621c0dfbcf1291b3610641a07380fefaa1d0e182890a1b2a7f13b446e8109a9"},
+    {file = "pyright-1.1.354-py3-none-any.whl", hash = "sha256:f28d61ae8ae035fc52ded1070e8d9e786051a26a4127bbd7a4ba0399b81b37b5"},
+    {file = "pyright-1.1.354.tar.gz", hash = "sha256:b1070dc774ff2e79eb0523fe87f4ba9a90550de7e4b030a2bc9e031864029a1f"},
 ]
 
 [package.dependencies]
@@ -902,6 +1228,7 @@ files = [
     {file = "PyYAML-6.0.1-cp311-cp311-win_amd64.whl", hash = "sha256:bf07ee2fef7014951eeb99f56f39c9bb4af143d8aa3c21b1677805985307da34"},
     {file = "PyYAML-6.0.1-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:855fb52b0dc35af121542a76b9a84f8d1cd886ea97c84703eaa6d88e37a2ad28"},
     {file = "PyYAML-6.0.1-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:40df9b996c2b73138957fe23a16a4f0ba614f4c0efce1e9406a184b6d07fa3a9"},
+    {file = "PyYAML-6.0.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:a08c6f0fe150303c1c6b71ebcd7213c2858041a7e01975da3a99aed1e7a378ef"},
     {file = "PyYAML-6.0.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:6c22bec3fbe2524cde73d7ada88f6566758a8f7227bfbf93a408a9d86bcc12a0"},
     {file = "PyYAML-6.0.1-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:8d4e9c88387b0f5c7d5f281e55304de64cf7f9c0021a3525bd3b1c542da3b0e4"},
     {file = "PyYAML-6.0.1-cp312-cp312-win32.whl", hash = "sha256:d483d2cdf104e7c9fa60c544d92981f12ad66a457afae824d146093b8c294c54"},
@@ -959,18 +1286,18 @@ use-chardet-on-py3 = ["chardet (>=3.0.2,<6)"]
 
 [[package]]
 name = "setuptools"
-version = "69.1.1"
+version = "69.2.0"
 description = "Easily download, build, install, upgrade, and uninstall Python packages"
 optional = false
 python-versions = ">=3.8"
 files = [
-    {file = "setuptools-69.1.1-py3-none-any.whl", hash = "sha256:02fa291a0471b3a18b2b2481ed902af520c69e8ae0919c13da936542754b4c56"},
-    {file = "setuptools-69.1.1.tar.gz", hash = "sha256:5c0806c7d9af348e6dd3777b4f4dbb42c7ad85b190104837488eab9a7c945cf8"},
+    {file = "setuptools-69.2.0-py3-none-any.whl", hash = "sha256:c21c49fb1042386df081cb5d86759792ab89efca84cf114889191cd09aacc80c"},
+    {file = "setuptools-69.2.0.tar.gz", hash = "sha256:0ff4183f8f42cd8fa3acea16c45205521a4ef28f73c6391d8a25e92893134f2e"},
 ]
 
 [package.extras]
 docs = ["furo", "jaraco.packaging (>=9.3)", "jaraco.tidelift (>=1.4)", "pygments-github-lexers (==0.0.5)", "rst.linker (>=1.9)", "sphinx (<7.2.5)", "sphinx (>=3.5)", "sphinx-favicon", "sphinx-inline-tabs", "sphinx-lint", "sphinx-notfound-page (>=1,<2)", "sphinx-reredirects", "sphinxcontrib-towncrier"]
-testing = ["build[virtualenv]", "filelock (>=3.4.0)", "flake8-2020", "ini2toml[lite] (>=0.9)", "jaraco.develop (>=7.21)", "jaraco.envs (>=2.2)", "jaraco.path (>=3.2.0)", "packaging (>=23.2)", "pip (>=19.1)", "pytest (>=6)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-home (>=0.5)", "pytest-mypy (>=0.9.1)", "pytest-perf", "pytest-ruff (>=0.2.1)", "pytest-timeout", "pytest-xdist", "tomli-w (>=1.0.0)", "virtualenv (>=13.0.0)", "wheel"]
+testing = ["build[virtualenv]", "filelock (>=3.4.0)", "importlib-metadata", "ini2toml[lite] (>=0.9)", "jaraco.develop (>=7.21)", "jaraco.envs (>=2.2)", "jaraco.path (>=3.2.0)", "mypy (==1.9)", "packaging (>=23.2)", "pip (>=19.1)", "pytest (>=6)", "pytest-checkdocs (>=2.4)", "pytest-cov", "pytest-enabler (>=2.2)", "pytest-home (>=0.5)", "pytest-mypy (>=0.9.1)", "pytest-perf", "pytest-ruff (>=0.2.1)", "pytest-timeout", "pytest-xdist (>=3)", "tomli", "tomli-w (>=1.0.0)", "virtualenv (>=13.0.0)", "wheel"]
 testing-integration = ["build[virtualenv] (>=1.0.3)", "filelock (>=3.4.0)", "jaraco.envs (>=2.2)", "jaraco.path (>=3.2.0)", "packaging (>=23.2)", "pytest", "pytest-enabler", "pytest-xdist", "tomli", "virtualenv (>=13.0.0)", "wheel"]
 
 [[package]]
@@ -1170,6 +1497,17 @@ notebook = ["ipywidgets (>=6)"]
 slack = ["slack-sdk"]
 telegram = ["requests"]
 
+[[package]]
+name = "types-pillow"
+version = "10.2.0.20240520"
+description = "Typing stubs for Pillow"
+optional = false
+python-versions = ">=3.8"
+files = [
+    {file = "types-Pillow-10.2.0.20240520.tar.gz", hash = "sha256:130b979195465fa1e1676d8e81c9c7c30319e8e95b12fae945e8f0d525213107"},
+    {file = "types_Pillow-10.2.0.20240520-py3-none-any.whl", hash = "sha256:33c36494b380e2a269bb742181bea5d9b00820367822dbd3760f07210a1da23d"},
+]
+
 [[package]]
 name = "typing-extensions"
 version = "4.10.0"
@@ -1318,7 +1656,110 @@ files = [
     {file = "websockets-12.0.tar.gz", hash = "sha256:81df9cbcbb6c260de1e007e58c011bfebe2dafc8435107b0537f393dd38c8b1b"},
 ]
 
+[[package]]
+name = "yarl"
+version = "1.9.4"
+description = "Yet another URL library"
+optional = false
+python-versions = ">=3.7"
+files = [
+    {file = "yarl-1.9.4-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:a8c1df72eb746f4136fe9a2e72b0c9dc1da1cbd23b5372f94b5820ff8ae30e0e"},
+    {file = "yarl-1.9.4-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:a3a6ed1d525bfb91b3fc9b690c5a21bb52de28c018530ad85093cc488bee2dd2"},
+    {file = "yarl-1.9.4-cp310-cp310-macosx_11_0_arm64.whl", hash = "sha256:c38c9ddb6103ceae4e4498f9c08fac9b590c5c71b0370f98714768e22ac6fa66"},
+    {file = "yarl-1.9.4-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d9e09c9d74f4566e905a0b8fa668c58109f7624db96a2171f21747abc7524234"},
+    {file = "yarl-1.9.4-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:b8477c1ee4bd47c57d49621a062121c3023609f7a13b8a46953eb6c9716ca392"},
+    {file = "yarl-1.9.4-cp310-cp310-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:d5ff2c858f5f6a42c2a8e751100f237c5e869cbde669a724f2062d4c4ef93551"},
+    {file = "yarl-1.9.4-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:357495293086c5b6d34ca9616a43d329317feab7917518bc97a08f9e55648455"},
+    {file = "yarl-1.9.4-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:54525ae423d7b7a8ee81ba189f131054defdb122cde31ff17477951464c1691c"},
+    {file = "yarl-1.9.4-cp310-cp310-musllinux_1_1_aarch64.whl", hash = "sha256:801e9264d19643548651b9db361ce3287176671fb0117f96b5ac0ee1c3530d53"},
+    {file = "yarl-1.9.4-cp310-cp310-musllinux_1_1_i686.whl", hash = "sha256:e516dc8baf7b380e6c1c26792610230f37147bb754d6426462ab115a02944385"},
+    {file = "yarl-1.9.4-cp310-cp310-musllinux_1_1_ppc64le.whl", hash = "sha256:7d5aaac37d19b2904bb9dfe12cdb08c8443e7ba7d2852894ad448d4b8f442863"},
+    {file = "yarl-1.9.4-cp310-cp310-musllinux_1_1_s390x.whl", hash = "sha256:54beabb809ffcacbd9d28ac57b0db46e42a6e341a030293fb3185c409e626b8b"},
+    {file = "yarl-1.9.4-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:bac8d525a8dbc2a1507ec731d2867025d11ceadcb4dd421423a5d42c56818541"},
+    {file = "yarl-1.9.4-cp310-cp310-win32.whl", hash = "sha256:7855426dfbddac81896b6e533ebefc0af2f132d4a47340cee6d22cac7190022d"},
+    {file = "yarl-1.9.4-cp310-cp310-win_amd64.whl", hash = "sha256:848cd2a1df56ddbffeb375535fb62c9d1645dde33ca4d51341378b3f5954429b"},
+    {file = "yarl-1.9.4-cp311-cp311-macosx_10_9_universal2.whl", hash = "sha256:35a2b9396879ce32754bd457d31a51ff0a9d426fd9e0e3c33394bf4b9036b099"},
+    {file = "yarl-1.9.4-cp311-cp311-macosx_10_9_x86_64.whl", hash = "sha256:4c7d56b293cc071e82532f70adcbd8b61909eec973ae9d2d1f9b233f3d943f2c"},
+    {file = "yarl-1.9.4-cp311-cp311-macosx_11_0_arm64.whl", hash = "sha256:d8a1c6c0be645c745a081c192e747c5de06e944a0d21245f4cf7c05e457c36e0"},
+    {file = "yarl-1.9.4-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4b3c1ffe10069f655ea2d731808e76e0f452fc6c749bea04781daf18e6039525"},
+    {file = "yarl-1.9.4-cp311-cp311-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:549d19c84c55d11687ddbd47eeb348a89df9cb30e1993f1b128f4685cd0ebbf8"},
+    {file = "yarl-1.9.4-cp311-cp311-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:a7409f968456111140c1c95301cadf071bd30a81cbd7ab829169fb9e3d72eae9"},
+    {file = "yarl-1.9.4-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:e23a6d84d9d1738dbc6e38167776107e63307dfc8ad108e580548d1f2c587f42"},
+    {file = "yarl-1.9.4-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d8b889777de69897406c9fb0b76cdf2fd0f31267861ae7501d93003d55f54fbe"},
+    {file = "yarl-1.9.4-cp311-cp311-musllinux_1_1_aarch64.whl", hash = "sha256:03caa9507d3d3c83bca08650678e25364e1843b484f19986a527630ca376ecce"},
+    {file = "yarl-1.9.4-cp311-cp311-musllinux_1_1_i686.whl", hash = "sha256:4e9035df8d0880b2f1c7f5031f33f69e071dfe72ee9310cfc76f7b605958ceb9"},
+    {file = "yarl-1.9.4-cp311-cp311-musllinux_1_1_ppc64le.whl", hash = "sha256:c0ec0ed476f77db9fb29bca17f0a8fcc7bc97ad4c6c1d8959c507decb22e8572"},
+    {file = "yarl-1.9.4-cp311-cp311-musllinux_1_1_s390x.whl", hash = "sha256:ee04010f26d5102399bd17f8df8bc38dc7ccd7701dc77f4a68c5b8d733406958"},
+    {file = "yarl-1.9.4-cp311-cp311-musllinux_1_1_x86_64.whl", hash = "sha256:49a180c2e0743d5d6e0b4d1a9e5f633c62eca3f8a86ba5dd3c471060e352ca98"},
+    {file = "yarl-1.9.4-cp311-cp311-win32.whl", hash = "sha256:81eb57278deb6098a5b62e88ad8281b2ba09f2f1147c4767522353eaa6260b31"},
+    {file = "yarl-1.9.4-cp311-cp311-win_amd64.whl", hash = "sha256:d1d2532b340b692880261c15aee4dc94dd22ca5d61b9db9a8a361953d36410b1"},
+    {file = "yarl-1.9.4-cp312-cp312-macosx_10_9_universal2.whl", hash = "sha256:0d2454f0aef65ea81037759be5ca9947539667eecebca092733b2eb43c965a81"},
+    {file = "yarl-1.9.4-cp312-cp312-macosx_10_9_x86_64.whl", hash = "sha256:44d8ffbb9c06e5a7f529f38f53eda23e50d1ed33c6c869e01481d3fafa6b8142"},
+    {file = "yarl-1.9.4-cp312-cp312-macosx_11_0_arm64.whl", hash = "sha256:aaaea1e536f98754a6e5c56091baa1b6ce2f2700cc4a00b0d49eca8dea471074"},
+    {file = "yarl-1.9.4-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:3777ce5536d17989c91696db1d459574e9a9bd37660ea7ee4d3344579bb6f129"},
+    {file = "yarl-1.9.4-cp312-cp312-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:9fc5fc1eeb029757349ad26bbc5880557389a03fa6ada41703db5e068881e5f2"},
+    {file = "yarl-1.9.4-cp312-cp312-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:ea65804b5dc88dacd4a40279af0cdadcfe74b3e5b4c897aa0d81cf86927fee78"},
+    {file = "yarl-1.9.4-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:aa102d6d280a5455ad6a0f9e6d769989638718e938a6a0a2ff3f4a7ff8c62cc4"},
+    {file = "yarl-1.9.4-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:09efe4615ada057ba2d30df871d2f668af661e971dfeedf0c159927d48bbeff0"},
+    {file = "yarl-1.9.4-cp312-cp312-musllinux_1_1_aarch64.whl", hash = "sha256:008d3e808d03ef28542372d01057fd09168419cdc8f848efe2804f894ae03e51"},
+    {file = "yarl-1.9.4-cp312-cp312-musllinux_1_1_i686.whl", hash = "sha256:6f5cb257bc2ec58f437da2b37a8cd48f666db96d47b8a3115c29f316313654ff"},
+    {file = "yarl-1.9.4-cp312-cp312-musllinux_1_1_ppc64le.whl", hash = "sha256:992f18e0ea248ee03b5a6e8b3b4738850ae7dbb172cc41c966462801cbf62cf7"},
+    {file = "yarl-1.9.4-cp312-cp312-musllinux_1_1_s390x.whl", hash = "sha256:0e9d124c191d5b881060a9e5060627694c3bdd1fe24c5eecc8d5d7d0eb6faabc"},
+    {file = "yarl-1.9.4-cp312-cp312-musllinux_1_1_x86_64.whl", hash = "sha256:3986b6f41ad22988e53d5778f91855dc0399b043fc8946d4f2e68af22ee9ff10"},
+    {file = "yarl-1.9.4-cp312-cp312-win32.whl", hash = "sha256:4b21516d181cd77ebd06ce160ef8cc2a5e9ad35fb1c5930882baff5ac865eee7"},
+    {file = "yarl-1.9.4-cp312-cp312-win_amd64.whl", hash = "sha256:a9bd00dc3bc395a662900f33f74feb3e757429e545d831eef5bb280252631984"},
+    {file = "yarl-1.9.4-cp37-cp37m-macosx_10_9_x86_64.whl", hash = "sha256:63b20738b5aac74e239622d2fe30df4fca4942a86e31bf47a81a0e94c14df94f"},
+    {file = "yarl-1.9.4-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:d7d7f7de27b8944f1fee2c26a88b4dabc2409d2fea7a9ed3df79b67277644e17"},
+    {file = "yarl-1.9.4-cp37-cp37m-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:c74018551e31269d56fab81a728f683667e7c28c04e807ba08f8c9e3bba32f14"},
+    {file = "yarl-1.9.4-cp37-cp37m-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:ca06675212f94e7a610e85ca36948bb8fc023e458dd6c63ef71abfd482481aa5"},
+    {file = "yarl-1.9.4-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5aef935237d60a51a62b86249839b51345f47564208c6ee615ed2a40878dccdd"},
+    {file = "yarl-1.9.4-cp37-cp37m-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:2b134fd795e2322b7684155b7855cc99409d10b2e408056db2b93b51a52accc7"},
+    {file = "yarl-1.9.4-cp37-cp37m-musllinux_1_1_aarch64.whl", hash = "sha256:d25039a474c4c72a5ad4b52495056f843a7ff07b632c1b92ea9043a3d9950f6e"},
+    {file = "yarl-1.9.4-cp37-cp37m-musllinux_1_1_i686.whl", hash = "sha256:f7d6b36dd2e029b6bcb8a13cf19664c7b8e19ab3a58e0fefbb5b8461447ed5ec"},
+    {file = "yarl-1.9.4-cp37-cp37m-musllinux_1_1_ppc64le.whl", hash = "sha256:957b4774373cf6f709359e5c8c4a0af9f6d7875db657adb0feaf8d6cb3c3964c"},
+    {file = "yarl-1.9.4-cp37-cp37m-musllinux_1_1_s390x.whl", hash = "sha256:d7eeb6d22331e2fd42fce928a81c697c9ee2d51400bd1a28803965883e13cead"},
+    {file = "yarl-1.9.4-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:6a962e04b8f91f8c4e5917e518d17958e3bdee71fd1d8b88cdce74dd0ebbf434"},
+    {file = "yarl-1.9.4-cp37-cp37m-win32.whl", hash = "sha256:f3bc6af6e2b8f92eced34ef6a96ffb248e863af20ef4fde9448cc8c9b858b749"},
+    {file = "yarl-1.9.4-cp37-cp37m-win_amd64.whl", hash = "sha256:ad4d7a90a92e528aadf4965d685c17dacff3df282db1121136c382dc0b6014d2"},
+    {file = "yarl-1.9.4-cp38-cp38-macosx_10_9_universal2.whl", hash = "sha256:ec61d826d80fc293ed46c9dd26995921e3a82146feacd952ef0757236fc137be"},
+    {file = "yarl-1.9.4-cp38-cp38-macosx_10_9_x86_64.whl", hash = "sha256:8be9e837ea9113676e5754b43b940b50cce76d9ed7d2461df1af39a8ee674d9f"},
+    {file = "yarl-1.9.4-cp38-cp38-macosx_11_0_arm64.whl", hash = "sha256:bef596fdaa8f26e3d66af846bbe77057237cb6e8efff8cd7cc8dff9a62278bbf"},
+    {file = "yarl-1.9.4-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:2d47552b6e52c3319fede1b60b3de120fe83bde9b7bddad11a69fb0af7db32f1"},
+    {file = "yarl-1.9.4-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:84fc30f71689d7fc9168b92788abc977dc8cefa806909565fc2951d02f6b7d57"},
+    {file = "yarl-1.9.4-cp38-cp38-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:4aa9741085f635934f3a2583e16fcf62ba835719a8b2b28fb2917bb0537c1dfa"},
+    {file = "yarl-1.9.4-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:206a55215e6d05dbc6c98ce598a59e6fbd0c493e2de4ea6cc2f4934d5a18d130"},
+    {file = "yarl-1.9.4-cp38-cp38-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:07574b007ee20e5c375a8fe4a0789fad26db905f9813be0f9fef5a68080de559"},
+    {file = "yarl-1.9.4-cp38-cp38-musllinux_1_1_aarch64.whl", hash = "sha256:5a2e2433eb9344a163aced6a5f6c9222c0786e5a9e9cac2c89f0b28433f56e23"},
+    {file = "yarl-1.9.4-cp38-cp38-musllinux_1_1_i686.whl", hash = "sha256:6ad6d10ed9b67a382b45f29ea028f92d25bc0bc1daf6c5b801b90b5aa70fb9ec"},
+    {file = "yarl-1.9.4-cp38-cp38-musllinux_1_1_ppc64le.whl", hash = "sha256:6fe79f998a4052d79e1c30eeb7d6c1c1056ad33300f682465e1b4e9b5a188b78"},
+    {file = "yarl-1.9.4-cp38-cp38-musllinux_1_1_s390x.whl", hash = "sha256:a825ec844298c791fd28ed14ed1bffc56a98d15b8c58a20e0e08c1f5f2bea1be"},
+    {file = "yarl-1.9.4-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:8619d6915b3b0b34420cf9b2bb6d81ef59d984cb0fde7544e9ece32b4b3043c3"},
+    {file = "yarl-1.9.4-cp38-cp38-win32.whl", hash = "sha256:686a0c2f85f83463272ddffd4deb5e591c98aac1897d65e92319f729c320eece"},
+    {file = "yarl-1.9.4-cp38-cp38-win_amd64.whl", hash = "sha256:a00862fb23195b6b8322f7d781b0dc1d82cb3bcac346d1e38689370cc1cc398b"},
+    {file = "yarl-1.9.4-cp39-cp39-macosx_10_9_universal2.whl", hash = "sha256:604f31d97fa493083ea21bd9b92c419012531c4e17ea6da0f65cacdcf5d0bd27"},
+    {file = "yarl-1.9.4-cp39-cp39-macosx_10_9_x86_64.whl", hash = "sha256:8a854227cf581330ffa2c4824d96e52ee621dd571078a252c25e3a3b3d94a1b1"},
+    {file = "yarl-1.9.4-cp39-cp39-macosx_11_0_arm64.whl", hash = "sha256:ba6f52cbc7809cd8d74604cce9c14868306ae4aa0282016b641c661f981a6e91"},
+    {file = "yarl-1.9.4-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:a6327976c7c2f4ee6816eff196e25385ccc02cb81427952414a64811037bbc8b"},
+    {file = "yarl-1.9.4-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:8397a3817d7dcdd14bb266283cd1d6fc7264a48c186b986f32e86d86d35fbac5"},
+    {file = "yarl-1.9.4-cp39-cp39-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:e0381b4ce23ff92f8170080c97678040fc5b08da85e9e292292aba67fdac6c34"},
+    {file = "yarl-1.9.4-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:23d32a2594cb5d565d358a92e151315d1b2268bc10f4610d098f96b147370136"},
+    {file = "yarl-1.9.4-cp39-cp39-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:ddb2a5c08a4eaaba605340fdee8fc08e406c56617566d9643ad8bf6852778fc7"},
+    {file = "yarl-1.9.4-cp39-cp39-musllinux_1_1_aarch64.whl", hash = "sha256:26a1dc6285e03f3cc9e839a2da83bcbf31dcb0d004c72d0730e755b33466c30e"},
+    {file = "yarl-1.9.4-cp39-cp39-musllinux_1_1_i686.whl", hash = "sha256:18580f672e44ce1238b82f7fb87d727c4a131f3a9d33a5e0e82b793362bf18b4"},
+    {file = "yarl-1.9.4-cp39-cp39-musllinux_1_1_ppc64le.whl", hash = "sha256:29e0f83f37610f173eb7e7b5562dd71467993495e568e708d99e9d1944f561ec"},
+    {file = "yarl-1.9.4-cp39-cp39-musllinux_1_1_s390x.whl", hash = "sha256:1f23e4fe1e8794f74b6027d7cf19dc25f8b63af1483d91d595d4a07eca1fb26c"},
+    {file = "yarl-1.9.4-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:db8e58b9d79200c76956cefd14d5c90af54416ff5353c5bfd7cbe58818e26ef0"},
+    {file = "yarl-1.9.4-cp39-cp39-win32.whl", hash = "sha256:c7224cab95645c7ab53791022ae77a4509472613e839dab722a72abe5a684575"},
+    {file = "yarl-1.9.4-cp39-cp39-win_amd64.whl", hash = "sha256:824d6c50492add5da9374875ce72db7a0733b29c2394890aef23d533106e2b15"},
+    {file = "yarl-1.9.4-py3-none-any.whl", hash = "sha256:928cecb0ef9d5a7946eb6ff58417ad2fe9375762382f1bf5c55e61645f2c43ad"},
+    {file = "yarl-1.9.4.tar.gz", hash = "sha256:566db86717cf8080b99b58b083b773a908ae40f06681e87e589a976faf8246bf"},
+]
+
+[package.dependencies]
+idna = ">=2.0"
+multidict = ">=4.0"
+
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.10"
-content-hash = "274fed55cf4a2f4e9954b3d196c103b72225409c6050759a939f0ca197ae3f79"
+content-hash = "9cec287b530dcda39515e7c76eed76d88eb0c725f37148af5ef93083cfb46ad8"

backend/prompts/__init__.py

@@ -1,10 +1,12 @@
-from typing import List, NoReturn, Union
-
+from typing import Union
 from openai.types.chat import ChatCompletionMessageParam, ChatCompletionContentPartParam
 
+from custom_types import InputMode
+from image_generation.core import create_alt_url_mapping
 from prompts.imported_code_prompts import IMPORTED_CODE_SYSTEM_PROMPTS
 from prompts.screenshot_system_prompts import SYSTEM_PROMPTS
 from prompts.types import Stack
+from video.utils import assemble_claude_prompt_video
 
 
 USER_PROMPT = """
@@ -16,9 +18,65 @@ Generate code for a SVG that looks exactly like this.
 """
 
 
+async def create_prompt(
+    params: dict[str, str], stack: Stack, input_mode: InputMode
+) -> tuple[list[ChatCompletionMessageParam], dict[str, str]]:
+
+    image_cache: dict[str, str] = {}
+
+    # If this generation started off with imported code, we need to assemble the prompt differently
+    if params.get("isImportedFromCode"):
+        original_imported_code = params["history"][0]
+        prompt_messages = assemble_imported_code_prompt(original_imported_code, stack)
+        for index, text in enumerate(params["history"][1:]):
+            if index % 2 == 0:
+                message: ChatCompletionMessageParam = {
+                    "role": "user",
+                    "content": text,
+                }
+            else:
+                message: ChatCompletionMessageParam = {
+                    "role": "assistant",
+                    "content": text,
+                }
+            prompt_messages.append(message)
+    else:
+        # Assemble the prompt for non-imported code
+        if params.get("resultImage"):
+            prompt_messages = assemble_prompt(
+                params["image"], stack, params["resultImage"]
+            )
+        else:
+            prompt_messages = assemble_prompt(params["image"], stack)
+
+        if params["generationType"] == "update":
+            # Transform the history tree into message format
+            # TODO: Move this to frontend
+            for index, text in enumerate(params["history"]):
+                if index % 2 == 0:
+                    message: ChatCompletionMessageParam = {
+                        "role": "assistant",
+                        "content": text,
+                    }
+                else:
+                    message: ChatCompletionMessageParam = {
+                        "role": "user",
+                        "content": text,
+                    }
+                prompt_messages.append(message)
+
+            image_cache = create_alt_url_mapping(params["history"][-2])
+
+    if input_mode == "video":
+        video_data_url = params["image"]
+        prompt_messages = await assemble_claude_prompt_video(video_data_url)
+
+    return prompt_messages, image_cache
+
+
 def assemble_imported_code_prompt(
-    code: str, stack: Stack, result_image_data_url: Union[str, None] = None
-) -> List[ChatCompletionMessageParam]:
+    code: str, stack: Stack
+) -> list[ChatCompletionMessageParam]:
     system_content = IMPORTED_CODE_SYSTEM_PROMPTS[stack]
 
     user_content = (
@@ -26,15 +84,12 @@ def assemble_imported_code_prompt(
         if stack != "svg"
         else "Here is the code of the SVG: " + code
     )
+
     return [
         {
             "role": "system",
-            "content": system_content,
-        },
-        {
-            "role": "user",
-            "content": user_content,
-        },
+            "content": system_content + "\n " + user_content,
+        }
     ]
     # TODO: Use result_image_data_url
 
@@ -43,11 +98,11 @@ def assemble_prompt(
     image_data_url: str,
     stack: Stack,
     result_image_data_url: Union[str, None] = None,
-) -> List[ChatCompletionMessageParam]:
+) -> list[ChatCompletionMessageParam]:
     system_content = SYSTEM_PROMPTS[stack]
     user_prompt = USER_PROMPT if stack != "svg" else SVG_USER_PROMPT
 
-    user_content: List[ChatCompletionContentPartParam] = [
+    user_content: list[ChatCompletionContentPartParam] = [
         {
             "type": "image_url",
             "image_url": {"url": image_data_url, "detail": "high"},

backend/prompts/imported_code_prompts.py

@@ -18,6 +18,22 @@ Return only the full code in <html></html> tags.
 Do not include markdown "```" or "```html" at the start or end.
 """
 
+IMPORTED_CODE_HTML_CSS_SYSTEM_PROMPT = """
+You are an expert CSS developer.
+
+- Do not add comments in the code such as "<!-- Add other navigation links as needed -->" and "<!-- ... other news items ... -->" in place of writing the full code. WRITE THE FULL CODE.
+- Repeat elements as needed. For example, if there are 15 items, the code should have 15 items. DO NOT LEAVE comments like "<!-- Repeat for each news item -->" or bad things will happen.
+- For images, use placeholder images from https://placehold.co and include a detailed description of the image in the alt text so that an image generation AI can generate the image later.
+
+In terms of libraries,
+
+- You can use Google Fonts
+- Font Awesome for icons: <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.15.3/css/all.min.css"></link>
+
+Return only the full code in <html></html> tags.
+Do not include markdown "```" or "```html" at the start or end.
+"""
+
 IMPORTED_CODE_REACT_TAILWIND_SYSTEM_PROMPT = """
 You are an expert React/Tailwind developer
 
@@ -128,6 +144,7 @@ Do not include markdown "```" or "```svg" at the start or end.
 
 IMPORTED_CODE_SYSTEM_PROMPTS = SystemPrompts(
     html_tailwind=IMPORTED_CODE_TAILWIND_SYSTEM_PROMPT,
+    html_css=IMPORTED_CODE_HTML_CSS_SYSTEM_PROMPT,
     react_tailwind=IMPORTED_CODE_REACT_TAILWIND_SYSTEM_PROMPT,
     bootstrap=IMPORTED_CODE_BOOTSTRAP_SYSTEM_PROMPT,
     ionic_tailwind=IMPORTED_CODE_IONIC_TAILWIND_SYSTEM_PROMPT,

backend/prompts/screenshot_system_prompts.py

@@ -26,6 +26,30 @@ Return only the full code in <html></html> tags.
 Do not include markdown "```" or "```html" at the start or end.
 """
 
+HTML_CSS_SYSTEM_PROMPT = """
+You are an expert CSS developer
+You take screenshots of a reference web page from the user, and then build single page apps 
+using CSS, HTML and JS.
+You might also be given a screenshot(The second image) of a web page that you have already built, and asked to
+update it to look more like the reference image(The first image).
+
+- Make sure the app looks exactly like the screenshot.
+- Pay close attention to background color, text color, font size, font family, 
+padding, margin, border, etc. Match the colors and sizes exactly.
+- Use the exact text from the screenshot.
+- Do not add comments in the code such as "<!-- Add other navigation links as needed -->" and "<!-- ... other news items ... -->" in place of writing the full code. WRITE THE FULL CODE.
+- Repeat elements as needed to match the screenshot. For example, if there are 15 items, the code should have 15 items. DO NOT LEAVE comments like "<!-- Repeat for each news item -->" or bad things will happen.
+- For images, use placeholder images from https://placehold.co and include a detailed description of the image in the alt text so that an image generation AI can generate the image later.
+
+In terms of libraries,
+
+- You can use Google Fonts
+- Font Awesome for icons: <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.15.3/css/all.min.css"></link>
+
+Return only the full code in <html></html> tags.
+Do not include markdown "```" or "```html" at the start or end.
+"""
+
 BOOTSTRAP_SYSTEM_PROMPT = """
 You are an expert Bootstrap developer
 You take screenshots of a reference web page from the user, and then build single page apps 
@@ -176,6 +200,7 @@ Do not include markdown "```" or "```svg" at the start or end.
 
 
 SYSTEM_PROMPTS = SystemPrompts(
+    html_css=HTML_CSS_SYSTEM_PROMPT,
     html_tailwind=HTML_TAILWIND_SYSTEM_PROMPT,
     react_tailwind=REACT_TAILWIND_SYSTEM_PROMPT,
     bootstrap=BOOTSTRAP_SYSTEM_PROMPT,

backend/prompts/test_prompts.py

@@ -1,3 +1,4 @@
+from llm import Llm
 from prompts import assemble_imported_code_prompt, assemble_prompt
 
 TAILWIND_SYSTEM_PROMPT = """
@@ -25,6 +26,30 @@ Return only the full code in <html></html> tags.
 Do not include markdown "```" or "```html" at the start or end.
 """
 
+HTML_CSS_SYSTEM_PROMPT = """
+You are an expert CSS developer
+You take screenshots of a reference web page from the user, and then build single page apps 
+using CSS, HTML and JS.
+You might also be given a screenshot(The second image) of a web page that you have already built, and asked to
+update it to look more like the reference image(The first image).
+
+- Make sure the app looks exactly like the screenshot.
+- Pay close attention to background color, text color, font size, font family, 
+padding, margin, border, etc. Match the colors and sizes exactly.
+- Use the exact text from the screenshot.
+- Do not add comments in the code such as "<!-- Add other navigation links as needed -->" and "<!-- ... other news items ... -->" in place of writing the full code. WRITE THE FULL CODE.
+- Repeat elements as needed to match the screenshot. For example, if there are 15 items, the code should have 15 items. DO NOT LEAVE comments like "<!-- Repeat for each news item -->" or bad things will happen.
+- For images, use placeholder images from https://placehold.co and include a detailed description of the image in the alt text so that an image generation AI can generate the image later.
+
+In terms of libraries,
+
+- You can use Google Fonts
+- Font Awesome for icons: <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.15.3/css/all.min.css"></link>
+
+Return only the full code in <html></html> tags.
+Do not include markdown "```" or "```html" at the start or end.
+"""
+
 BOOTSTRAP_SYSTEM_PROMPT = """
 You are an expert Bootstrap developer
 You take screenshots of a reference web page from the user, and then build single page apps 
@@ -189,6 +214,22 @@ Return only the full code in <html></html> tags.
 Do not include markdown "```" or "```html" at the start or end.
 """
 
+IMPORTED_CODE_HTML_CSS_SYSTEM_PROMPT = """
+You are an expert CSS developer.
+
+- Do not add comments in the code such as "<!-- Add other navigation links as needed -->" and "<!-- ... other news items ... -->" in place of writing the full code. WRITE THE FULL CODE.
+- Repeat elements as needed. For example, if there are 15 items, the code should have 15 items. DO NOT LEAVE comments like "<!-- Repeat for each news item -->" or bad things will happen.
+- For images, use placeholder images from https://placehold.co and include a detailed description of the image in the alt text so that an image generation AI can generate the image later.
+
+In terms of libraries,
+
+- You can use Google Fonts
+- Font Awesome for icons: <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/5.15.3/css/all.min.css"></link>
+
+Return only the full code in <html></html> tags.
+Do not include markdown "```" or "```html" at the start or end.
+"""
+
 IMPORTED_CODE_REACT_TAILWIND_SYSTEM_PROMPT = """
 You are an expert React/Tailwind developer
 
@@ -311,87 +352,120 @@ def test_prompts():
     tailwind_prompt = assemble_prompt(
         "image_data_url", "html_tailwind", "result_image_data_url"
     )
-    assert tailwind_prompt[0]["content"] == TAILWIND_SYSTEM_PROMPT
+    assert tailwind_prompt[0].get("content") == TAILWIND_SYSTEM_PROMPT
     assert tailwind_prompt[1]["content"][2]["text"] == USER_PROMPT  # type: ignore
 
+    html_css_prompt = assemble_prompt(
+        "image_data_url", "html_css", "result_image_data_url"
+    )
+    assert html_css_prompt[0].get("content") == HTML_CSS_SYSTEM_PROMPT
+    assert html_css_prompt[1]["content"][2]["text"] == USER_PROMPT  # type: ignore
+
     react_tailwind_prompt = assemble_prompt(
         "image_data_url", "react_tailwind", "result_image_data_url"
     )
-    assert react_tailwind_prompt[0]["content"] == REACT_TAILWIND_SYSTEM_PROMPT
+    assert react_tailwind_prompt[0].get("content") == REACT_TAILWIND_SYSTEM_PROMPT
     assert react_tailwind_prompt[1]["content"][2]["text"] == USER_PROMPT  # type: ignore
 
     bootstrap_prompt = assemble_prompt(
         "image_data_url", "bootstrap", "result_image_data_url"
     )
-    assert bootstrap_prompt[0]["content"] == BOOTSTRAP_SYSTEM_PROMPT
+    assert bootstrap_prompt[0].get("content") == BOOTSTRAP_SYSTEM_PROMPT
     assert bootstrap_prompt[1]["content"][2]["text"] == USER_PROMPT  # type: ignore
 
     ionic_tailwind = assemble_prompt(
         "image_data_url", "ionic_tailwind", "result_image_data_url"
     )
-    assert ionic_tailwind[0]["content"] == IONIC_TAILWIND_SYSTEM_PROMPT
+    assert ionic_tailwind[0].get("content") == IONIC_TAILWIND_SYSTEM_PROMPT
     assert ionic_tailwind[1]["content"][2]["text"] == USER_PROMPT  # type: ignore
 
     vue_tailwind = assemble_prompt(
         "image_data_url", "vue_tailwind", "result_image_data_url"
     )
-    assert vue_tailwind[0]["content"] == VUE_TAILWIND_SYSTEM_PROMPT
+    assert vue_tailwind[0].get("content") == VUE_TAILWIND_SYSTEM_PROMPT
     assert vue_tailwind[1]["content"][2]["text"] == USER_PROMPT  # type: ignore
 
     svg_prompt = assemble_prompt("image_data_url", "svg", "result_image_data_url")
-    assert svg_prompt[0]["content"] == SVG_SYSTEM_PROMPT
+    assert svg_prompt[0].get("content") == SVG_SYSTEM_PROMPT
     assert svg_prompt[1]["content"][2]["text"] == SVG_USER_PROMPT  # type: ignore
 
 
 def test_imported_code_prompts():
-    tailwind_prompt = assemble_imported_code_prompt(
-        "code", "html_tailwind", "result_image_data_url"
-    )
+    code = "Sample code"
+
+    tailwind_prompt = assemble_imported_code_prompt(code, "html_tailwind")
     expected_tailwind_prompt = [
-        {"role": "system", "content": IMPORTED_CODE_TAILWIND_SYSTEM_PROMPT},
-        {"role": "user", "content": "Here is the code of the app: code"},
+        {
+            "role": "system",
+            "content": IMPORTED_CODE_TAILWIND_SYSTEM_PROMPT
+            + "\n Here is the code of the app: "
+            + code,
+        }
     ]
     assert tailwind_prompt == expected_tailwind_prompt
 
-    react_tailwind_prompt = assemble_imported_code_prompt(
-        "code", "react_tailwind", "result_image_data_url"
-    )
+    html_css_prompt = assemble_imported_code_prompt(code, "html_css")
+    expected_html_css_prompt = [
+        {
+            "role": "system",
+            "content": IMPORTED_CODE_HTML_CSS_SYSTEM_PROMPT
+            + "\n Here is the code of the app: "
+            + code,
+        }
+    ]
+    assert html_css_prompt == expected_html_css_prompt
+
+    react_tailwind_prompt = assemble_imported_code_prompt(code, "react_tailwind")
     expected_react_tailwind_prompt = [
-        {"role": "system", "content": IMPORTED_CODE_REACT_TAILWIND_SYSTEM_PROMPT},
-        {"role": "user", "content": "Here is the code of the app: code"},
+        {
+            "role": "system",
+            "content": IMPORTED_CODE_REACT_TAILWIND_SYSTEM_PROMPT
+            + "\n Here is the code of the app: "
+            + code,
+        }
     ]
     assert react_tailwind_prompt == expected_react_tailwind_prompt
 
-    bootstrap_prompt = assemble_imported_code_prompt(
-        "code", "bootstrap", "result_image_data_url"
-    )
+    bootstrap_prompt = assemble_imported_code_prompt(code, "bootstrap")
     expected_bootstrap_prompt = [
-        {"role": "system", "content": IMPORTED_CODE_BOOTSTRAP_SYSTEM_PROMPT},
-        {"role": "user", "content": "Here is the code of the app: code"},
+        {
+            "role": "system",
+            "content": IMPORTED_CODE_BOOTSTRAP_SYSTEM_PROMPT
+            + "\n Here is the code of the app: "
+            + code,
+        }
     ]
     assert bootstrap_prompt == expected_bootstrap_prompt
 
-    ionic_tailwind = assemble_imported_code_prompt(
-        "code", "ionic_tailwind", "result_image_data_url"
-    )
+    ionic_tailwind = assemble_imported_code_prompt(code, "ionic_tailwind")
     expected_ionic_tailwind = [
-        {"role": "system", "content": IMPORTED_CODE_IONIC_TAILWIND_SYSTEM_PROMPT},
-        {"role": "user", "content": "Here is the code of the app: code"},
+        {
+            "role": "system",
+            "content": IMPORTED_CODE_IONIC_TAILWIND_SYSTEM_PROMPT
+            + "\n Here is the code of the app: "
+            + code,
+        }
     ]
     assert ionic_tailwind == expected_ionic_tailwind
 
-    vue_tailwind = assemble_imported_code_prompt(
-        "code", "vue_tailwind", "result_image_data_url"
-    )
+    vue_tailwind = assemble_imported_code_prompt(code, "vue_tailwind")
     expected_vue_tailwind = [
-        {"role": "system", "content": IMPORTED_CODE_VUE_TAILWIND_PROMPT},
-        {"role": "user", "content": "Here is the code of the app: code"},
+        {
+            "role": "system",
+            "content": IMPORTED_CODE_VUE_TAILWIND_PROMPT
+            + "\n Here is the code of the app: "
+            + code,
+        }
     ]
     assert vue_tailwind == expected_vue_tailwind
 
-    svg = assemble_imported_code_prompt("code", "svg", "result_image_data_url")
+    svg = assemble_imported_code_prompt(code, "svg")
     expected_svg = [
-        {"role": "system", "content": IMPORTED_CODE_SVG_SYSTEM_PROMPT},
-        {"role": "user", "content": "Here is the code of the SVG: code"},
+        {
+            "role": "system",
+            "content": IMPORTED_CODE_SVG_SYSTEM_PROMPT
+            + "\n Here is the code of the SVG: "
+            + code,
+        }
     ]
     assert svg == expected_svg

backend/prompts/types.py

@@ -2,6 +2,7 @@ from typing import Literal, TypedDict
 
 
 class SystemPrompts(TypedDict):
+    html_css: str
     html_tailwind: str
     react_tailwind: str
     bootstrap: str
@@ -11,6 +12,7 @@ class SystemPrompts(TypedDict):
 
 
 Stack = Literal[
+    "html_css",
     "html_tailwind",
     "react_tailwind",
     "bootstrap",

backend/pyproject.toml

@@ -17,6 +17,9 @@ httpx = "^0.25.1"
 pre-commit = "^3.6.2"
 anthropic = "^0.18.0"
 moviepy = "^1.0.3"
+pillow = "^10.3.0"
+types-pillow = "^10.2.0.20240520"
+aiohttp = "^3.9.5"
 
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.4.3"

backend/routes/evals.py

@@ -7,10 +7,13 @@ from evals.config import EVALS_DIR
 
 router = APIRouter()
 
+# Update this if the number of outputs generated per input changes
+N = 1
+
 
 class Eval(BaseModel):
     input: str
-    output: str
+    outputs: list[str]
 
 
 @router.get("/evals")
@@ -25,21 +28,27 @@ async def get_evals():
             input_file_path = os.path.join(input_dir, file)
             input_file = await image_to_data_url(input_file_path)
 
-            # Construct the corresponding output file name
-            output_file_name = file.replace(".png", ".html")
-            output_file_path = os.path.join(output_dir, output_file_name)
+            # Construct the corresponding output file names
+            output_file_names = [
+                file.replace(".png", f"_{i}.html") for i in range(0, N)
+            ]  # Assuming 3 outputs for each input
 
-            # Check if the output file exists
-            if os.path.exists(output_file_path):
-                with open(output_file_path, "r") as f:
-                    output_file_data = f.read()
-            else:
-                output_file_data = "Output file not found."
+            output_files_data: list[str] = []
+            for output_file_name in output_file_names:
+                output_file_path = os.path.join(output_dir, output_file_name)
+                # Check if the output file exists
+                if os.path.exists(output_file_path):
+                    with open(output_file_path, "r") as f:
+                        output_files_data.append(f.read())
+                else:
+                    output_files_data.append(
+                        "<html><h1>Output file not found.</h1></html>"
+                    )
 
             evals.append(
                 Eval(
                     input=input_file,
-                    output=output_file_data,
+                    outputs=output_files_data,
                 )
             )
 

backend/routes/generate_code.py

@@ -1,265 +1,343 @@
-import os
+import asyncio
+from dataclasses import dataclass
 import traceback
 from fastapi import APIRouter, WebSocket
 import openai
-from config import ANTHROPIC_API_KEY, IS_PROD, SHOULD_MOCK_AI_RESPONSE
+from codegen.utils import extract_html_content
+from config import (
+    ANTHROPIC_API_KEY,
+    IS_PROD,
+    NUM_VARIANTS,
+    OPENAI_API_KEY,
+    OPENAI_BASE_URL,
+    REPLICATE_API_KEY,
+    SHOULD_MOCK_AI_RESPONSE,
+)
 from custom_types import InputMode
 from llm import (
-    CODE_GENERATION_MODELS,
-    MODEL_CLAUDE_OPUS,
+    Llm,
+    convert_frontend_str_to_llm,
     stream_claude_response,
     stream_claude_response_native,
     stream_openai_response,
 )
-from openai.types.chat import ChatCompletionMessageParam
+from fs_logging.core import write_logs
 from mock_llm import mock_completion
-from typing import Dict, List, cast, get_args
-from image_generation import create_alt_url_mapping, generate_images
-from prompts import assemble_imported_code_prompt, assemble_prompt
-from access_token import validate_access_token
-from datetime import datetime
-import json
+from typing import Any, Callable, Coroutine, Dict, List, Literal, cast, get_args
+from image_generation.core import generate_images
+from prompts import create_prompt
 from prompts.claude_prompts import VIDEO_PROMPT
 from prompts.types import Stack
 
 # from utils import pprint_prompt
-from video.utils import extract_tag_content, assemble_claude_prompt_video  # type: ignore
+from ws.constants import APP_ERROR_WEB_SOCKET_CODE  # type: ignore
 
 
 router = APIRouter()
 
 
-def write_logs(prompt_messages: List[ChatCompletionMessageParam], completion: str):
-    # Get the logs path from environment, default to the current working directory
-    logs_path = os.environ.get("LOGS_PATH", os.getcwd())
+# Auto-upgrade usage of older models
+def auto_upgrade_model(code_generation_model: Llm) -> Llm:
+    if code_generation_model in {Llm.GPT_4_VISION, Llm.GPT_4_TURBO_2024_04_09}:
+        print(
+            f"Initial deprecated model: {code_generation_model}. Auto-updating code generation model to GPT-4O-2024-05-13"
+        )
+        return Llm.GPT_4O_2024_05_13
+    elif code_generation_model == Llm.CLAUDE_3_SONNET:
+        print(
+            f"Initial deprecated model: {code_generation_model}. Auto-updating code generation model to CLAUDE-3.5-SONNET-2024-06-20"
+        )
+        return Llm.CLAUDE_3_5_SONNET_2024_06_20
+    return code_generation_model
 
-    # Create run_logs directory if it doesn't exist within the specified logs path
-    logs_directory = os.path.join(logs_path, "run_logs")
-    if not os.path.exists(logs_directory):
-        os.makedirs(logs_directory)
 
-    print("Writing to logs directory:", logs_directory)
+# Generate images, if needed
+async def perform_image_generation(
+    completion: str,
+    should_generate_images: bool,
+    openai_api_key: str | None,
+    openai_base_url: str | None,
+    image_cache: dict[str, str],
+):
+    replicate_api_key = REPLICATE_API_KEY
+    if not should_generate_images:
+        return completion
 
-    # Generate a unique filename using the current timestamp within the logs directory
-    filename = datetime.now().strftime(f"{logs_directory}/messages_%Y%m%d_%H%M%S.json")
+    if replicate_api_key:
+        image_generation_model = "flux"
+        api_key = replicate_api_key
+    else:
+        if not openai_api_key:
+            print(
+                "No OpenAI API key and Replicate key found. Skipping image generation."
+            )
+            return completion
+        image_generation_model = "dalle3"
+        api_key = openai_api_key
 
-    # Write the messages dict into a new file for each run
-    with open(filename, "w") as f:
-        f.write(json.dumps({"prompt": prompt_messages, "completion": completion}))
+    print("Generating images with model: ", image_generation_model)
+
+    return await generate_images(
+        completion,
+        api_key=api_key,
+        base_url=openai_base_url,
+        image_cache=image_cache,
+        model=image_generation_model,
+    )
+
+
+@dataclass
+class ExtractedParams:
+    stack: Stack
+    input_mode: InputMode
+    code_generation_model: Llm
+    should_generate_images: bool
+    openai_api_key: str | None
+    anthropic_api_key: str | None
+    openai_base_url: str | None
+
+
+async def extract_params(
+    params: Dict[str, str], throw_error: Callable[[str], Coroutine[Any, Any, None]]
+) -> ExtractedParams:
+    # Read the code config settings (stack) from the request.
+    generated_code_config = params.get("generatedCodeConfig", "")
+    if generated_code_config not in get_args(Stack):
+        await throw_error(f"Invalid generated code config: {generated_code_config}")
+        raise ValueError(f"Invalid generated code config: {generated_code_config}")
+    validated_stack = cast(Stack, generated_code_config)
+
+    # Validate the input mode
+    input_mode = params.get("inputMode")
+    if input_mode not in get_args(InputMode):
+        await throw_error(f"Invalid input mode: {input_mode}")
+        raise ValueError(f"Invalid input mode: {input_mode}")
+    validated_input_mode = cast(InputMode, input_mode)
+
+    # Read the model from the request. Fall back to default if not provided.
+    code_generation_model_str = params.get(
+        "codeGenerationModel", Llm.GPT_4O_2024_05_13.value
+    )
+    try:
+        code_generation_model = convert_frontend_str_to_llm(code_generation_model_str)
+    except ValueError:
+        await throw_error(f"Invalid model: {code_generation_model_str}")
+        raise ValueError(f"Invalid model: {code_generation_model_str}")
+
+    openai_api_key = get_from_settings_dialog_or_env(
+        params, "openAiApiKey", OPENAI_API_KEY
+    )
+
+    # If neither is provided, we throw an error later only if Claude is used.
+    anthropic_api_key = get_from_settings_dialog_or_env(
+        params, "anthropicApiKey", ANTHROPIC_API_KEY
+    )
+
+    # Base URL for OpenAI API
+    openai_base_url: str | None = None
+    # Disable user-specified OpenAI Base URL in prod
+    if not IS_PROD:
+        openai_base_url = get_from_settings_dialog_or_env(
+            params, "openAiBaseURL", OPENAI_BASE_URL
+        )
+    if not openai_base_url:
+        print("Using official OpenAI URL")
+
+    # Get the image generation flag from the request. Fall back to True if not provided.
+    should_generate_images = bool(params.get("isImageGenerationEnabled", True))
+
+    return ExtractedParams(
+        stack=validated_stack,
+        input_mode=validated_input_mode,
+        code_generation_model=code_generation_model,
+        should_generate_images=should_generate_images,
+        openai_api_key=openai_api_key,
+        anthropic_api_key=anthropic_api_key,
+        openai_base_url=openai_base_url,
+    )
+
+
+def get_from_settings_dialog_or_env(
+    params: dict[str, str], key: str, env_var: str | None
+) -> str | None:
+    value = params.get(key)
+    if value:
+        print(f"Using {key} from client-side settings dialog")
+        return value
+
+    if env_var:
+        print(f"Using {key} from environment variable")
+        return env_var
+
+    return None
 
 
 @router.websocket("/generate-code")
 async def stream_code(websocket: WebSocket):
     await websocket.accept()
-
     print("Incoming websocket connection...")
 
+    ## Communication protocol setup
     async def throw_error(
         message: str,
     ):
+        print(message)
         await websocket.send_json({"type": "error", "value": message})
-        await websocket.close()
+        await websocket.close(APP_ERROR_WEB_SOCKET_CODE)
+
+    async def send_message(
+        type: Literal["chunk", "status", "setCode", "error"],
+        value: str,
+        variantIndex: int,
+    ):
+        # Print for debugging on the backend
+        if type == "error":
+            print(f"Error (variant {variantIndex}): {value}")
+        elif type == "status":
+            print(f"Status (variant {variantIndex}): {value}")
+
+        await websocket.send_json(
+            {"type": type, "value": value, "variantIndex": variantIndex}
+        )
+
+    ## Parameter extract and validation
 
     # TODO: Are the values always strings?
-    params: Dict[str, str] = await websocket.receive_json()
-
+    params: dict[str, str] = await websocket.receive_json()
     print("Received params")
 
-    # Read the code config settings from the request. Fall back to default if not provided.
-    generated_code_config = ""
-    if "generatedCodeConfig" in params and params["generatedCodeConfig"]:
-        generated_code_config = params["generatedCodeConfig"]
-    if not generated_code_config in get_args(Stack):
-        await throw_error(f"Invalid generated code config: {generated_code_config}")
-        return
-    # Cast the variable to the Stack type
-    valid_stack = cast(Stack, generated_code_config)
+    extracted_params = await extract_params(params, throw_error)
+    stack = extracted_params.stack
+    input_mode = extracted_params.input_mode
+    code_generation_model = extracted_params.code_generation_model
+    openai_api_key = extracted_params.openai_api_key
+    openai_base_url = extracted_params.openai_base_url
+    anthropic_api_key = extracted_params.anthropic_api_key
+    should_generate_images = extracted_params.should_generate_images
 
-    # Validate the input mode
-    input_mode = params.get("inputMode")
-    if not input_mode in get_args(InputMode):
-        await throw_error(f"Invalid input mode: {input_mode}")
-        raise Exception(f"Invalid input mode: {input_mode}")
-    # Cast the variable to the right type
-    validated_input_mode = cast(InputMode, input_mode)
-
-    # Read the model from the request. Fall back to default if not provided.
-    code_generation_model = params.get("codeGenerationModel", "gpt_4_vision")
-    if code_generation_model not in CODE_GENERATION_MODELS:
-        await throw_error(f"Invalid model: {code_generation_model}")
-        raise Exception(f"Invalid model: {code_generation_model}")
+    # Auto-upgrade usage of older models
+    code_generation_model = auto_upgrade_model(code_generation_model)
 
     print(
-        f"Generating {generated_code_config} code for uploaded {input_mode} using {code_generation_model} model..."
+        f"Generating {stack} code in {input_mode} mode using {code_generation_model}..."
     )
 
-    # Get the OpenAI API key from the request. Fall back to environment variable if not provided.
-    # If neither is provided, we throw an error.
-    openai_api_key = None
-    if "accessCode" in params and params["accessCode"]:
-        print("Access code - using platform API key")
-        res = await validate_access_token(params["accessCode"])
-        if res["success"]:
-            openai_api_key = os.environ.get("PLATFORM_OPENAI_API_KEY")
-        else:
-            await websocket.send_json(
-                {
-                    "type": "error",
-                    "value": res["failure_reason"],
-                }
-            )
-            return
-    else:
-        if params["openAiApiKey"]:
-            openai_api_key = params["openAiApiKey"]
-            print("Using OpenAI API key from client-side settings dialog")
-        else:
-            openai_api_key = os.environ.get("OPENAI_API_KEY")
-            if openai_api_key:
-                print("Using OpenAI API key from environment variable")
+    for i in range(NUM_VARIANTS):
+        await send_message("status", "Generating code...", i)
 
-    if not openai_api_key:
-        print("OpenAI API key not found")
-        await websocket.send_json(
-            {
-                "type": "error",
-                "value": "No OpenAI API key found. Please add your API key in the settings dialog or add it to backend/.env file. If you add it to .env, make sure to restart the backend server.",
-            }
-        )
-        return
-
-    # Get the OpenAI Base URL from the request. Fall back to environment variable if not provided.
-    openai_base_url = None
-    # Disable user-specified OpenAI Base URL in prod
-    if not os.environ.get("IS_PROD"):
-        if "openAiBaseURL" in params and params["openAiBaseURL"]:
-            openai_base_url = params["openAiBaseURL"]
-            print("Using OpenAI Base URL from client-side settings dialog")
-        else:
-            openai_base_url = os.environ.get("OPENAI_BASE_URL")
-            if openai_base_url:
-                print("Using OpenAI Base URL from environment variable")
-
-    if not openai_base_url:
-        print("Using official OpenAI URL")
-
-    # Get the image generation flag from the request. Fall back to True if not provided.
-    should_generate_images = (
-        params["isImageGenerationEnabled"]
-        if "isImageGenerationEnabled" in params
-        else True
-    )
-
-    print("generating code...")
-    await websocket.send_json({"type": "status", "value": "Generating code..."})
-
-    async def process_chunk(content: str):
-        await websocket.send_json({"type": "chunk", "value": content})
+    ### Prompt creation
 
     # Image cache for updates so that we don't have to regenerate images
     image_cache: Dict[str, str] = {}
 
-    # If this generation started off with imported code, we need to assemble the prompt differently
-    if params.get("isImportedFromCode") and params["isImportedFromCode"]:
-        original_imported_code = params["history"][0]
-        prompt_messages = assemble_imported_code_prompt(
-            original_imported_code, valid_stack
+    try:
+        prompt_messages, image_cache = await create_prompt(params, stack, input_mode)
+    except:
+        await throw_error(
+            "Error assembling prompt. Contact support at support@picoapps.xyz"
         )
-        for index, text in enumerate(params["history"][1:]):
-            if index % 2 == 0:
-                message: ChatCompletionMessageParam = {
-                    "role": "user",
-                    "content": text,
-                }
-            else:
-                message: ChatCompletionMessageParam = {
-                    "role": "assistant",
-                    "content": text,
-                }
-            prompt_messages.append(message)
-    else:
-        # Assemble the prompt
-        try:
-            if params.get("resultImage") and params["resultImage"]:
-                prompt_messages = assemble_prompt(
-                    params["image"], valid_stack, params["resultImage"]
-                )
-            else:
-                prompt_messages = assemble_prompt(params["image"], valid_stack)
-        except:
-            await websocket.send_json(
-                {
-                    "type": "error",
-                    "value": "Error assembling prompt. Contact support at support@picoapps.xyz",
-                }
-            )
-            await websocket.close()
-            return
-
-        if params["generationType"] == "update":
-            # Transform the history tree into message format
-            # TODO: Move this to frontend
-            for index, text in enumerate(params["history"]):
-                if index % 2 == 0:
-                    message: ChatCompletionMessageParam = {
-                        "role": "assistant",
-                        "content": text,
-                    }
-                else:
-                    message: ChatCompletionMessageParam = {
-                        "role": "user",
-                        "content": text,
-                    }
-                prompt_messages.append(message)
-
-            image_cache = create_alt_url_mapping(params["history"][-2])
-
-    if validated_input_mode == "video":
-        video_data_url = params["image"]
-        prompt_messages = await assemble_claude_prompt_video(video_data_url)
+        raise
 
     # pprint_prompt(prompt_messages)  # type: ignore
 
+    ### Code generation
+
+    async def process_chunk(content: str, variantIndex: int):
+        await send_message("chunk", content, variantIndex)
+
     if SHOULD_MOCK_AI_RESPONSE:
-        completion = await mock_completion(
-            process_chunk, input_mode=validated_input_mode
-        )
+        completions = [await mock_completion(process_chunk, input_mode=input_mode)]
     else:
         try:
-            if validated_input_mode == "video":
-                if not ANTHROPIC_API_KEY:
+            if input_mode == "video":
+                if not anthropic_api_key:
                     await throw_error(
-                        "No Anthropic API key found. Please add the environment variable ANTHROPIC_API_KEY to backend/.env"
+                        "Video only works with Anthropic models. No Anthropic API key found. Please add the environment variable ANTHROPIC_API_KEY to backend/.env or in the settings dialog"
                     )
                     raise Exception("No Anthropic key")
 
-                completion = await stream_claude_response_native(
-                    system_prompt=VIDEO_PROMPT,
-                    messages=prompt_messages,  # type: ignore
-                    api_key=ANTHROPIC_API_KEY,
-                    callback=lambda x: process_chunk(x),
-                    model=MODEL_CLAUDE_OPUS,
-                    include_thinking=True,
-                )
-            elif code_generation_model == "claude_3_sonnet":
-                if not ANTHROPIC_API_KEY:
-                    await throw_error(
-                        "No Anthropic API key found. Please add the environment variable ANTHROPIC_API_KEY to backend/.env"
+                completions = [
+                    await stream_claude_response_native(
+                        system_prompt=VIDEO_PROMPT,
+                        messages=prompt_messages,  # type: ignore
+                        api_key=anthropic_api_key,
+                        callback=lambda x: process_chunk(x, 0),
+                        model=Llm.CLAUDE_3_OPUS,
+                        include_thinking=True,
                     )
-                    raise Exception("No Anthropic key")
-
-                completion = await stream_claude_response(
-                    prompt_messages,  # type: ignore
-                    api_key=ANTHROPIC_API_KEY,
-                    callback=lambda x: process_chunk(x),
-                )
+                ]
             else:
-                completion = await stream_openai_response(
-                    prompt_messages,  # type: ignore
-                    api_key=openai_api_key,
-                    base_url=openai_base_url,
-                    callback=lambda x: process_chunk(x),
+
+                # Depending on the presence and absence of various keys,
+                # we decide which models to run
+                variant_models = []
+                if openai_api_key and anthropic_api_key:
+                    variant_models = ["anthropic", "openai"]
+                elif openai_api_key:
+                    variant_models = ["openai", "openai"]
+                elif anthropic_api_key:
+                    variant_models = ["anthropic", "anthropic"]
+                else:
+                    await throw_error(
+                        "No OpenAI or Anthropic API key found. Please add the environment variable OPENAI_API_KEY or ANTHROPIC_API_KEY to backend/.env or in the settings dialog. If you add it to .env, make sure to restart the backend server."
+                    )
+                    raise Exception("No OpenAI or Anthropic key")
+
+                tasks: List[Coroutine[Any, Any, str]] = []
+                for index, model in enumerate(variant_models):
+                    if model == "openai":
+                        if openai_api_key is None:
+                            await throw_error("OpenAI API key is missing.")
+                            raise Exception("OpenAI API key is missing.")
+
+                        tasks.append(
+                            stream_openai_response(
+                                prompt_messages,
+                                api_key=openai_api_key,
+                                base_url=openai_base_url,
+                                callback=lambda x, i=index: process_chunk(x, i),
+                                model=Llm.GPT_4O_2024_05_13,
+                            )
+                        )
+                    elif model == "anthropic":
+                        if anthropic_api_key is None:
+                            await throw_error("Anthropic API key is missing.")
+                            raise Exception("Anthropic API key is missing.")
+
+                        tasks.append(
+                            stream_claude_response(
+                                prompt_messages,
+                                api_key=anthropic_api_key,
+                                callback=lambda x, i=index: process_chunk(x, i),
+                                model=Llm.CLAUDE_3_5_SONNET_2024_06_20,
+                            )
+                        )
+
+                # Run the models in parallel and capture exceptions if any
+                completions = await asyncio.gather(*tasks, return_exceptions=True)
+
+                # If all generations failed, throw an error
+                all_generations_failed = all(
+                    isinstance(completion, Exception) for completion in completions
                 )
+                if all_generations_failed:
+                    await throw_error("Error generating code. Please contact support.")
+
+                    # Print the all the underlying exceptions for debugging
+                    for completion in completions:
+                        traceback.print_exception(
+                            type(completion), completion, completion.__traceback__
+                        )
+                    raise Exception("All generations failed")
+
+                # If some completions failed, replace them with empty strings
+                for index, completion in enumerate(completions):
+                    if isinstance(completion, Exception):
+                        completions[index] = ""
+                        print("Generation failed for variant", index)
+
+                print("Models used for generation: ", variant_models)
+
         except openai.AuthenticationError as e:
             print("[GENERATE_CODE] Authentication failed", e)
             error_message = (
@@ -295,37 +373,34 @@ async def stream_code(websocket: WebSocket):
             )
             return await throw_error(error_message)
 
-    if validated_input_mode == "video":
-        completion = extract_tag_content("html", completion)
+    ## Post-processing
+
+    # Strip the completion of everything except the HTML content
+    completions = [extract_html_content(completion) for completion in completions]
 
     # Write the messages dict into a log so that we can debug later
-    write_logs(prompt_messages, completion)  # type: ignore
+    write_logs(prompt_messages, completions[0])
 
-    try:
-        if should_generate_images:
-            await websocket.send_json(
-                {"type": "status", "value": "Generating images..."}
-            )
-            updated_html = await generate_images(
-                completion,
-                api_key=openai_api_key,
-                base_url=openai_base_url,
-                image_cache=image_cache,
-            )
-        else:
-            updated_html = completion
-        await websocket.send_json({"type": "setCode", "value": updated_html})
-        await websocket.send_json(
-            {"type": "status", "value": "Code generation complete."}
-        )
-    except Exception as e:
-        traceback.print_exc()
-        print("Image generation failed", e)
-        # Send set code even if image generation fails since that triggers
-        # the frontend to update history
-        await websocket.send_json({"type": "setCode", "value": completion})
-        await websocket.send_json(
-            {"type": "status", "value": "Image generation failed but code is complete."}
+    ## Image Generation
+
+    for index, _ in enumerate(completions):
+        await send_message("status", "Generating images...", index)
+
+    image_generation_tasks = [
+        perform_image_generation(
+            completion,
+            should_generate_images,
+            openai_api_key,
+            openai_base_url,
+            image_cache,
         )
+        for completion in completions
+    ]
+
+    updated_completions = await asyncio.gather(*image_generation_tasks)
+
+    for index, updated_html in enumerate(updated_completions):
+        await send_message("setCode", updated_html, index)
+        await send_message("status", "Code generation complete.", index)
 
     await websocket.close()

backend/run_evals.py

@@ -1,6 +1,8 @@
 # Load environment variables first
 from dotenv import load_dotenv
 
+from llm import Llm
+
 load_dotenv()
 
 import os
@@ -8,10 +10,12 @@ from typing import Any, Coroutine
 import asyncio
 
 from evals.config import EVALS_DIR
-from evals.core import generate_code_core
+from evals.core import generate_code_for_image
 from evals.utils import image_to_data_url
 
 STACK = "html_tailwind"
+# MODEL = Llm.CLAUDE_3_5_SONNET_2024_06_20
+N = 2  # Number of outputs to generate
 
 
 async def main():
@@ -25,19 +29,79 @@ async def main():
     for filename in evals:
         filepath = os.path.join(INPUT_DIR, filename)
         data_url = await image_to_data_url(filepath)
-        task = generate_code_core(data_url, STACK)
-        tasks.append(task)
+        for n in range(N):  # Generate N tasks for each input
+            if n == 0:
+                task = generate_code_for_image(
+                    image_url=data_url,
+                    stack=STACK,
+                    model=Llm.CLAUDE_3_5_SONNET_2024_06_20,
+                )
+            else:
+                task = generate_code_for_image(
+                    image_url=data_url, stack=STACK, model=Llm.GPT_4O_2024_05_13
+                )
+            tasks.append(task)
+
+    print(f"Generating {len(tasks)} codes")
 
     results = await asyncio.gather(*tasks)
 
     os.makedirs(OUTPUT_DIR, exist_ok=True)
 
-    for filename, content in zip(evals, results):
-        # File name is derived from the original filename in evals
-        output_filename = f"{os.path.splitext(filename)[0]}.html"
+    for i, content in enumerate(results):
+        # Calculate index for filename and output number
+        eval_index = i // N
+        output_number = i % N
+        filename = evals[eval_index]
+        # File name is derived from the original filename in evals with an added output number
+        output_filename = f"{os.path.splitext(filename)[0]}_{output_number}.html"
         output_filepath = os.path.join(OUTPUT_DIR, output_filename)
         with open(output_filepath, "w") as file:
             file.write(content)
 
 
+# async def text_main():
+#     OUTPUT_DIR = EVALS_DIR + "/outputs"
+
+#     GENERAL_TEXT_V1 = [
+#         "Login form",
+#         "Simple notification",
+#         "button",
+#         "saas dashboard",
+#         "landing page for barber shop",
+#     ]
+
+#     tasks: list[Coroutine[Any, Any, str]] = []
+#     for prompt in GENERAL_TEXT_V1:
+#         for n in range(N):  # Generate N tasks for each input
+#             if n == 0:
+#                 task = generate_code_for_text(
+#                     text=prompt,
+#                     stack=STACK,
+#                     model=Llm.CLAUDE_3_5_SONNET_2024_06_20,
+#                 )
+#             else:
+#                 task = generate_code_for_text(
+#                     text=prompt, stack=STACK, model=Llm.GPT_4O_2024_05_13
+#                 )
+#             tasks.append(task)
+
+#     print(f"Generating {len(tasks)} codes")
+
+#     results = await asyncio.gather(*tasks)
+
+#     os.makedirs(OUTPUT_DIR, exist_ok=True)
+
+#     for i, content in enumerate(results):
+#         # Calculate index for filename and output number
+#         eval_index = i // N
+#         output_number = i % N
+#         filename = GENERAL_TEXT_V1[eval_index]
+#         # File name is derived from the original filename in evals with an added output number
+#         output_filename = f"{os.path.splitext(filename)[0]}_{output_number}.html"
+#         output_filepath = os.path.join(OUTPUT_DIR, output_filename)
+#         with open(output_filepath, "w") as file:
+#             file.write(content)
+
+
 asyncio.run(main())

backend/run_image_generation_evals.py

@@ -0,0 +1,85 @@
+import asyncio
+import os
+from typing import List, Optional, Literal
+from dotenv import load_dotenv
+import aiohttp
+from image_generation.core import process_tasks
+
+EVALS = [
+    "Romantic Background",
+    "Company logo: A stylized green sprout emerging from a circle",
+    "Placeholder image of a PDF cover with abstract design",
+    "A complex bubble diagram showing various interconnected features and aspects of FestivalPro, with a large central bubble surrounded by smaller bubbles of different colors representing different categories and functionalities",
+    "A vibrant, abstract visualization of the RhythmRise experience ecosystem, featuring interconnected neon elements representing music, technology, and human connection",
+    "Banner with text 'LiblibAI学院 课程入口'",
+    "Profile picture of Pierre-Louis Labonne",
+    "Two hands holding iPhone 14 models with colorful displays",
+    "Portrait of a woman with long dark hair smiling at the camera",
+    "Threadless logo on a gradient background from light pink to coral",
+    "Jordan Schlansky Shows Conan His Favorite Nose Hair Trimmer",
+    "Team Coco",
+    "Intro to Large Language Models",
+    "Andrej Karpathy",
+    "He built a $200 million toy company",
+    "CNBC International",
+    "What will happen in year three of the war?",
+    "Channel",
+    "This is it",
+    "How ASML Dominates Chip Machines",
+]
+
+# Load environment variables
+load_dotenv()
+
+# Get API keys from environment variables
+OPENAI_API_KEY: Optional[str] = os.getenv("OPENAI_API_KEY")
+REPLICATE_API_TOKEN: Optional[str] = os.getenv("REPLICATE_API_TOKEN")
+
+# Directory to save generated images
+OUTPUT_DIR: str = "generated_images"
+
+
+async def generate_and_save_images(
+    prompts: List[str],
+    model: Literal["dalle3", "flux"],
+    api_key: Optional[str],
+) -> None:
+    # Ensure the output directory exists
+    os.makedirs(OUTPUT_DIR, exist_ok=True)
+
+    if api_key is None:
+        raise ValueError(f"API key for {model} is not set in the environment variables")
+
+    # Generate images
+    results: List[Optional[str]] = await process_tasks(
+        prompts, api_key, None, model=model
+    )
+
+    # Save images to disk
+    async with aiohttp.ClientSession() as session:
+        for i, image_url in enumerate(results):
+            if image_url:
+                # Get the image data
+                async with session.get(image_url) as response:
+                    image_data: bytes = await response.read()
+
+                # Save the image with a filename based on the input eval
+                prefix = "replicate_" if model == "flux" else "dalle3_"
+                filename: str = (
+                    f"{prefix}{prompts[i][:50].replace(' ', '_').replace(':', '')}.png"
+                )
+                filepath: str = os.path.join(OUTPUT_DIR, filename)
+                with open(filepath, "wb") as f:
+                    f.write(image_data)
+                print(f"Saved {model} image: {filepath}")
+            else:
+                print(f"Failed to generate {model} image for prompt: {prompts[i]}")
+
+
+async def main() -> None:
+    # await generate_and_save_images(EVALS, "dalle3", OPENAI_API_KEY)
+    await generate_and_save_images(EVALS, "flux", REPLICATE_API_TOKEN)
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

backend/test_llm.py

@@ -0,0 +1,41 @@
+import unittest
+from llm import convert_frontend_str_to_llm, Llm
+
+
+class TestConvertFrontendStrToLlm(unittest.TestCase):
+    def test_convert_valid_strings(self):
+        self.assertEqual(
+            convert_frontend_str_to_llm("gpt_4_vision"),
+            Llm.GPT_4_VISION,
+            "Should convert 'gpt_4_vision' to Llm.GPT_4_VISION",
+        )
+        self.assertEqual(
+            convert_frontend_str_to_llm("claude_3_sonnet"),
+            Llm.CLAUDE_3_SONNET,
+            "Should convert 'claude_3_sonnet' to Llm.CLAUDE_3_SONNET",
+        )
+        self.assertEqual(
+            convert_frontend_str_to_llm("claude-3-opus-20240229"),
+            Llm.CLAUDE_3_OPUS,
+            "Should convert 'claude-3-opus-20240229' to Llm.CLAUDE_3_OPUS",
+        )
+        self.assertEqual(
+            convert_frontend_str_to_llm("gpt-4-turbo-2024-04-09"),
+            Llm.GPT_4_TURBO_2024_04_09,
+            "Should convert 'gpt-4-turbo-2024-04-09' to Llm.GPT_4_TURBO_2024_04_09",
+        )
+        self.assertEqual(
+            convert_frontend_str_to_llm("gpt-4o-2024-05-13"),
+            Llm.GPT_4O_2024_05_13,
+            "Should convert 'gpt-4o-2024-05-13' to Llm.GPT_4O_2024_05_13",
+        )
+
+    def test_convert_invalid_string_raises_exception(self):
+        with self.assertRaises(ValueError):
+            convert_frontend_str_to_llm("invalid_string")
+        with self.assertRaises(ValueError):
+            convert_frontend_str_to_llm("another_invalid_string")
+
+
+if __name__ == "__main__":
+    unittest.main()

backend/video/utils.py

@@ -5,7 +5,7 @@ import mimetypes
 import os
 import tempfile
 import uuid
-from typing import Union, cast
+from typing import Any, Union, cast
 from moviepy.editor import VideoFileClip  # type: ignore
 from PIL import Image
 import math
@@ -17,7 +17,7 @@ TARGET_NUM_SCREENSHOTS = (
 )
 
 
-async def assemble_claude_prompt_video(video_data_url: str):
+async def assemble_claude_prompt_video(video_data_url: str) -> list[Any]:
     images = split_video_into_screenshots(video_data_url)
 
     # Save images to tmp if we're debugging
@@ -28,7 +28,7 @@ async def assemble_claude_prompt_video(video_data_url: str):
     print(f"Number of frames extracted from video: {len(images)}")
     if len(images) > 20:
         print(f"Too many screenshots: {len(images)}")
-        return
+        raise ValueError("Too many screenshots extracted from video")
 
     # Convert images to the message format for Claude
     content_messages: list[dict[str, Union[dict[str, str], str]]] = []

backend/video_to_app.py

@@ -17,8 +17,7 @@ from utils import pprint_prompt
 from config import ANTHROPIC_API_KEY
 from video.utils import extract_tag_content, assemble_claude_prompt_video
 from llm import (
-    MODEL_CLAUDE_OPUS,
-    # MODEL_CLAUDE_SONNET,
+    Llm,
     stream_claude_response_native,
 )
 
@@ -87,7 +86,7 @@ async def main():
         messages=prompt_messages,
         api_key=ANTHROPIC_API_KEY,
         callback=lambda x: process_chunk(x),
-        model=MODEL_CLAUDE_OPUS,
+        model=Llm.CLAUDE_3_OPUS,
         include_thinking=True,
     )
 

backend/ws/constants.py

@@ -0,0 +1,2 @@
+# WebSocket protocol (RFC 6455) allows for the use of custom close codes in the range 4000-4999
+APP_ERROR_WEB_SOCKET_CODE = 4332

blog/evaluating-claude.md

@@ -1,21 +1,24 @@
-# Evaluating Claude 3 at converting screenshots to code
+# Claude 3 for converting screenshots to code
 
-[video running both]
+Claude 3 dropped yesterday, claiming to rival GPT-4 on a wide variety of tasks. I maintain a very popular open source project called “screenshot-to-code” (this one!) that uses GPT-4 vision to convert screenshots/designs into clean code. Naturally, I was excited to see how good Claude 3 was at this task.
 
-Claude 3 dropped yesterday, claiming to rival GPT-4 on a wide variety of tasks. I maintain a very popular open source project (41k+ Github stars) called “screenshot-to-code” (this one!) that uses GPT-4 vision to convert screenshots/designs into clean code. Naturally, I was excited to see how good Claude 3 was at this task!
+**TLDR:** Claude 3 is on par with GPT-4 vision for screenshot to code, better in some ways but worse in others.
 
 ## Evaluation Setup
 
-I don’t know of a public benchmark for the “screenshot to code” task so I created simple evaluation set up for the purposes of testing\*:
+I don’t know of a public benchmark for “screenshot to code” so I created simple evaluation setup for the purposes of testing:
 
 - **Evaluation Dataset**: 16 screenshots with a mix of UI elements, landing pages, dashboards and popular websites.
-- **Evaluation Metric**: Replication accuracy, as in “How close does the generated code look to the screenshot?” There are other metrics that are important for sure but this is by far the #1 thing that most developers and users of the repo care about. Each output is subjectively rated by a human (me!) on a rating scale from 0 to 4. 4 = very close to an exact replica while 0 = nothing like the screenshot. With 16 screenshots, the maximum any model can score is 64. I like to compare the percentage of the maximum possible score for each run.
+<img width="784" alt="Screenshot 2024-03-05 at 3 05 52 PM" src="https://github.com/abi/screenshot-to-code/assets/23818/c32af2db-eb5a-44c1-9a19-2f0c3dd11ab4">
 
-\*I’ve used it mostly for testing prompt variations until now since no model has come close to GPT-4 vision yet.
+- **Evaluation Metric**: Replication accuracy, as in “How close does the generated code look to the screenshot?” While there are other metrics that are important like code quality, speed and so on, this is by far the #1 thing most users of the repo care about.
+- **Evaluation Mechanism**: Each output is subjectively rated by a human on a rating scale from 0 to 4. 4 = very close to an exact replica while 0 = nothing like the screenshot. With 16 screenshots, the maximum any model can score is 64.
 
-To make the evaluation process easy, I created [a Python script](https://github.com/abi/screenshot-to-code/blob/main/backend/run_evals.py) that runs code for all the inputs in parallel. I also have a simple UI to do a side-by-side comparison of the input and output, shown below.
 
-[video of input/output]
+To make the evaluation process easy, I created [a Python script](https://github.com/abi/screenshot-to-code/blob/main/backend/run_evals.py) that runs code for all the inputs in parallel. I also made a simple UI to do a side-by-side comparison of the input and output.
+
+![Google Chrome](https://github.com/abi/screenshot-to-code/assets/23818/38126f8f-205d-4ed1-b8cf-039e81dcc3d0)
+
 
 ## Results
 
@@ -26,21 +29,31 @@ I only ran the evals on HTML/Tailwind here which is the stack where GPT-4 vision
 Here are the results (average of 3 runs for each model):
 
 - GPT-4 Vision obtains a score of **65.10%** - this is what we’re trying to beat
-- Claude 3 Sonnet receives a score of **70.31%**, which is ~7.5% better.
-- Surprisingly, Claude 3 Opus which is supposed to be the smarter and slower model scores worse than both GPT-4 vision and Claude 3 Sonnet, coming in at **61.46%**. Strange result.
+- Claude 3 Sonnet receives a score of **70.31%**, which is a bit better.
+- Surprisingly, Claude 3 Opus which is supposed to be the smarter and slower model scores worse than both GPT-4 vision and Claude 3 Sonnet, comes in at **61.46%**.
 
-Overall, a very strong showing for Claude 3! Obviously, there's a lot of subjectivity involved in this evaluation but Claude 3 is definitely on par with GPT-4 Vision, if not better. It's also a lot less lazy than GPT-4 Vision.
+Overall, a very strong showing for Claude 3. Obviously, there's a lot of subjectivity involved in this evaluation but Claude 3 is definitely on par with GPT-4 Vision, if not better.
 
-You can see the results for a run of Claude 3 Sonnet here. You can see the results for a run of GPT-4 Vision here.
+You can see the [side-by-side comparison for a run of Claude 3 Sonnet here](https://github.com/abi/screenshot-to-code-files/blob/main/sonnet%20results.png). And for [a run of GPT-4 Vision here](https://github.com/abi/screenshot-to-code-files/blob/main/gpt%204%20vision%20results.png).
 
-Some other notes:
+Other notes:
 
-- The prompts used are optimized for GPT-4 vision. I played around with adjusting the prompts for Claude and that did yield a small improvement. Nothing game-changing and potentially not worth the trade-off of maintaining two sets of prompts.
-- All these models excel at code quality - the quality is usually comparable to a human or better
-- All the models seem to struggle with side-by-side layouts (the struggle is mostly in seeing, not in implementing) TODO: add iamge
-- Claude 3 is much less lazy than GPT-4 Vision. TODO: Hacker News example
-- Claude 3 Sonnet is faster and more steerable
+- The prompts used are optimized for GPT-4 vision. Adjusting the prompts a bit for Claude did yield a small improvement. But nothing game-changing and potentially not worth the trade-off of maintaining two sets of prompts.
+- All the models excel at code quality - the quality is usually comparable to a human or better.
+- Claude 3 is much less lazy than GPT-4 Vision. When asked to recreate Hacker News, GPT-4 Vision will only create two items in the list and leave comments in this code like `<!-- Repeat for each news item -->` and `<!-- ... other news items ... -->`.
+<img width="699" alt="Screenshot 2024-03-05 at 9 25 04 PM" src="https://github.com/abi/screenshot-to-code/assets/23818/04b03155-45e0-40b0-8de0-b1f0b4382bee">
 
-Overall, I'm very impressed with Claude 3 Sonnet as a multimodal model for this use case. I've added it as an alternative to GPT-4 Vision in the open source repo (hosted version update coming soon). I'm sure it's possible to get Opus to be quite good as well.
+While Claude 3 Sonnet can sometimes be lazy too, most of the time, it does what you ask it to do.
 
-If you’d like to contribute to this effort, I have some documentation on [running these evals yourself here](https://github.com/abi/screenshot-to-code/blob/main/Evaluation.md). If you know of a good tool for running evals like this (image input, code output), it would save me a lot of effort from having to build the tool from scratch.
+<img width="904" alt="Screenshot 2024-03-05 at 9 30 23 PM" src="https://github.com/abi/screenshot-to-code/assets/23818/b7c7d1ba-47c1-414d-928f-6989e81cf41d">
+
+- For some reason, all the models struggle with side-by-side "flex" layouts
+<img width="1090" alt="Screenshot 2024-03-05 at 9 20 58 PM" src="https://github.com/abi/screenshot-to-code/assets/23818/8957bb3a-da66-467d-997d-1c7cc24e6d9a">
+
+- Claude 3 Sonnet is a lot faster
+- Claude 3 gets background and text colors wrong quite often! (like in the Hacker News image above)
+- My suspicion is that Claude 3 Opus results can be improved to be on par with the other models through better prompting
+  
+Overall, I'm very impressed with Claude 3 Sonnet for this use case. I've added it as an alternative to GPT-4 Vision in the open source repo (hosted version update coming soon).
+
+If you’d like to contribute to this effort, I have some documentation on [running these evals yourself here](https://github.com/abi/screenshot-to-code/blob/main/Evaluation.md). I'm also working on a better evaluation mechanism with Elo ratings and would love some help on that.

frontend/.gitignore

@@ -25,3 +25,6 @@ dist-ssr
 
 # Env files
 .env*
+
+# Test files
+src/tests/results/

frontend/Dockerfile

@@ -1,4 +1,4 @@
-FROM node:20.9-bullseye-slim
+FROM node:22-bullseye-slim
 
 # Set the working directory in the container
 WORKDIR /app
@@ -6,6 +6,9 @@ WORKDIR /app
 # Copy package.json and yarn.lock
 COPY package.json yarn.lock /app/
 
+# Set the environment variable to skip Puppeteer download
+ENV PUPPETEER_SKIP_DOWNLOAD=true
+
 # Install dependencies
 RUN yarn install
 

frontend/index.html

@@ -2,11 +2,7 @@
 <html lang="en">
   <head>
     <meta charset="UTF-8" />
-    <link
-      rel="icon"
-      type="image/svg+xml"
-      href="https://picoapps.xyz/favicon.png"
-    />
+    <link rel="icon" type="image/png" href="/favicon/main.png" />
     <meta name="viewport" content="width=device-width, initial-scale=1.0" />
 
     <!-- Google Fonts -->

frontend/jest.config.js

@@ -0,0 +1,9 @@
+export default {
+  preset: "ts-jest",
+  testEnvironment: "node",
+  setupFiles: ["<rootDir>/src/setupTests.ts"],
+  transform: {
+    "^.+\\.tsx?$": "ts-jest",
+  },
+  testTimeout: 30000,
+};

frontend/package.json

@@ -10,7 +10,7 @@
     "build-hosted": "tsc && vite build --mode prod",
     "lint": "eslint . --ext ts,tsx --report-unused-disable-directives --max-warnings 0",
     "preview": "vite preview",
-    "test": "vitest"
+    "test": "jest"
   },
   "dependencies": {
     "@codemirror/lang-html": "^6.4.6",
@@ -36,6 +36,7 @@
     "codemirror": "^6.0.1",
     "copy-to-clipboard": "^3.3.3",
     "html2canvas": "^1.4.1",
+    "nanoid": "^5.0.7",
     "react": "^18.2.0",
     "react-dom": "^18.2.0",
     "react-dropzone": "^14.2.3",
@@ -45,21 +46,29 @@
     "tailwind-merge": "^2.0.0",
     "tailwindcss-animate": "^1.0.7",
     "thememirror": "^2.0.1",
-    "vite-plugin-checker": "^0.6.2"
+    "vite-plugin-checker": "^0.6.2",
+    "webm-duration-fix": "^1.0.4",
+    "zustand": "^4.5.2"
   },
   "devDependencies": {
+    "@types/jest": "^29.5.12",
     "@types/node": "^20.9.0",
+    "@types/puppeteer": "^7.0.4",
     "@types/react": "^18.2.15",
     "@types/react-dom": "^18.2.7",
     "@typescript-eslint/eslint-plugin": "^6.0.0",
     "@typescript-eslint/parser": "^6.0.0",
     "@vitejs/plugin-react": "^4.0.3",
     "autoprefixer": "^10.4.16",
+    "dotenv": "^16.4.5",
     "eslint": "^8.45.0",
     "eslint-plugin-react-hooks": "^4.6.0",
     "eslint-plugin-react-refresh": "^0.4.3",
+    "jest": "^29.7.0",
     "postcss": "^8.4.31",
+    "puppeteer": "^22.6.4",
     "tailwindcss": "^3.3.5",
+    "ts-jest": "^29.1.2",
     "typescript": "^5.0.2",
     "vite": "^4.4.5",
     "vite-plugin-html": "^3.2.0",

frontend/src/.env.jest.example

@@ -0,0 +1,2 @@
+TEST_SCREENSHOTONE_API_KEY=
+TEST_ROOT_PATH=

frontend/src/App.tsx

@@ -1,86 +1,100 @@
-import { useEffect, useRef, useState } from "react";
-import ImageUpload from "./components/ImageUpload";
-import CodePreview from "./components/CodePreview";
-import Preview from "./components/Preview";
+import { useEffect, useRef } from "react";
 import { generateCode } from "./generateCode";
-import Spinner from "./components/Spinner";
-import classNames from "classnames";
-import {
-  FaCode,
-  FaDesktop,
-  FaDownload,
-  FaMobile,
-  FaUndo,
-} from "react-icons/fa";
-
-import { Switch } from "./components/ui/switch";
-import { Button } from "@/components/ui/button";
-import { Textarea } from "@/components/ui/textarea";
-import { Tabs, TabsContent, TabsList, TabsTrigger } from "./components/ui/tabs";
-import SettingsDialog from "./components/SettingsDialog";
+import SettingsDialog from "./components/settings/SettingsDialog";
 import { AppState, CodeGenerationParams, EditorTheme, Settings } from "./types";
 import { IS_RUNNING_ON_CLOUD } from "./config";
-import { PicoBadge } from "./components/PicoBadge";
-import { OnboardingNote } from "./components/OnboardingNote";
+import { PicoBadge } from "./components/messages/PicoBadge";
+import { OnboardingNote } from "./components/messages/OnboardingNote";
 import { usePersistedState } from "./hooks/usePersistedState";
-import { UrlInputSection } from "./components/UrlInputSection";
 import TermsOfServiceDialog from "./components/TermsOfServiceDialog";
-import html2canvas from "html2canvas";
 import { USER_CLOSE_WEB_SOCKET_CODE } from "./constants";
-import CodeTab from "./components/CodeTab";
-import OutputSettingsSection from "./components/OutputSettingsSection";
-import { History } from "./components/history/history_types";
-import HistoryDisplay from "./components/history/HistoryDisplay";
-import { extractHistoryTree } from "./components/history/utils";
+import { extractHistory } from "./components/history/utils";
 import toast from "react-hot-toast";
-import ImportCodeSection from "./components/ImportCodeSection";
 import { Stack } from "./lib/stacks";
 import { CodeGenerationModel } from "./lib/models";
-import ModelSettingsSection from "./components/ModelSettingsSection";
-
-const IS_OPENAI_DOWN = false;
+import useBrowserTabIndicator from "./hooks/useBrowserTabIndicator";
+import TipLink from "./components/messages/TipLink";
+import { useAppStore } from "./store/app-store";
+import { useProjectStore } from "./store/project-store";
+import Sidebar from "./components/sidebar/Sidebar";
+import PreviewPane from "./components/preview/PreviewPane";
+import DeprecationMessage from "./components/messages/DeprecationMessage";
+import { GenerationSettings } from "./components/settings/GenerationSettings";
+import StartPane from "./components/start-pane/StartPane";
+import { takeScreenshot } from "./lib/takeScreenshot";
+import { Commit } from "./components/commits/types";
+import { createCommit } from "./components/commits/utils";
 
 function App() {
-  const [appState, setAppState] = useState<AppState>(AppState.INITIAL);
-  const [generatedCode, setGeneratedCode] = useState<string>("");
+  const {
+    // Inputs
+    inputMode,
+    setInputMode,
+    isImportedFromCode,
+    setIsImportedFromCode,
+    referenceImages,
+    setReferenceImages,
 
-  const [inputMode, setInputMode] = useState<"image" | "video">("image");
+    head,
+    commits,
+    addCommit,
+    removeCommit,
+    setHead,
+    appendCommitCode,
+    setCommitCode,
+    resetCommits,
+    resetHead,
 
-  const [referenceImages, setReferenceImages] = useState<string[]>([]);
-  const [executionConsole, setExecutionConsole] = useState<string[]>([]);
-  const [updateInstruction, setUpdateInstruction] = useState("");
-  const [isImportedFromCode, setIsImportedFromCode] = useState<boolean>(false);
+    // Outputs
+    appendExecutionConsole,
+    resetExecutionConsoles,
+  } = useProjectStore();
+
+  const {
+    disableInSelectAndEditMode,
+    setUpdateInstruction,
+    appState,
+    setAppState,
+    shouldIncludeResultImage,
+    setShouldIncludeResultImage,
+  } = useAppStore();
 
   // Settings
   const [settings, setSettings] = usePersistedState<Settings>(
     {
       openAiApiKey: null,
       openAiBaseURL: null,
+      anthropicApiKey: null,
       screenshotOneApiKey: null,
       isImageGenerationEnabled: true,
       editorTheme: EditorTheme.COBALT,
       generatedCodeConfig: Stack.HTML_TAILWIND,
-      codeGenerationModel: CodeGenerationModel.GPT_4_VISION,
+      codeGenerationModel: CodeGenerationModel.CLAUDE_3_5_SONNET_2024_06_20,
       // Only relevant for hosted version
       isTermOfServiceAccepted: false,
-      accessCode: null,
     },
     "setting"
   );
 
+  const wsRef = useRef<WebSocket>(null);
+
   // Code generation model from local storage or the default value
-  const selectedCodeGenerationModel =
+  const model =
     settings.codeGenerationModel || CodeGenerationModel.GPT_4_VISION;
 
-  // App history
-  const [appHistory, setAppHistory] = useState<History>([]);
-  // Tracks the currently shown version from app history
-  const [currentVersion, setCurrentVersion] = useState<number | null>(null);
+  const showBetterModelMessage =
+    model !== CodeGenerationModel.GPT_4O_2024_05_13 &&
+    model !== CodeGenerationModel.CLAUDE_3_5_SONNET_2024_06_20 &&
+    appState === AppState.INITIAL;
 
-  const [shouldIncludeResultImage, setShouldIncludeResultImage] =
-    useState<boolean>(false);
+  const showSelectAndEditFeature =
+    (model === CodeGenerationModel.GPT_4O_2024_05_13 ||
+      model === CodeGenerationModel.CLAUDE_3_5_SONNET_2024_06_20) &&
+    (settings.generatedCodeConfig === Stack.HTML_TAILWIND ||
+      settings.generatedCodeConfig === Stack.HTML_CSS);
 
-  const wsRef = useRef<WebSocket>(null);
+  // Indicate coding state using the browser tab's favicon and title
+  useBrowserTabIndicator(appState === AppState.CODING);
 
   // When the user already has the settings in local storage, newly added keys
   // do not get added to the settings so if it's falsy, we populate it with the default
@@ -94,128 +108,122 @@ function App() {
     }
   }, [settings.generatedCodeConfig, setSettings]);
 
-  const takeScreenshot = async (): Promise<string> => {
-    const iframeElement = document.querySelector(
-      "#preview-desktop"
-    ) as HTMLIFrameElement;
-    if (!iframeElement?.contentWindow?.document.body) {
-      return "";
-    }
-
-    const canvas = await html2canvas(iframeElement.contentWindow.document.body);
-    const png = canvas.toDataURL("image/png");
-    return png;
-  };
-
-  const downloadCode = () => {
-    // Create a blob from the generated code
-    const blob = new Blob([generatedCode], { type: "text/html" });
-    const url = URL.createObjectURL(blob);
-
-    // Create an anchor element and set properties for download
-    const a = document.createElement("a");
-    a.href = url;
-    a.download = "index.html"; // Set the file name for download
-    document.body.appendChild(a); // Append to the document
-    a.click(); // Programmatically click the anchor to trigger download
-
-    // Clean up by removing the anchor and revoking the Blob URL
-    document.body.removeChild(a);
-    URL.revokeObjectURL(url);
-  };
-
+  // Functions
   const reset = () => {
     setAppState(AppState.INITIAL);
-    setGeneratedCode("");
-    setReferenceImages([]);
-    setExecutionConsole([]);
-    setUpdateInstruction("");
-    setIsImportedFromCode(false);
-    setAppHistory([]);
-    setCurrentVersion(null);
     setShouldIncludeResultImage(false);
+    setUpdateInstruction("");
+    disableInSelectAndEditMode();
+    resetExecutionConsoles();
+
+    resetCommits();
+    resetHead();
+
+    // Inputs
+    setInputMode("image");
+    setReferenceImages([]);
+    setIsImportedFromCode(false);
   };
 
+  const regenerate = () => {
+    if (head === null) {
+      toast.error(
+        "No current version set. Please contact support via chat or Github."
+      );
+      throw new Error("Regenerate called with no head");
+    }
+
+    // Retrieve the previous command
+    const currentCommit = commits[head];
+    if (currentCommit.type !== "ai_create") {
+      toast.error("Only the first version can be regenerated.");
+      return;
+    }
+
+    // Re-run the create
+    doCreate(referenceImages, inputMode);
+  };
+
+  // Used when the user cancels the code generation
   const cancelCodeGeneration = () => {
     wsRef.current?.close?.(USER_CLOSE_WEB_SOCKET_CODE);
-    // make sure stop can correct the state even if the websocket is already closed
-    cancelCodeGenerationAndReset();
   };
 
-  const shouldDisablePreview = inputMode === "video";
-  const previewCode =
-    shouldDisablePreview && appState === AppState.CODING ? "" : generatedCode;
-
-  const cancelCodeGenerationAndReset = () => {
-    // When this is the first version, reset the entire app state
-    if (currentVersion === null) {
+  // Used for code generation failure as well
+  const cancelCodeGenerationAndReset = (commit: Commit) => {
+    // When the current commit is the first version, reset the entire app state
+    if (commit.type === "ai_create") {
       reset();
     } else {
-      // Otherwise, revert to the last version
-      setGeneratedCode(appHistory[currentVersion].code);
+      // Otherwise, remove current commit from commits
+      removeCommit(commit.hash);
+
+      // Revert to parent commit
+      const parentCommitHash = commit.parentHash;
+      if (parentCommitHash) {
+        setHead(parentCommitHash);
+      } else {
+        throw new Error("Parent commit not found");
+      }
+
       setAppState(AppState.CODE_READY);
     }
   };
 
-  function doGenerateCode(
-    params: CodeGenerationParams,
-    parentVersion: number | null
-  ) {
-    setExecutionConsole([]);
+  function doGenerateCode(params: CodeGenerationParams) {
+    // Reset the execution console
+    resetExecutionConsoles();
+
+    // Set the app state
     setAppState(AppState.CODING);
 
     // Merge settings with params
     const updatedParams = { ...params, ...settings };
 
+    const baseCommitObject = {
+      variants: [{ code: "" }, { code: "" }],
+    };
+
+    const commitInputObject =
+      params.generationType === "create"
+        ? {
+            ...baseCommitObject,
+            type: "ai_create" as const,
+            parentHash: null,
+            inputs: { image_url: referenceImages[0] },
+          }
+        : {
+            ...baseCommitObject,
+            type: "ai_edit" as const,
+            parentHash: head,
+            inputs: {
+              prompt: params.history
+                ? params.history[params.history.length - 1]
+                : "",
+            },
+          };
+
+    // Create a new commit and set it as the head
+    const commit = createCommit(commitInputObject);
+    addCommit(commit);
+    setHead(commit.hash);
+
     generateCode(
       wsRef,
       updatedParams,
       // On change
-      (token) => setGeneratedCode((prev) => prev + token),
+      (token, variantIndex) => {
+        appendCommitCode(commit.hash, variantIndex, token);
+      },
       // On set code
-      (code) => {
-        setGeneratedCode(code);
-        if (params.generationType === "create") {
-          setAppHistory([
-            {
-              type: "ai_create",
-              parentIndex: null,
-              code,
-              inputs: { image_url: referenceImages[0] },
-            },
-          ]);
-          setCurrentVersion(0);
-        } else {
-          setAppHistory((prev) => {
-            // Validate parent version
-            if (parentVersion === null) {
-              toast.error(
-                "No parent version set. Contact support or open a Github issue."
-              );
-              return prev;
-            }
-
-            const newHistory: History = [
-              ...prev,
-              {
-                type: "ai_edit",
-                parentIndex: parentVersion,
-                code,
-                inputs: {
-                  prompt: updateInstruction,
-                },
-              },
-            ];
-            setCurrentVersion(newHistory.length - 1);
-            return newHistory;
-          });
-        }
+      (code, variantIndex) => {
+        setCommitCode(commit.hash, variantIndex, code);
       },
       // On status update
-      (line) => setExecutionConsole((prev) => [...prev, line]),
+      (line, variantIndex) => appendExecutionConsole(variantIndex, line),
       // On cancel
       () => {
-        cancelCodeGenerationAndReset();
+        cancelCodeGenerationAndReset(commit);
       },
       // On complete
       () => {
@@ -229,68 +237,71 @@ function App() {
     // Reset any existing state
     reset();
 
+    // Set the input states
     setReferenceImages(referenceImages);
     setInputMode(inputMode);
+
+    // Kick off the code generation
     if (referenceImages.length > 0) {
-      doGenerateCode(
-        {
-          generationType: "create",
-          image: referenceImages[0],
-          inputMode,
-        },
-        currentVersion
-      );
+      doGenerateCode({
+        generationType: "create",
+        image: referenceImages[0],
+        inputMode,
+      });
     }
   }
 
   // Subsequent updates
-  async function doUpdate() {
-    if (currentVersion === null) {
+  async function doUpdate(
+    updateInstruction: string,
+    selectedElement?: HTMLElement
+  ) {
+    if (updateInstruction.trim() === "") {
+      toast.error("Please include some instructions for AI on what to update.");
+      return;
+    }
+
+    if (head === null) {
       toast.error(
         "No current version set. Contact support or open a Github issue."
       );
-      return;
+      throw new Error("Update called with no head");
     }
 
     let historyTree;
     try {
-      historyTree = extractHistoryTree(appHistory, currentVersion);
+      historyTree = extractHistory(head, commits);
     } catch {
       toast.error(
         "Version history is invalid. This shouldn't happen. Please contact support or open a Github issue."
       );
-      return;
+      throw new Error("Invalid version history");
     }
 
-    const updatedHistory = [...historyTree, updateInstruction];
+    let modifiedUpdateInstruction = updateInstruction;
 
-    if (shouldIncludeResultImage) {
-      const resultImage = await takeScreenshot();
-      doGenerateCode(
-        {
-          generationType: "update",
-          inputMode,
-          image: referenceImages[0],
-          resultImage: resultImage,
-          history: updatedHistory,
-          isImportedFromCode,
-        },
-        currentVersion
-      );
-    } else {
-      doGenerateCode(
-        {
-          generationType: "update",
-          inputMode,
-          image: referenceImages[0],
-          history: updatedHistory,
-          isImportedFromCode,
-        },
-        currentVersion
-      );
+    // Send in a reference to the selected element if it exists
+    if (selectedElement) {
+      modifiedUpdateInstruction =
+        updateInstruction +
+        " referring to this element specifically: " +
+        selectedElement.outerHTML;
     }
 
-    setGeneratedCode("");
+    const updatedHistory = [...historyTree, modifiedUpdateInstruction];
+    const resultImage = shouldIncludeResultImage
+      ? await takeScreenshot()
+      : undefined;
+
+    doGenerateCode({
+      generationType: "update",
+      inputMode,
+      image: referenceImages[0],
+      resultImage,
+      history: updatedHistory,
+      isImportedFromCode,
+    });
+
     setUpdateInstruction("");
   }
 
@@ -308,35 +319,30 @@ function App() {
     }));
   }
 
-  function setCodeGenerationModel(codeGenerationModel: CodeGenerationModel) {
-    setSettings((prev) => ({
-      ...prev,
-      codeGenerationModel,
-    }));
-  }
-
   function importFromCode(code: string, stack: Stack) {
+    // Set input state
     setIsImportedFromCode(true);
 
     // Set up this project
-    setGeneratedCode(code);
     setStack(stack);
-    setAppHistory([
-      {
-        type: "code_create",
-        parentIndex: null,
-        code,
-        inputs: { code },
-      },
-    ]);
-    setCurrentVersion(0);
 
+    // Create a new commit and set it as the head
+    const commit = createCommit({
+      type: "code_create",
+      parentHash: null,
+      variants: [{ code }],
+      inputs: null,
+    });
+    addCommit(commit);
+    setHead(commit.hash);
+
+    // Set the app state
     setAppState(AppState.CODE_READY);
   }
 
   return (
     <div className="mt-2 dark:bg-black dark:text-white">
-      {IS_RUNNING_ON_CLOUD && <PicoBadge settings={settings} />}
+      {IS_RUNNING_ON_CLOUD && <PicoBadge />}
       {IS_RUNNING_ON_CLOUD && (
         <TermsOfServiceDialog
           open={!settings.isTermOfServiceAccepted}
@@ -345,204 +351,47 @@ function App() {
       )}
       <div className="lg:fixed lg:inset-y-0 lg:z-40 lg:flex lg:w-96 lg:flex-col">
         <div className="flex grow flex-col gap-y-2 overflow-y-auto border-r border-gray-200 bg-white px-6 dark:bg-zinc-950 dark:text-white">
+          {/* Header with access to settings */}
           <div className="flex items-center justify-between mt-10 mb-2">
             <h1 className="text-2xl ">Screenshot to Code</h1>
             <SettingsDialog settings={settings} setSettings={setSettings} />
           </div>
 
-          <OutputSettingsSection
-            stack={settings.generatedCodeConfig}
-            setStack={(config) => setStack(config)}
-            shouldDisableUpdates={
-              appState === AppState.CODING || appState === AppState.CODE_READY
-            }
-          />
+          {/* Generation settings like stack and model */}
+          <GenerationSettings settings={settings} setSettings={setSettings} />
 
-          <ModelSettingsSection
-            codeGenerationModel={selectedCodeGenerationModel}
-            setCodeGenerationModel={setCodeGenerationModel}
-            shouldDisableUpdates={
-              appState === AppState.CODING || appState === AppState.CODE_READY
-            }
-          />
+          {/* Show auto updated message when older models are choosen */}
+          {showBetterModelMessage && <DeprecationMessage />}
 
-          {IS_RUNNING_ON_CLOUD &&
-            !(settings.openAiApiKey || settings.accessCode) && (
-              <OnboardingNote />
-            )}
+          {/* Show tip link until coding is complete */}
+          {appState !== AppState.CODE_READY && <TipLink />}
 
-          {IS_OPENAI_DOWN && (
-            <div className="bg-black text-white dark:bg-white dark:text-black p-3 rounded">
-              OpenAI API is currently down. Try back in 30 minutes or later. We
-              apologize for the inconvenience.
-            </div>
-          )}
+          {IS_RUNNING_ON_CLOUD && !settings.openAiApiKey && <OnboardingNote />}
 
+          {/* Rest of the sidebar when we're not in the initial state */}
           {(appState === AppState.CODING ||
             appState === AppState.CODE_READY) && (
-            <>
-              {/* Show code preview only when coding */}
-              {appState === AppState.CODING && (
-                <div className="flex flex-col">
-                  <div className="flex items-center gap-x-1">
-                    <Spinner />
-                    {executionConsole.slice(-1)[0]}
-                  </div>
-                  <div className="flex mt-4 w-full">
-                    <Button
-                      onClick={cancelCodeGeneration}
-                      className="w-full dark:text-white dark:bg-gray-700"
-                    >
-                      Cancel
-                    </Button>
-                  </div>
-                  <CodePreview code={generatedCode} />
-                </div>
-              )}
-
-              {appState === AppState.CODE_READY && (
-                <div>
-                  <div className="grid w-full gap-2">
-                    <Textarea
-                      placeholder="Tell the AI what to change..."
-                      onChange={(e) => setUpdateInstruction(e.target.value)}
-                      value={updateInstruction}
-                    />
-                    <div className="flex justify-between items-center gap-x-2">
-                      <div className="font-500 text-xs text-slate-700 dark:text-white">
-                        Include screenshot of current version?
-                      </div>
-                      <Switch
-                        checked={shouldIncludeResultImage}
-                        onCheckedChange={setShouldIncludeResultImage}
-                        className="dark:bg-gray-700"
-                      />
-                    </div>
-                    <Button
-                      onClick={doUpdate}
-                      className="dark:text-white dark:bg-gray-700"
-                    >
-                      Update
-                    </Button>
-                  </div>
-                  <div className="flex items-center gap-x-2 mt-2">
-                    <Button
-                      onClick={downloadCode}
-                      className="flex items-center gap-x-2 dark:text-white dark:bg-gray-700"
-                    >
-                      <FaDownload /> Download
-                    </Button>
-                    <Button
-                      onClick={reset}
-                      className="flex items-center gap-x-2 dark:text-white dark:bg-gray-700"
-                    >
-                      <FaUndo />
-                      Reset
-                    </Button>
-                  </div>
-                </div>
-              )}
-
-              {/* Reference image display */}
-              <div className="flex gap-x-2 mt-2">
-                {referenceImages.length > 0 && (
-                  <div className="flex flex-col">
-                    <div
-                      className={classNames({
-                        "scanning relative": appState === AppState.CODING,
-                      })}
-                    >
-                      <img
-                        className="w-[340px] border border-gray-200 rounded-md"
-                        src={referenceImages[0]}
-                        alt="Reference"
-                      />
-                    </div>
-                    <div className="text-gray-400 uppercase text-sm text-center mt-1">
-                      Original Screenshot
-                    </div>
-                  </div>
-                )}
-                <div className="bg-gray-400 px-4 py-2 rounded text-sm hidden">
-                  <h2 className="text-lg mb-4 border-b border-gray-800">
-                    Console
-                  </h2>
-                  {executionConsole.map((line, index) => (
-                    <div
-                      key={index}
-                      className="border-b border-gray-400 mb-2 text-gray-600 font-mono"
-                    >
-                      {line}
-                    </div>
-                  ))}
-                </div>
-              </div>
-            </>
-          )}
-          {
-            <HistoryDisplay
-              history={appHistory}
-              currentVersion={currentVersion}
-              revertToVersion={(index) => {
-                if (
-                  index < 0 ||
-                  index >= appHistory.length ||
-                  !appHistory[index]
-                )
-                  return;
-                setCurrentVersion(index);
-                setGeneratedCode(appHistory[index].code);
-              }}
-              shouldDisableReverts={appState === AppState.CODING}
+            <Sidebar
+              showSelectAndEditFeature={showSelectAndEditFeature}
+              doUpdate={doUpdate}
+              regenerate={regenerate}
+              cancelCodeGeneration={cancelCodeGeneration}
             />
-          }
+          )}
         </div>
       </div>
 
       <main className="py-2 lg:pl-96">
         {appState === AppState.INITIAL && (
-          <div className="flex flex-col justify-center items-center gap-y-10">
-            <ImageUpload setReferenceImages={doCreate} />
-            <UrlInputSection
-              doCreate={doCreate}
-              screenshotOneApiKey={settings.screenshotOneApiKey}
-            />
-            <ImportCodeSection importFromCode={importFromCode} />
-          </div>
+          <StartPane
+            doCreate={doCreate}
+            importFromCode={importFromCode}
+            settings={settings}
+          />
         )}
 
         {(appState === AppState.CODING || appState === AppState.CODE_READY) && (
-          <div className="ml-4">
-            <Tabs defaultValue="desktop">
-              <div className="flex justify-end mr-8 mb-4">
-                <TabsList>
-                  <TabsTrigger value="desktop" className="flex gap-x-2">
-                    <FaDesktop /> Desktop
-                  </TabsTrigger>
-                  <TabsTrigger value="mobile" className="flex gap-x-2">
-                    <FaMobile /> Mobile
-                  </TabsTrigger>
-                  <TabsTrigger value="code" className="flex gap-x-2">
-                    <FaCode />
-                    Code
-                  </TabsTrigger>
-                </TabsList>
-              </div>
-              <TabsContent value="desktop">
-                <Preview code={previewCode} device="desktop" />
-              </TabsContent>
-              <TabsContent value="mobile">
-                <Preview code={previewCode} device="mobile" />
-              </TabsContent>
-              <TabsContent value="code">
-                <CodeTab
-                  code={previewCode}
-                  setCode={setGeneratedCode}
-                  settings={settings}
-                />
-              </TabsContent>
-            </Tabs>
-          </div>
+          <PreviewPane doUpdate={doUpdate} reset={reset} settings={settings} />
         )}
       </main>
     </div>

frontend/src/components/ImageUpload.tsx

@@ -3,6 +3,10 @@ import { useState, useEffect, useMemo } from "react";
 import { useDropzone } from "react-dropzone";
 // import { PromptImage } from "../../../types";
 import { toast } from "react-hot-toast";
+import { URLS } from "../urls";
+import { Badge } from "./ui/badge";
+import ScreenRecorder from "./recording/ScreenRecorder";
+import { ScreenRecorderState } from "../types";
 
 const baseStyle = {
   flex: 1,
@@ -59,16 +63,23 @@ interface Props {
 
 function ImageUpload({ setReferenceImages }: Props) {
   const [files, setFiles] = useState<FileWithPreview[]>([]);
+  // TODO: Switch to Zustand
+  const [screenRecorderState, setScreenRecorderState] =
+    useState<ScreenRecorderState>(ScreenRecorderState.INITIAL);
+
   const { getRootProps, getInputProps, isFocused, isDragAccept, isDragReject } =
     useDropzone({
       maxFiles: 1,
       maxSize: 1024 * 1024 * 20, // 20 MB
       accept: {
+        // Image formats
         "image/png": [".png"],
         "image/jpeg": [".jpeg"],
         "image/jpg": [".jpg"],
+        // Video formats
         "video/quicktime": [".mov"],
         "video/mp4": [".mp4"],
+        "video/webm": [".webm"],
       },
       onDrop: (acceptedFiles) => {
         // Set up the preview thumbnail images
@@ -152,15 +163,34 @@ function ImageUpload({ setReferenceImages }: Props) {
 
   return (
     <section className="container">
-      {/* eslint-disable-next-line @typescript-eslint/no-explicit-any */}
-      <div {...getRootProps({ style: style as any })}>
-        <input {...getInputProps()} />
-        <p className="text-slate-700 text-lg">
-          Drag & drop a screenshot here, <br />
-          or paste from clipboard, <br />
-          or click to upload
-        </p>
-      </div>
+      {screenRecorderState === ScreenRecorderState.INITIAL && (
+        /* eslint-disable-next-line @typescript-eslint/no-explicit-any */
+        <div {...getRootProps({ style: style as any })}>
+          <input {...getInputProps()} className="file-input" />
+          <p className="text-slate-700 text-lg">
+            Drag & drop a screenshot here, <br />
+            or click to upload
+          </p>
+        </div>
+      )}
+      {screenRecorderState === ScreenRecorderState.INITIAL && (
+        <div className="text-center text-sm text-slate-800 mt-4">
+          <Badge>New!</Badge> Upload a screen recording (.mp4, .mov) or record
+          your screen to clone a whole app (experimental).{" "}
+          <a
+            className="underline"
+            href={URLS["intro-to-video"]}
+            target="_blank"
+          >
+            Learn more.
+          </a>
+        </div>
+      )}
+      <ScreenRecorder
+        screenRecorderState={screenRecorderState}
+        setScreenRecorderState={setScreenRecorderState}
+        generateCode={setReferenceImages}
+      />
     </section>
   );
 }

frontend/src/components/ImportCodeSection.tsx

@@ -10,7 +10,7 @@ import {
   DialogTrigger,
 } from "./ui/dialog";
 import { Textarea } from "./ui/textarea";
-import OutputSettingsSection from "./OutputSettingsSection";
+import OutputSettingsSection from "./settings/OutputSettingsSection";
 import toast from "react-hot-toast";
 import { Stack } from "../lib/stacks";
 
@@ -38,7 +38,9 @@ function ImportCodeSection({ importFromCode }: Props) {
   return (
     <Dialog>
       <DialogTrigger asChild>
-        <Button variant="secondary">Import from Code</Button>
+        <Button className="import-from-code-btn" variant="secondary">
+          Import from Code
+        </Button>
       </DialogTrigger>
       <DialogContent className="sm:max-w-[425px]">
         <DialogHeader>
@@ -62,7 +64,7 @@ function ImportCodeSection({ importFromCode }: Props) {
         />
 
         <DialogFooter>
-          <Button type="submit" onClick={doImport}>
+          <Button className="import-btn" type="submit" onClick={doImport}>
             Import
           </Button>
         </DialogFooter>

frontend/src/components/ModelSettingsSection.tsx

@@ -1,65 +0,0 @@
-import {
-  Select,
-  SelectContent,
-  SelectGroup,
-  SelectItem,
-  SelectTrigger,
-} from "./ui/select";
-import {
-  CODE_GENERATION_MODEL_DESCRIPTIONS,
-  CodeGenerationModel,
-} from "../lib/models";
-import { Badge } from "./ui/badge";
-
-interface Props {
-  codeGenerationModel: CodeGenerationModel;
-  setCodeGenerationModel: (codeGenerationModel: CodeGenerationModel) => void;
-  shouldDisableUpdates?: boolean;
-}
-
-function ModelSettingsSection({
-  codeGenerationModel,
-  setCodeGenerationModel,
-  shouldDisableUpdates = false,
-}: Props) {
-  return (
-    <div className="flex flex-col gap-y-2 justify-between text-sm">
-      <div className="grid grid-cols-3 items-center gap-4">
-        <span>Model:</span>
-        <Select
-          value={codeGenerationModel}
-          onValueChange={(value: string) =>
-            setCodeGenerationModel(value as CodeGenerationModel)
-          }
-          disabled={shouldDisableUpdates}
-        >
-          <SelectTrigger className="col-span-2" id="output-settings-js">
-            <span className="font-semibold">
-              {CODE_GENERATION_MODEL_DESCRIPTIONS[codeGenerationModel].name}
-            </span>
-          </SelectTrigger>
-          <SelectContent>
-            <SelectGroup>
-              {Object.values(CodeGenerationModel).map((model) => (
-                <SelectItem key={model} value={model}>
-                  <div className="flex items-center">
-                    <span className="font-semibold">
-                      {CODE_GENERATION_MODEL_DESCRIPTIONS[model].name}
-                    </span>
-                    {CODE_GENERATION_MODEL_DESCRIPTIONS[model].inBeta && (
-                      <Badge className="ml-2" variant="secondary">
-                        Beta
-                      </Badge>
-                    )}
-                  </div>
-                </SelectItem>
-              ))}
-            </SelectGroup>
-          </SelectContent>
-        </Select>
-      </div>
-    </div>
-  );
-}
-
-export default ModelSettingsSection;

frontend/src/components/PicoBadge.tsx

@@ -1,39 +0,0 @@
-import { Settings } from "../types";
-
-export function PicoBadge({ settings }: { settings: Settings }) {
-  return (
-    <>
-      <a
-        href="https://screenshot-to-code.canny.io/feature-requests"
-        target="_blank"
-      >
-        <div
-          className="fixed z-50 bottom-16 right-5 rounded-md shadow bg-black
-         text-white px-4 text-xs py-3 cursor-pointer"
-        >
-          feature requests?
-        </div>
-      </a>
-      {!settings.accessCode && (
-        <a href="https://picoapps.xyz?ref=screenshot-to-code" target="_blank">
-          <div
-            className="fixed z-50 bottom-5 right-5 rounded-md shadow text-black
-         bg-white px-4 text-xs py-3 cursor-pointer"
-          >
-            an open source project by Pico
-          </div>
-        </a>
-      )}
-      {settings.accessCode && (
-        <a href="mailto:support@picoapps.xyz" target="_blank">
-          <div
-            className="fixed z-50 bottom-5 right-5 rounded-md shadow text-black
-         bg-white px-4 text-xs py-3 cursor-pointer"
-          >
-            email support
-          </div>
-        </a>
-      )}
-    </>
-  );
-}

frontend/src/components/Preview.tsx

@@ -1,44 +0,0 @@
-import { useEffect, useRef } from "react";
-import classNames from "classnames";
-// import useThrottle from "../hooks/useThrottle";
-
-interface Props {
-  code: string;
-  device: "mobile" | "desktop";
-}
-
-function Preview({ code, device }: Props) {
-  const throttledCode = code;
-  // Temporary disable throttling for the preview not updating when the code changes
-  // useThrottle(code, 200);
-  const iframeRef = useRef<HTMLIFrameElement | null>(null);
-
-  useEffect(() => {
-    const iframe = iframeRef.current;
-    if (iframe && iframe.contentDocument) {
-      iframe.contentDocument.open();
-      iframe.contentDocument.write(throttledCode);
-      iframe.contentDocument.close();
-    }
-  }, [throttledCode]);
-
-  return (
-    <div className="flex justify-center mx-2">
-      <iframe
-        id={`preview-${device}`}
-        ref={iframeRef}
-        title="Preview"
-        className={classNames(
-          "border-[4px] border-black rounded-[20px] shadow-lg",
-          "transform scale-[0.9] origin-top",
-          {
-            "w-full h-[832px]": device === "desktop",
-            "w-[400px] h-[832px]": device === "mobile",
-          }
-        )}
-      ></iframe>
-    </div>
-  );
-}
-
-export default Preview;

frontend/src/components/UrlInputSection.tsx

@@ -69,7 +69,7 @@ export function UrlInputSection({ doCreate, screenshotOneApiKey }: Props) {
       <Button
         onClick={takeScreenshot}
         disabled={isLoading}
-        className="bg-slate-400"
+        className="bg-slate-400 capture-btn"
       >
         {isLoading ? "Capturing..." : "Capture"}
       </Button>

frontend/src/components/commits/types.ts

@@ -0,0 +1,37 @@
+export type CommitHash = string;
+
+export type Variant = {
+  code: string;
+};
+
+export type BaseCommit = {
+  hash: CommitHash;
+  parentHash: CommitHash | null;
+  dateCreated: Date;
+  isCommitted: boolean;
+  variants: Variant[];
+  selectedVariantIndex: number;
+};
+
+export type CommitType = "ai_create" | "ai_edit" | "code_create";
+
+export type AiCreateCommit = BaseCommit & {
+  type: "ai_create";
+  inputs: {
+    image_url: string;
+  };
+};
+
+export type AiEditCommit = BaseCommit & {
+  type: "ai_edit";
+  inputs: {
+    prompt: string;
+  };
+};
+
+export type CodeCreateCommit = BaseCommit & {
+  type: "code_create";
+  inputs: null;
+};
+
+export type Commit = AiCreateCommit | AiEditCommit | CodeCreateCommit;

frontend/src/components/commits/utils.ts

@@ -0,0 +1,32 @@
+import { nanoid } from "nanoid";
+import {
+  AiCreateCommit,
+  AiEditCommit,
+  CodeCreateCommit,
+  Commit,
+} from "./types";
+
+export function createCommit(
+  commit:
+    | Omit<
+        AiCreateCommit,
+        "hash" | "dateCreated" | "selectedVariantIndex" | "isCommitted"
+      >
+    | Omit<
+        AiEditCommit,
+        "hash" | "dateCreated" | "selectedVariantIndex" | "isCommitted"
+      >
+    | Omit<
+        CodeCreateCommit,
+        "hash" | "dateCreated" | "selectedVariantIndex" | "isCommitted"
+      >
+): Commit {
+  const hash = nanoid();
+  return {
+    ...commit,
+    hash,
+    isCommitted: false,
+    dateCreated: new Date(),
+    selectedVariantIndex: 0,
+  };
+}

frontend/src/components/core/KeyboardShortcutBadge.tsx

@@ -0,0 +1,25 @@
+import React from "react";
+import { BsArrowReturnLeft } from "react-icons/bs";
+
+interface KeyboardShortcutBadgeProps {
+  letter: string;
+}
+
+const KeyboardShortcutBadge: React.FC<KeyboardShortcutBadgeProps> = ({
+  letter,
+}) => {
+  const icon =
+    letter.toLowerCase() === "enter" || letter.toLowerCase() === "return" ? (
+      <BsArrowReturnLeft />
+    ) : (
+      letter.toUpperCase()
+    );
+
+  return (
+    <span className="font-mono text-xs ml-2 rounded bg-gray-700 dark:bg-gray-900 text-white py-[2px] px-2">
+      {icon}
+    </span>
+  );
+};
+
+export default KeyboardShortcutBadge;

frontend/src/components/evals/EvalsPage.tsx

@@ -4,7 +4,7 @@ import RatingPicker from "./RatingPicker";
 
 interface Eval {
   input: string;
-  output: string;
+  outputs: string[];
 }
 
 function EvalsPage() {
@@ -38,18 +38,22 @@ function EvalsPage() {
       <div className="flex flex-col gap-y-4 mt-4 mx-auto justify-center">
         {evals.map((e, index) => (
           <div className="flex flex-col justify-center" key={index}>
-            <div className="flex gap-x-2 justify-center">
+            <h2 className="font-bold text-lg ml-4">{index}</h2>
+            <div className="flex gap-x-2 justify-center ml-4">
+              {/* Update w if N changes to a fixed number like w-[600px] */}
               <div className="w-1/2 p-1 border">
-                <img src={e.input} />
-              </div>
-              <div className="w-1/2 p-1 border">
-                {/* Put output into an iframe */}
-                <iframe
-                  srcDoc={e.output}
-                  className="w-[1200px] h-[800px] transform scale-[0.60]"
-                  style={{ transformOrigin: "top left" }}
-                ></iframe>
+                <img src={e.input} alt={`Input for eval ${index}`} />
               </div>
+              {e.outputs.map((output, outputIndex) => (
+                <div className="w-1/2 p-1 border" key={outputIndex}>
+                  {/* Put output into an iframe */}
+                  <iframe
+                    srcDoc={output}
+                    className="w-[1200px] h-[800px] transform scale-[0.60]"
+                    style={{ transformOrigin: "top left" }}
+                  ></iframe>
+                </div>
+              ))}
             </div>
             <div className="ml-8 mt-4 flex justify-center">
               <RatingPicker

frontend/src/components/history/HistoryDisplay.tsx

@@ -1,4 +1,3 @@
-import { History } from "./history_types";
 import toast from "react-hot-toast";
 import classNames from "classnames";
 
@@ -11,21 +10,23 @@ import {
 } from "../ui/collapsible";
 import { Button } from "../ui/button";
 import { CaretSortIcon } from "@radix-ui/react-icons";
+import { useProjectStore } from "../../store/project-store";
 
 interface Props {
-  history: History;
-  currentVersion: number | null;
-  revertToVersion: (version: number) => void;
   shouldDisableReverts: boolean;
 }
 
-export default function HistoryDisplay({
-  history,
-  currentVersion,
-  revertToVersion,
-  shouldDisableReverts,
-}: Props) {
-  const renderedHistory = renderHistory(history, currentVersion);
+export default function HistoryDisplay({ shouldDisableReverts }: Props) {
+  const { commits, head, setHead } = useProjectStore();
+
+  // Put all commits into an array and sort by created date (oldest first)
+  const flatHistory = Object.values(commits).sort(
+    (a, b) =>
+      new Date(a.dateCreated).getTime() - new Date(b.dateCreated).getTime()
+  );
+
+  // Annotate history items with a summary, parent version, etc.
+  const renderedHistory = renderHistory(flatHistory);
 
   return renderedHistory.length === 0 ? null : (
     <div className="flex flex-col h-screen">
@@ -39,8 +40,8 @@ export default function HistoryDisplay({
                   "flex items-center justify-between space-x-2 w-full pr-2",
                   "border-b cursor-pointer",
                   {
-                    " hover:bg-black hover:text-white": !item.isActive,
-                    "bg-slate-500 text-white": item.isActive,
+                    " hover:bg-black hover:text-white": item.hash === head,
+                    "bg-slate-500 text-white": item.hash === head,
                   }
                 )}
               >
@@ -51,14 +52,14 @@ export default function HistoryDisplay({
                       ? toast.error(
                           "Please wait for code generation to complete before viewing an older version."
                         )
-                      : revertToVersion(index)
+                      : setHead(item.hash)
                   }
                 >
                   <div className="flex gap-x-1 truncate">
                     <h2 className="text-sm truncate">{item.summary}</h2>
                     {item.parentVersion !== null && (
                       <h2 className="text-sm">
-                        (parent: {item.parentVersion})
+                        (parent: v{item.parentVersion})
                       </h2>
                     )}
                   </div>

frontend/src/components/history/history_types.ts

@@ -1,41 +0,0 @@
-export type HistoryItemType = "ai_create" | "ai_edit" | "code_create";
-
-type CommonHistoryItem = {
-  parentIndex: null | number;
-  code: string;
-};
-
-export type HistoryItem =
-  | ({
-      type: "ai_create";
-      inputs: AiCreateInputs;
-    } & CommonHistoryItem)
-  | ({
-      type: "ai_edit";
-      inputs: AiEditInputs;
-    } & CommonHistoryItem)
-  | ({
-      type: "code_create";
-      inputs: CodeCreateInputs;
-    } & CommonHistoryItem);
-
-export type AiCreateInputs = {
-  image_url: string;
-};
-
-export type AiEditInputs = {
-  prompt: string;
-};
-
-export type CodeCreateInputs = {
-  code: string;
-};
-
-export type History = HistoryItem[];
-
-export type RenderedHistoryItem = {
-  type: string;
-  summary: string;
-  parentVersion: string | null;
-  isActive: boolean;
-};

frontend/src/components/history/utils.test.ts

@@ -1,230 +1,242 @@
-import { expect, test } from "vitest";
-import { extractHistoryTree, renderHistory } from "./utils";
-import type { History } from "./history_types";
+import { extractHistory, renderHistory } from "./utils";
+import { Commit, CommitHash } from "../commits/types";
 
-const basicLinearHistory: History = [
-  {
+const basicLinearHistory: Record<CommitHash, Commit> = {
+  "0": {
+    hash: "0",
+    dateCreated: new Date(),
+    isCommitted: false,
     type: "ai_create",
-    parentIndex: null,
-    code: "<html>1. create</html>",
+    parentHash: null,
+    variants: [{ code: "<html>1. create</html>" }],
+    selectedVariantIndex: 0,
     inputs: {
       image_url: "",
     },
   },
-  {
+  "1": {
+    hash: "1",
+    dateCreated: new Date(),
+    isCommitted: false,
     type: "ai_edit",
-    parentIndex: 0,
-    code: "<html>2. edit with better icons</html>",
+    parentHash: "0",
+    variants: [{ code: "<html>2. edit with better icons</html>" }],
+    selectedVariantIndex: 0,
     inputs: {
       prompt: "use better icons",
     },
   },
-  {
+  "2": {
+    hash: "2",
+    dateCreated: new Date(),
+    isCommitted: false,
     type: "ai_edit",
-    parentIndex: 1,
-    code: "<html>3. edit with better icons and red text</html>",
+    parentHash: "1",
+    variants: [{ code: "<html>3. edit with better icons and red text</html>" }],
+    selectedVariantIndex: 0,
     inputs: {
       prompt: "make text red",
     },
   },
-];
+};
 
-const basicLinearHistoryWithCode: History = [
-  {
+const basicLinearHistoryWithCode: Record<CommitHash, Commit> = {
+  "0": {
+    hash: "0",
+    dateCreated: new Date(),
+    isCommitted: false,
     type: "code_create",
-    parentIndex: null,
-    code: "<html>1. create</html>",
-    inputs: {
-      code: "<html>1. create</html>",
-    },
+    parentHash: null,
+    variants: [{ code: "<html>1. create</html>" }],
+    selectedVariantIndex: 0,
+    inputs: null,
   },
-  ...basicLinearHistory.slice(1),
-];
+  ...Object.fromEntries(Object.entries(basicLinearHistory).slice(1)),
+};
 
-const basicBranchingHistory: History = [
+const basicBranchingHistory: Record<CommitHash, Commit> = {
   ...basicLinearHistory,
-  {
+  "3": {
+    hash: "3",
+    dateCreated: new Date(),
+    isCommitted: false,
     type: "ai_edit",
-    parentIndex: 1,
-    code: "<html>4. edit with better icons and green text</html>",
+    parentHash: "1",
+    variants: [
+      { code: "<html>4. edit with better icons and green text</html>" },
+    ],
+    selectedVariantIndex: 0,
     inputs: {
       prompt: "make text green",
     },
   },
-];
+};
 
-const longerBranchingHistory: History = [
+const longerBranchingHistory: Record<CommitHash, Commit> = {
   ...basicBranchingHistory,
-  {
+  "4": {
+    hash: "4",
+    dateCreated: new Date(),
+    isCommitted: false,
     type: "ai_edit",
-    parentIndex: 3,
-    code: "<html>5. edit with better icons and green, bold text</html>",
+    parentHash: "3",
+    variants: [
+      { code: "<html>5. edit with better icons and green, bold text</html>" },
+    ],
+    selectedVariantIndex: 0,
     inputs: {
       prompt: "make text bold",
     },
   },
-];
+};
 
-const basicBadHistory: History = [
-  {
+const basicBadHistory: Record<CommitHash, Commit> = {
+  "0": {
+    hash: "0",
+    dateCreated: new Date(),
+    isCommitted: false,
     type: "ai_create",
-    parentIndex: null,
-    code: "<html>1. create</html>",
+    parentHash: null,
+    variants: [{ code: "<html>1. create</html>" }],
+    selectedVariantIndex: 0,
     inputs: {
       image_url: "",
     },
   },
-  {
+  "1": {
+    hash: "1",
+    dateCreated: new Date(),
+    isCommitted: false,
     type: "ai_edit",
-    parentIndex: 2, // <- Bad parent index
-    code: "<html>2. edit with better icons</html>",
+    parentHash: "2", // <- Bad parent hash
+    variants: [{ code: "<html>2. edit with better icons</html>" }],
+    selectedVariantIndex: 0,
     inputs: {
       prompt: "use better icons",
     },
   },
-];
+};
 
-test("should correctly extract the history tree", () => {
-  expect(extractHistoryTree(basicLinearHistory, 2)).toEqual([
-    "<html>1. create</html>",
-    "use better icons",
-    "<html>2. edit with better icons</html>",
-    "make text red",
-    "<html>3. edit with better icons and red text</html>",
-  ]);
+describe("History Utils", () => {
+  test("should correctly extract the history tree", () => {
+    expect(extractHistory("2", basicLinearHistory)).toEqual([
+      "<html>1. create</html>",
+      "use better icons",
+      "<html>2. edit with better icons</html>",
+      "make text red",
+      "<html>3. edit with better icons and red text</html>",
+    ]);
 
-  expect(extractHistoryTree(basicLinearHistory, 0)).toEqual([
-    "<html>1. create</html>",
-  ]);
+    expect(extractHistory("0", basicLinearHistory)).toEqual([
+      "<html>1. create</html>",
+    ]);
 
-  // Test branching
-  expect(extractHistoryTree(basicBranchingHistory, 3)).toEqual([
-    "<html>1. create</html>",
-    "use better icons",
-    "<html>2. edit with better icons</html>",
-    "make text green",
-    "<html>4. edit with better icons and green text</html>",
-  ]);
+    // Test branching
+    expect(extractHistory("3", basicBranchingHistory)).toEqual([
+      "<html>1. create</html>",
+      "use better icons",
+      "<html>2. edit with better icons</html>",
+      "make text green",
+      "<html>4. edit with better icons and green text</html>",
+    ]);
 
-  expect(extractHistoryTree(longerBranchingHistory, 4)).toEqual([
-    "<html>1. create</html>",
-    "use better icons",
-    "<html>2. edit with better icons</html>",
-    "make text green",
-    "<html>4. edit with better icons and green text</html>",
-    "make text bold",
-    "<html>5. edit with better icons and green, bold text</html>",
-  ]);
+    expect(extractHistory("4", longerBranchingHistory)).toEqual([
+      "<html>1. create</html>",
+      "use better icons",
+      "<html>2. edit with better icons</html>",
+      "make text green",
+      "<html>4. edit with better icons and green text</html>",
+      "make text bold",
+      "<html>5. edit with better icons and green, bold text</html>",
+    ]);
 
-  expect(extractHistoryTree(longerBranchingHistory, 2)).toEqual([
-    "<html>1. create</html>",
-    "use better icons",
-    "<html>2. edit with better icons</html>",
-    "make text red",
-    "<html>3. edit with better icons and red text</html>",
-  ]);
+    expect(extractHistory("2", longerBranchingHistory)).toEqual([
+      "<html>1. create</html>",
+      "use better icons",
+      "<html>2. edit with better icons</html>",
+      "make text red",
+      "<html>3. edit with better icons and red text</html>",
+    ]);
 
-  // Errors
+    // Errors
 
-  // Bad index
-  expect(() => extractHistoryTree(basicLinearHistory, 100)).toThrow();
-  expect(() => extractHistoryTree(basicLinearHistory, -2)).toThrow();
+    // Bad hash
+    expect(() => extractHistory("100", basicLinearHistory)).toThrow();
 
-  // Bad tree
-  expect(() => extractHistoryTree(basicBadHistory, 1)).toThrow();
-});
-
-test("should correctly render the history tree", () => {
-  expect(renderHistory(basicLinearHistory, 2)).toEqual([
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "Create",
-      type: "Create",
-    },
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "use better icons",
-      type: "Edit",
-    },
-    {
-      isActive: true,
-      parentVersion: null,
-      summary: "make text red",
-      type: "Edit",
-    },
-  ]);
-
-  // Current version is the first version
-  expect(renderHistory(basicLinearHistory, 0)).toEqual([
-    {
-      isActive: true,
-      parentVersion: null,
-      summary: "Create",
-      type: "Create",
-    },
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "use better icons",
-      type: "Edit",
-    },
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "make text red",
-      type: "Edit",
-    },
-  ]);
-
-  // Render a history with code
-  expect(renderHistory(basicLinearHistoryWithCode, 0)).toEqual([
-    {
-      isActive: true,
-      parentVersion: null,
-      summary: "Imported from code",
-      type: "Imported from code",
-    },
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "use better icons",
-      type: "Edit",
-    },
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "make text red",
-      type: "Edit",
-    },
-  ]);
-
-  // Render a non-linear history
-  expect(renderHistory(basicBranchingHistory, 3)).toEqual([
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "Create",
-      type: "Create",
-    },
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "use better icons",
-      type: "Edit",
-    },
-    {
-      isActive: false,
-      parentVersion: null,
-      summary: "make text red",
-      type: "Edit",
-    },
-    {
-      isActive: true,
-      parentVersion: "v2",
-      summary: "make text green",
-      type: "Edit",
-    },
-  ]);
+    // Bad tree
+    expect(() => extractHistory("1", basicBadHistory)).toThrow();
+  });
+
+  test("should correctly render the history tree", () => {
+    expect(renderHistory(Object.values(basicLinearHistory))).toEqual([
+      {
+        ...basicLinearHistory["0"],
+        type: "Create",
+        summary: "Create",
+        parentVersion: null,
+      },
+      {
+        ...basicLinearHistory["1"],
+        type: "Edit",
+        summary: "use better icons",
+        parentVersion: null,
+      },
+      {
+        ...basicLinearHistory["2"],
+        type: "Edit",
+        summary: "make text red",
+        parentVersion: null,
+      },
+    ]);
+
+    // Render a history with code
+    expect(renderHistory(Object.values(basicLinearHistoryWithCode))).toEqual([
+      {
+        ...basicLinearHistoryWithCode["0"],
+        type: "Imported from code",
+        summary: "Imported from code",
+        parentVersion: null,
+      },
+      {
+        ...basicLinearHistoryWithCode["1"],
+        type: "Edit",
+        summary: "use better icons",
+        parentVersion: null,
+      },
+      {
+        ...basicLinearHistoryWithCode["2"],
+        type: "Edit",
+        summary: "make text red",
+        parentVersion: null,
+      },
+    ]);
+
+    // Render a non-linear history
+    expect(renderHistory(Object.values(basicBranchingHistory))).toEqual([
+      {
+        ...basicBranchingHistory["0"],
+        type: "Create",
+        summary: "Create",
+        parentVersion: null,
+      },
+      {
+        ...basicBranchingHistory["1"],
+        type: "Edit",
+        summary: "use better icons",
+        parentVersion: null,
+      },
+      {
+        ...basicBranchingHistory["2"],
+        type: "Edit",
+        summary: "make text red",
+        parentVersion: null,
+      },
+      {
+        ...basicBranchingHistory["3"],
+        type: "Edit",
+        summary: "make text green",
+        parentVersion: 2,
+      },
+    ]);
+  });
 });

frontend/src/components/history/utils.ts

@@ -1,33 +1,25 @@
-import {
-  History,
-  HistoryItem,
-  HistoryItemType,
-  RenderedHistoryItem,
-} from "./history_types";
+import { Commit, CommitHash, CommitType } from "../commits/types";
 
-export function extractHistoryTree(
-  history: History,
-  version: number
+export function extractHistory(
+  hash: CommitHash,
+  commits: Record<CommitHash, Commit>
 ): string[] {
   const flatHistory: string[] = [];
 
-  let currentIndex: number | null = version;
-  while (currentIndex !== null) {
-    const item: HistoryItem = history[currentIndex];
+  let currentCommitHash: CommitHash | null = hash;
+  while (currentCommitHash !== null) {
+    const commit: Commit | null = commits[currentCommitHash];
 
-    if (item) {
-      if (item.type === "ai_create") {
-        // Don't include the image for ai_create
-        flatHistory.unshift(item.code);
-      } else if (item.type === "ai_edit") {
-        flatHistory.unshift(item.code);
-        flatHistory.unshift(item.inputs.prompt);
-      } else if (item.type === "code_create") {
-        flatHistory.unshift(item.code);
+    if (commit) {
+      flatHistory.unshift(commit.variants[commit.selectedVariantIndex].code);
+
+      // For edits, add the prompt to the history
+      if (commit.type === "ai_edit") {
+        flatHistory.unshift(commit.inputs.prompt);
       }
 
       // Move to the parent of the current item
-      currentIndex = item.parentIndex;
+      currentCommitHash = commit.parentHash;
     } else {
       throw new Error("Malformed history: missing parent index");
     }
@@ -36,7 +28,7 @@ export function extractHistoryTree(
   return flatHistory;
 }
 
-function displayHistoryItemType(itemType: HistoryItemType) {
+function displayHistoryItemType(itemType: CommitType) {
   switch (itemType) {
     case "ai_create":
       return "Create";
@@ -51,44 +43,48 @@ function displayHistoryItemType(itemType: HistoryItemType) {
   }
 }
 
-function summarizeHistoryItem(item: HistoryItem) {
-  const itemType = item.type;
-  switch (itemType) {
+const setParentVersion = (commit: Commit, history: Commit[]) => {
+  // If the commit has no parent, return null
+  if (!commit.parentHash) return null;
+
+  const parentIndex = history.findIndex(
+    (item) => item.hash === commit.parentHash
+  );
+  const currentIndex = history.findIndex((item) => item.hash === commit.hash);
+
+  // Only set parent version if the parent is not the previous commit
+  // and parent exists
+  return parentIndex !== -1 && parentIndex != currentIndex - 1
+    ? parentIndex + 1
+    : null;
+};
+
+export function summarizeHistoryItem(commit: Commit) {
+  const commitType = commit.type;
+  switch (commitType) {
     case "ai_create":
       return "Create";
     case "ai_edit":
-      return item.inputs.prompt;
+      return commit.inputs.prompt;
     case "code_create":
       return "Imported from code";
     default: {
-      const exhaustiveCheck: never = itemType;
+      const exhaustiveCheck: never = commitType;
       throw new Error(`Unhandled case: ${exhaustiveCheck}`);
     }
   }
 }
 
-export const renderHistory = (
-  history: History,
-  currentVersion: number | null
-) => {
-  const renderedHistory: RenderedHistoryItem[] = [];
+export const renderHistory = (history: Commit[]) => {
+  const renderedHistory = [];
 
   for (let i = 0; i < history.length; i++) {
-    const item = history[i];
-    // Only show the parent version if it's not the previous version
-    // (i.e. it's the branching point) and if it's not the first version
-    const parentVersion =
-      item.parentIndex !== null && item.parentIndex !== i - 1
-        ? `v${(item.parentIndex || 0) + 1}`
-        : null;
-    const type = displayHistoryItemType(item.type);
-    const isActive = i === currentVersion;
-    const summary = summarizeHistoryItem(item);
+    const commit = history[i];
     renderedHistory.push({
-      isActive,
-      summary: summary,
-      parentVersion,
-      type,
+      ...commit,
+      type: displayHistoryItemType(commit.type),
+      summary: summarizeHistoryItem(commit),
+      parentVersion: setParentVersion(commit, history),
     });
   }
 

frontend/src/components/messages/DeprecationMessage.tsx

@@ -0,0 +1,16 @@
+import React from "react";
+
+interface DeprecationMessageProps {}
+
+const DeprecationMessage: React.FC<DeprecationMessageProps> = () => {
+  return (
+    <div className="rounded-lg p-2 bg-fuchsia-200">
+      <p className="text-gray-800 text-sm">
+        We no longer support this model. Instead, code generation will use
+        GPT-4o or Claude Sonnet 3.5, the 2 state-of-the-art models.
+      </p>
+    </div>
+  );
+};
+
+export default DeprecationMessage;

frontend/src/components/messages/PicoBadge.tsx

@@ -0,0 +1,25 @@
+export function PicoBadge() {
+  return (
+    <>
+      <a
+        href="https://screenshot-to-code.canny.io/feature-requests"
+        target="_blank"
+      >
+        <div
+          className="fixed z-50 bottom-16 right-5 rounded-md shadow bg-black
+         text-white px-4 text-xs py-3 cursor-pointer"
+        >
+          feature requests?
+        </div>
+      </a>
+      <a href="https://picoapps.xyz?ref=screenshot-to-code" target="_blank">
+        <div
+          className="fixed z-50 bottom-5 right-5 rounded-md shadow text-black
+         bg-white px-4 text-xs py-3 cursor-pointer"
+        >
+          an open source project by Pico
+        </div>
+      </a>
+    </>
+  );
+}

frontend/src/components/messages/TipLink.tsx

@@ -0,0 +1,16 @@
+import { URLS } from "../../urls";
+
+function TipLink() {
+  return (
+    <a
+      className="text-xs underline text-gray-500 text-right"
+      href={URLS.tips}
+      target="_blank"
+      rel="noopener"
+    >
+      Tips for better results
+    </a>
+  );
+}
+
+export default TipLink;

frontend/src/components/preview/CodeTab.tsx

@@ -1,7 +1,7 @@
 import { FaCopy } from "react-icons/fa";
 import CodeMirror from "./CodeMirror";
-import { Button } from "./ui/button";
-import { Settings } from "../types";
+import { Button } from "../ui/button";
+import { Settings } from "../../types";
 import copy from "copy-to-clipboard";
 import { useCallback } from "react";
 import toast from "react-hot-toast";

frontend/src/components/preview/PreviewComponent.tsx

@@ -0,0 +1,56 @@
+import { useEffect, useRef, useState } from "react";
+import classNames from "classnames";
+import useThrottle from "../../hooks/useThrottle";
+import EditPopup from "../select-and-edit/EditPopup";
+
+interface Props {
+  code: string;
+  device: "mobile" | "desktop";
+  doUpdate: (updateInstruction: string, selectedElement?: HTMLElement) => void;
+}
+
+function PreviewComponent({ code, device, doUpdate }: Props) {
+  const iframeRef = useRef<HTMLIFrameElement | null>(null);
+
+  // Don't update code more often than every 200ms.
+  const throttledCode = useThrottle(code, 200);
+
+  // Select and edit functionality
+  const [clickEvent, setClickEvent] = useState<MouseEvent | null>(null);
+
+  useEffect(() => {
+    const iframe = iframeRef.current;
+    if (iframe) {
+      iframe.srcdoc = throttledCode;
+
+      // Set up click handler for select and edit funtionality
+      iframe.addEventListener("load", function () {
+        iframe.contentWindow?.document.body.addEventListener(
+          "click",
+          setClickEvent
+        );
+      });
+    }
+  }, [throttledCode]);
+
+  return (
+    <div className="flex justify-center mx-2">
+      <iframe
+        id={`preview-${device}`}
+        ref={iframeRef}
+        title="Preview"
+        className={classNames(
+          "border-[4px] border-black rounded-[20px] shadow-lg",
+          "transform scale-[0.9] origin-top",
+          {
+            "w-full h-[832px]": device === "desktop",
+            "w-[400px] h-[832px]": device === "mobile",
+          }
+        )}
+      ></iframe>
+      <EditPopup event={clickEvent} iframeRef={iframeRef} doUpdate={doUpdate} />
+    </div>
+  );
+}
+
+export default PreviewComponent;

frontend/src/components/preview/PreviewPane.tsx

@@ -0,0 +1,99 @@
+import { Tabs, TabsList, TabsTrigger, TabsContent } from "../ui/tabs";
+import {
+  FaUndo,
+  FaDownload,
+  FaDesktop,
+  FaMobile,
+  FaCode,
+} from "react-icons/fa";
+import { AppState, Settings } from "../../types";
+import CodeTab from "./CodeTab";
+import { Button } from "../ui/button";
+import { useAppStore } from "../../store/app-store";
+import { useProjectStore } from "../../store/project-store";
+import { extractHtml } from "./extractHtml";
+import PreviewComponent from "./PreviewComponent";
+import { downloadCode } from "./download";
+
+interface Props {
+  doUpdate: (instruction: string) => void;
+  reset: () => void;
+  settings: Settings;
+}
+
+function PreviewPane({ doUpdate, reset, settings }: Props) {
+  const { appState } = useAppStore();
+  const { inputMode, head, commits } = useProjectStore();
+
+  const currentCommit = head && commits[head] ? commits[head] : "";
+  const currentCode = currentCommit
+    ? currentCommit.variants[currentCommit.selectedVariantIndex].code
+    : "";
+
+  const previewCode =
+    inputMode === "video" && appState === AppState.CODING
+      ? extractHtml(currentCode)
+      : currentCode;
+
+  return (
+    <div className="ml-4">
+      <Tabs defaultValue="desktop">
+        <div className="flex justify-between mr-8 mb-4">
+          <div className="flex items-center gap-x-2">
+            {appState === AppState.CODE_READY && (
+              <>
+                <Button
+                  onClick={reset}
+                  className="flex items-center ml-4 gap-x-2 dark:text-white dark:bg-gray-700"
+                >
+                  <FaUndo />
+                  Reset
+                </Button>
+                <Button
+                  onClick={() => downloadCode(previewCode)}
+                  variant="secondary"
+                  className="flex items-center gap-x-2 mr-4 dark:text-white dark:bg-gray-700 download-btn"
+                >
+                  <FaDownload /> Download
+                </Button>
+              </>
+            )}
+          </div>
+          <div className="flex items-center">
+            <TabsList>
+              <TabsTrigger value="desktop" className="flex gap-x-2">
+                <FaDesktop /> Desktop
+              </TabsTrigger>
+              <TabsTrigger value="mobile" className="flex gap-x-2">
+                <FaMobile /> Mobile
+              </TabsTrigger>
+              <TabsTrigger value="code" className="flex gap-x-2">
+                <FaCode />
+                Code
+              </TabsTrigger>
+            </TabsList>
+          </div>
+        </div>
+        <TabsContent value="desktop">
+          <PreviewComponent
+            code={previewCode}
+            device="desktop"
+            doUpdate={doUpdate}
+          />
+        </TabsContent>
+        <TabsContent value="mobile">
+          <PreviewComponent
+            code={previewCode}
+            device="mobile"
+            doUpdate={doUpdate}
+          />
+        </TabsContent>
+        <TabsContent value="code">
+          <CodeTab code={previewCode} setCode={() => {}} settings={settings} />
+        </TabsContent>
+      </Tabs>
+    </div>
+  );
+}
+
+export default PreviewPane;

frontend/src/components/preview/download.ts

@@ -0,0 +1,16 @@
+export const downloadCode = (code: string) => {
+  // Create a blob from the generated code
+  const blob = new Blob([code], { type: "text/html" });
+  const url = URL.createObjectURL(blob);
+
+  // Create an anchor element and set properties for download
+  const a = document.createElement("a");
+  a.href = url;
+  a.download = "index.html"; // Set the file name for download
+  document.body.appendChild(a); // Append to the document
+  a.click(); // Programmatically click the anchor to trigger download
+
+  // Clean up by removing the anchor and revoking the Blob URL
+  document.body.removeChild(a);
+  URL.revokeObjectURL(url);
+};

frontend/src/components/preview/extractHtml.ts

@@ -0,0 +1,16 @@
+// Not robust enough to support <html lang='en'> for instance
+export function extractHtml(code: string): string {
+  const lastHtmlStartIndex = code.lastIndexOf("<html>");
+  let htmlEndIndex = code.indexOf("</html>", lastHtmlStartIndex);
+
+  if (lastHtmlStartIndex !== -1) {
+    // If "</html>" is found, adjust htmlEndIndex to include the "</html>" tag
+    if (htmlEndIndex !== -1) {
+      htmlEndIndex += "</html>".length;
+      return code.slice(lastHtmlStartIndex, htmlEndIndex);
+    }
+    // If "</html>" is not found, return the rest of the string starting from the last "<html>"
+    return code.slice(lastHtmlStartIndex);
+  }
+  return "";
+}

frontend/src/components/preview/simpleHash.ts

@@ -0,0 +1,10 @@
+
+export function simpleHash(str: string, seed = 0) {
+  let hash = seed;
+  for (let i = 0; i < str.length; i++) {
+    const char = str.charCodeAt(i);
+    hash = (hash << 5) - hash + char;
+    hash |= 0; // Convert to 32bit integer
+  }
+  return hash;
+}

frontend/src/components/recording/ScreenRecorder.tsx

@@ -0,0 +1,145 @@
+import { useState } from "react";
+import { Button } from "../ui/button";
+import { ScreenRecorderState } from "../../types";
+import { blobToBase64DataUrl } from "./utils";
+import fixWebmDuration from "webm-duration-fix";
+import toast from "react-hot-toast";
+
+interface Props {
+  screenRecorderState: ScreenRecorderState;
+  setScreenRecorderState: (state: ScreenRecorderState) => void;
+  generateCode: (
+    referenceImages: string[],
+    inputMode: "image" | "video"
+  ) => void;
+}
+
+function ScreenRecorder({
+  screenRecorderState,
+  setScreenRecorderState,
+  generateCode,
+}: Props) {
+  const [mediaStream, setMediaStream] = useState<MediaStream | null>(null);
+  const [mediaRecorder, setMediaRecorder] = useState<MediaRecorder | null>(
+    null
+  );
+  const [screenRecordingDataUrl, setScreenRecordingDataUrl] = useState<
+    string | null
+  >(null);
+
+  const startScreenRecording = async () => {
+    try {
+      // Get the screen recording stream
+      const stream = await navigator.mediaDevices.getDisplayMedia({
+        video: true,
+        audio: { echoCancellation: true },
+      });
+      setMediaStream(stream);
+
+      // TODO: Test across different browsers
+      // Create the media recorder
+      const options = { mimeType: "video/webm" };
+      const mediaRecorder = new MediaRecorder(stream, options);
+      setMediaRecorder(mediaRecorder);
+
+      const chunks: BlobPart[] = [];
+
+      // Accumalate chunks as data is available
+      mediaRecorder.ondataavailable = (e: BlobEvent) => chunks.push(e.data);
+
+      // When media recorder is stopped, create a data URL
+      mediaRecorder.onstop = async () => {
+        // TODO: Do I need to fix duration if it's not a webm?
+        const completeBlob = await fixWebmDuration(
+          new Blob(chunks, {
+            type: options.mimeType,
+          })
+        );
+
+        const dataUrl = await blobToBase64DataUrl(completeBlob);
+
+        setScreenRecordingDataUrl(dataUrl);
+        setScreenRecorderState(ScreenRecorderState.FINISHED);
+      };
+
+      // Start recording
+      mediaRecorder.start();
+      setScreenRecorderState(ScreenRecorderState.RECORDING);
+    } catch (error) {
+      toast.error("Could not start screen recording");
+      throw error;
+    }
+  };
+
+  const stopScreenRecording = () => {
+    // Stop the recorder
+    if (mediaRecorder) {
+      mediaRecorder.stop();
+      setMediaRecorder(null);
+    }
+
+    // Stop the screen sharing stream
+    if (mediaStream) {
+      mediaStream.getTracks().forEach((track) => {
+        track.stop();
+      });
+    }
+  };
+
+  const kickoffGeneration = () => {
+    if (screenRecordingDataUrl) {
+      generateCode([screenRecordingDataUrl], "video");
+    } else {
+      toast.error("Screen recording does not exist. Please try again.");
+      throw new Error("No screen recording data url");
+    }
+  };
+
+  return (
+    <div className="flex items-center justify-center my-3">
+      {screenRecorderState === ScreenRecorderState.INITIAL && (
+        <Button onClick={startScreenRecording}>Record Screen</Button>
+      )}
+
+      {screenRecorderState === ScreenRecorderState.RECORDING && (
+        <div className="flex items-center flex-col gap-y-4">
+          <div className="flex items-center mr-2 text-xl gap-x-1">
+            <span className="block h-10 w-10 bg-red-600 rounded-full mr-1 animate-pulse"></span>
+            <span>Recording...</span>
+          </div>
+          <Button onClick={stopScreenRecording}>Finish Recording</Button>
+        </div>
+      )}
+
+      {screenRecorderState === ScreenRecorderState.FINISHED && (
+        <div className="flex items-center flex-col gap-y-4">
+          <div className="flex items-center mr-2 text-xl gap-x-1">
+            <span>Screen Recording Captured.</span>
+          </div>
+          {screenRecordingDataUrl && (
+            <video
+              muted
+              autoPlay
+              loop
+              className="w-[340px] border border-gray-200 rounded-md"
+              src={screenRecordingDataUrl}
+            />
+          )}
+          <div className="flex gap-x-2">
+            <Button
+              variant="secondary"
+              onClick={() =>
+                setScreenRecorderState(ScreenRecorderState.INITIAL)
+              }
+            >
+              Re-record
+            </Button>
+            <Button onClick={kickoffGeneration}>Generate</Button>
+          </div>
+        </div>
+      )}
+    </div>
+  );
+}
+
+export default ScreenRecorder;

frontend/src/components/recording/utils.ts

@@ -0,0 +1,31 @@
+export function downloadBlob(blob: Blob) {
+  // Create a URL for the blob object
+  const videoURL = URL.createObjectURL(blob);
+
+  // Create a temporary anchor element and trigger the download
+  const a = document.createElement("a");
+  a.href = videoURL;
+  a.download = "recording.webm";
+  document.body.appendChild(a);
+  a.click();
+  document.body.removeChild(a);
+
+  // Clear object URL
+  URL.revokeObjectURL(videoURL);
+}
+
+export function blobToBase64DataUrl(blob: Blob): Promise<string> {
+  return new Promise((resolve, reject) => {
+    const reader = new FileReader();
+    reader.onloadend = () => {
+      if (reader.result) {
+        resolve(reader.result as string);
+      } else {
+        reject(new Error("FileReader did not return a result."));
+      }
+    };
+    reader.onerror = () =>
+      reject(new Error("FileReader encountered an error."));
+    reader.readAsDataURL(blob);
+  });
+}

frontend/src/components/select-and-edit/EditPopup.tsx

@@ -0,0 +1,150 @@
+import React, { useEffect, useRef, useState } from "react";
+import { Textarea } from "../ui/textarea";
+import { Button } from "../ui/button";
+import { addHighlight, getAdjustedCoordinates, removeHighlight } from "./utils";
+import { useAppStore } from "../../store/app-store";
+import KeyboardShortcutBadge from "../core/KeyboardShortcutBadge";
+
+interface EditPopupProps {
+  event: MouseEvent | null;
+  iframeRef: React.RefObject<HTMLIFrameElement>;
+  doUpdate: (updateInstruction: string, selectedElement?: HTMLElement) => void;
+}
+
+const EditPopup: React.FC<EditPopupProps> = ({
+  event,
+  iframeRef,
+  doUpdate,
+}) => {
+  // App state
+  const { inSelectAndEditMode } = useAppStore();
+
+  // Create a wrapper ref to store inSelectAndEditMode so the value is not stale
+  // in a event listener
+  const inSelectAndEditModeRef = useRef(inSelectAndEditMode);
+
+  // Update the ref whenever the state changes
+  useEffect(() => {
+    inSelectAndEditModeRef.current = inSelectAndEditMode;
+  }, [inSelectAndEditMode]);
+
+  // Popup state
+  const [popupVisible, setPopupVisible] = useState(false);
+  const [popupPosition, setPopupPosition] = useState({ x: 0, y: 0 });
+
+  // Edit state
+  const [selectedElement, setSelectedElement] = useState<
+    HTMLElement | undefined
+  >(undefined);
+  const [updateText, setUpdateText] = useState("");
+
+  // Textarea ref for focusing
+  const textareaRef = useRef<HTMLTextAreaElement | null>(null);
+
+  function onUpdate(updateText: string) {
+    // Perform the update
+    doUpdate(
+      updateText,
+      selectedElement ? removeHighlight(selectedElement) : selectedElement
+    );
+
+    // Unselect the element
+    setSelectedElement(undefined);
+
+    // Hide the popup
+    setPopupVisible(false);
+  }
+
+  // Remove highlight and reset state when not in select and edit mode
+  useEffect(() => {
+    if (!inSelectAndEditMode) {
+      if (selectedElement) removeHighlight(selectedElement);
+      setSelectedElement(undefined);
+      setPopupVisible(false);
+    }
+  }, [inSelectAndEditMode, selectedElement]);
+
+  // Handle the click event
+  useEffect(() => {
+    // Return if not in select and edit mode
+    if (!inSelectAndEditModeRef.current || !event) {
+      return;
+    }
+
+    // Prevent default to avoid issues like label clicks triggering textareas, etc.
+    event.preventDefault();
+
+    const targetElement = event.target as HTMLElement;
+
+    // Return if no target element
+    if (!targetElement) return;
+
+    // Highlight and set the selected element
+    setSelectedElement((prev) => {
+      // Remove style from previous element
+      if (prev) {
+        removeHighlight(prev);
+      }
+      return addHighlight(targetElement);
+    });
+
+    // Calculate adjusted coordinates
+    const adjustedCoordinates = getAdjustedCoordinates(
+      event.clientX,
+      event.clientY,
+      iframeRef.current?.getBoundingClientRect()
+    );
+
+    // Show the popup at the click position
+    setPopupVisible(true);
+    setPopupPosition({ x: adjustedCoordinates.x, y: adjustedCoordinates.y });
+
+    // Reset the update text
+    setUpdateText("");
+
+    // Focus the textarea
+    textareaRef.current?.focus();
+  }, [event, iframeRef]);
+
+  // Focus the textarea when the popup is visible (we can't do this only when handling the click event
+  // because the textarea is not rendered yet)
+  // We need to also do it in the click event because popupVisible doesn't change values in that event
+  useEffect(() => {
+    if (popupVisible) {
+      textareaRef.current?.focus();
+    }
+  }, [popupVisible]);
+
+  if (!popupVisible) return;
+
+  return (
+    <div
+      className="absolute bg-white dark:bg-gray-800 p-4 border border-gray-300 dark:border-gray-600 rounded shadow-lg w-60"
+      style={{ top: popupPosition.y, left: popupPosition.x }}
+    >
+      <Textarea
+        ref={textareaRef}
+        value={updateText}
+        onChange={(e) => setUpdateText(e.target.value)}
+        placeholder="Tell the AI what to change about this element..."
+        className="dark:bg-gray-700 dark:text-white"
+        onKeyDown={(e) => {
+          if (e.key === "Enter") {
+            e.preventDefault();
+            onUpdate(updateText);
+          }
+        }}
+      />
+      <div className="flex justify-end mt-2">
+        <Button
+          className="dark:bg-gray-700 dark:text-white"
+          onClick={() => onUpdate(updateText)}
+        >
+          Update <KeyboardShortcutBadge letter="enter" />
+        </Button>
+      </div>
+    </div>
+  );
+};
+
+export default EditPopup;

frontend/src/components/select-and-edit/SelectAndEditModeToggleButton.tsx

@@ -0,0 +1,22 @@
+import { GiClick } from "react-icons/gi";
+import { useAppStore } from "../../store/app-store";
+import { Button } from "../ui/button";
+
+function SelectAndEditModeToggleButton() {
+  const { inSelectAndEditMode, toggleInSelectAndEditMode } = useAppStore();
+
+  return (
+    <Button
+      onClick={toggleInSelectAndEditMode}
+      className="flex items-center gap-x-2 dark:text-white dark:bg-gray-700 regenerate-btn"
+      variant={inSelectAndEditMode ? "destructive" : "default"}
+    >
+      <GiClick className="text-lg" />
+      <span>
+        {inSelectAndEditMode ? "Exit selection mode" : "Select and update"}
+      </span>
+    </Button>
+  );
+}
+
+export default SelectAndEditModeToggleButton;

frontend/src/components/select-and-edit/utils.ts

@@ -0,0 +1,22 @@
+export function removeHighlight(element: HTMLElement) {
+  element.style.outline = "";
+  element.style.backgroundColor = "";
+  return element;
+}
+
+export function addHighlight(element: HTMLElement) {
+  element.style.outline = "2px dashed #1846db";
+  element.style.backgroundColor = "#bfcbf5";
+  return element;
+}
+
+export function getAdjustedCoordinates(
+  x: number,
+  y: number,
+  rect: DOMRect | undefined
+) {
+  const offsetX = rect ? rect.left : 0;
+  const offsetY = rect ? rect.top : 0;
+
+  return { x: x + offsetX, y: y + offsetY };
+}

frontend/src/components/settings/AccessCodeSection.tsx

@@ -1,142 +0,0 @@
-import { useEffect, useState } from "react";
-import { Settings } from "../../types";
-import { Button } from "../ui/button";
-import { Input } from "../ui/input";
-import { Label } from "../ui/label";
-import useThrottle from "../../hooks/useThrottle";
-import { Progress } from "../ui/progress";
-import { PICO_BACKEND_FORM_SECRET } from "../../config";
-
-interface Props {
-  settings: Settings;
-  setSettings: React.Dispatch<React.SetStateAction<Settings>>;
-}
-
-interface UsageResponse {
-  used_credits: number;
-  total_credits: number;
-  is_valid: boolean;
-}
-
-enum FetchState {
-  EMPTY = "EMPTY",
-  LOADING = "LOADING",
-  INVALID = "INVALID",
-  VALID = "VALID",
-}
-
-function AccessCodeSection({ settings, setSettings }: Props) {
-  const [isLoading, setIsLoading] = useState(false);
-  const [isValid, setIsValid] = useState(false);
-  const [usedCredits, setUsedCredits] = useState(0);
-  const [totalCredits, setTotalCredits] = useState(0);
-  const throttledAccessCode = useThrottle(settings.accessCode || "", 500);
-
-  const fetchState = (() => {
-    if (!settings.accessCode) return FetchState.EMPTY;
-    if (isLoading) return FetchState.LOADING;
-    if (!isValid) return FetchState.INVALID;
-    return FetchState.VALID;
-  })();
-
-  async function fetchUsage(accessCode: string) {
-    const res = await fetch(
-      "https://backend.buildpicoapps.com/screenshot_to_code/get_access_code_usage",
-      {
-        method: "POST",
-        headers: {
-          "Content-Type": "application/json",
-        },
-        body: JSON.stringify({
-          access_code: accessCode,
-          secret: PICO_BACKEND_FORM_SECRET,
-        }),
-      }
-    );
-    const usage = (await res.json()) as UsageResponse;
-
-    if (!usage.is_valid) {
-      setIsValid(false);
-    } else {
-      setIsValid(true);
-      setUsedCredits(usage.used_credits);
-      setTotalCredits(usage.total_credits);
-    }
-
-    setIsLoading(false);
-  }
-
-  useEffect(() => {
-    // Don't do anything if access code is empty
-    if (!throttledAccessCode) return;
-
-    setIsLoading(true);
-    setIsValid(true);
-
-    // Wait for 500 ms before fetching usage
-    setTimeout(async () => {
-      await fetchUsage(throttledAccessCode);
-    }, 500);
-  }, [throttledAccessCode]);
-
-  return (
-    <div className="flex flex-col space-y-4 bg-slate-200 p-4 rounded dark:text-white dark:bg-slate-800">
-      <Label htmlFor="access-code">
-        <div>Access Code</div>
-      </Label>
-
-      <Input
-        id="access-code"
-        className="border-gray-700 dark:border-gray-700 dark:bg-gray-800 dark:text-white"
-        placeholder="Enter your Screenshot to Code access code"
-        value={settings.accessCode || ""}
-        onChange={(e) =>
-          setSettings((s) => ({
-            ...s,
-            accessCode: e.target.value,
-          }))
-        }
-      />
-
-      {fetchState === "EMPTY" && (
-        <div className="flex items-center justify-between">
-          <a href="https://buy.stripe.com/8wM6sre70gBW1nqaEE" target="_blank">
-            <Button size="sm" variant="secondary">
-              Buy credits
-            </Button>
-          </a>
-        </div>
-      )}
-
-      {fetchState === "LOADING" && (
-        <div className="flex items-center justify-between">
-          <span className="text-xs text-gray-700">Loading...</span>
-        </div>
-      )}
-
-      {fetchState === "INVALID" && (
-        <>
-          <div className="flex items-center justify-between">
-            <span className="text-xs text-gray-700">Invalid access code</span>
-          </div>
-        </>
-      )}
-
-      {fetchState === "VALID" && (
-        <>
-          <Progress value={(usedCredits / totalCredits) * 100} />
-          <div className="flex items-center justify-between">
-            <span className="text-xs text-gray-700">
-              {usedCredits} out of {totalCredits} credits used
-            </span>
-            <a href="https://buy.stripe.com/8wM6sre70gBW1nqaEE" target="_blank">
-              <Button size="sm">Add credits</Button>
-            </a>
-          </div>
-        </>
-      )}
-    </div>
-  );
-}
-
-export default AccessCodeSection;

frontend/src/components/settings/GenerationSettings.tsx

@@ -0,0 +1,37 @@
+import React from "react";
+import { useAppStore } from "../../store/app-store";
+import { AppState, Settings } from "../../types";
+import OutputSettingsSection from "./OutputSettingsSection";
+import { Stack } from "../../lib/stacks";
+
+interface GenerationSettingsProps {
+  settings: Settings;
+  setSettings: React.Dispatch<React.SetStateAction<Settings>>;
+}
+
+export const GenerationSettings: React.FC<GenerationSettingsProps> = ({
+  settings,
+  setSettings,
+}) => {
+  const { appState } = useAppStore();
+
+  function setStack(stack: Stack) {
+    setSettings((prev: Settings) => ({
+      ...prev,
+      generatedCodeConfig: stack,
+    }));
+  }
+
+  const shouldDisableUpdates =
+    appState === AppState.CODING || appState === AppState.CODE_READY;
+
+  return (
+    <div className="flex flex-col gap-y-2">
+      <OutputSettingsSection
+        stack={settings.generatedCodeConfig}
+        setStack={setStack}
+        shouldDisableUpdates={shouldDisableUpdates}
+      />
+    </div>
+  );
+};

frontend/src/components/settings/OutputSettingsSection.tsx

@@ -5,9 +5,9 @@ import {
   SelectGroup,
   SelectItem,
   SelectTrigger,
-} from "./ui/select";
-import { Badge } from "./ui/badge";
-import { Stack, STACK_DESCRIPTIONS } from "../lib/stacks";
+} from "../ui/select";
+import { Badge } from "../ui/badge";
+import { Stack, STACK_DESCRIPTIONS } from "../../lib/stacks";
 
 function generateDisplayComponent(stack: Stack) {
   const stackComponents = STACK_DESCRIPTIONS[stack].components;

frontend/src/components/settings/SettingsDialog.tsx

@@ -9,20 +9,19 @@ import {
   DialogTrigger,
 } from "@/components/ui/dialog";
 import { FaCog } from "react-icons/fa";
-import { EditorTheme, Settings } from "../types";
-import { Switch } from "./ui/switch";
-import { Label } from "./ui/label";
-import { Input } from "./ui/input";
-import { Select, SelectContent, SelectItem, SelectTrigger } from "./ui/select";
-import { capitalize } from "../lib/utils";
-import { IS_RUNNING_ON_CLOUD } from "../config";
+import { EditorTheme, Settings } from "../../types";
+import { Switch } from "../ui/switch";
+import { Label } from "../ui/label";
+import { Input } from "../ui/input";
+import { Select, SelectContent, SelectItem, SelectTrigger } from "../ui/select";
+import { capitalize } from "../../lib/utils";
+import { IS_RUNNING_ON_CLOUD } from "../../config";
 import {
   Accordion,
   AccordionContent,
   AccordionItem,
   AccordionTrigger,
-} from "./ui/accordion";
-import AccessCodeSection from "./settings/AccessCodeSection";
+} from "../ui/accordion";
 
 interface Props {
   settings: Settings;
@@ -47,15 +46,10 @@ function SettingsDialog({ settings, setSettings }: Props) {
           <DialogTitle className="mb-4">Settings</DialogTitle>
         </DialogHeader>
 
-        {/* Access code */}
-        {IS_RUNNING_ON_CLOUD && (
-          <AccessCodeSection settings={settings} setSettings={setSettings} />
-        )}
-
         <div className="flex items-center space-x-2">
           <Label htmlFor="image-generation">
             <div>DALL-E Placeholder Image Generation</div>
-            <div className="font-light mt-2">
+            <div className="font-light mt-2 text-xs">
               More fun with it but if you want to save money, turn it off.
             </div>
           </Label>
@@ -70,29 +64,31 @@ function SettingsDialog({ settings, setSettings }: Props) {
             }
           />
         </div>
-        <div className="flex flex-col space-y-4">
-          <Label htmlFor="openai-api-key">
-            <div>OpenAI API key</div>
-            <div className="font-light mt-2 leading-relaxed">
-              Only stored in your browser. Never stored on servers. Overrides
-              your .env config.
-            </div>
-          </Label>
+        <div className="flex flex-col space-y-6">
+          <div>
+            <Label htmlFor="openai-api-key">
+              <div>OpenAI API key</div>
+              <div className="font-light mt-1 mb-2 text-xs leading-relaxed">
+                Only stored in your browser. Never stored on servers. Overrides
+                your .env config.
+              </div>
+            </Label>
 
-          <Input
-            id="openai-api-key"
-            placeholder="OpenAI API key"
-            value={settings.openAiApiKey || ""}
-            onChange={(e) =>
-              setSettings((s) => ({
-                ...s,
-                openAiApiKey: e.target.value,
-              }))
-            }
-          />
+            <Input
+              id="openai-api-key"
+              placeholder="OpenAI API key"
+              value={settings.openAiApiKey || ""}
+              onChange={(e) =>
+                setSettings((s) => ({
+                  ...s,
+                  openAiApiKey: e.target.value,
+                }))
+              }
+            />
+          </div>
 
           {!IS_RUNNING_ON_CLOUD && (
-            <>
+            <div>
               <Label htmlFor="openai-api-key">
                 <div>OpenAI Base URL (optional)</div>
                 <div className="font-light mt-2 leading-relaxed">
@@ -111,9 +107,31 @@ function SettingsDialog({ settings, setSettings }: Props) {
                   }))
                 }
               />
-            </>
+            </div>
           )}
 
+          <div>
+            <Label htmlFor="anthropic-api-key">
+              <div>Anthropic API key</div>
+              <div className="font-light mt-1 text-xs leading-relaxed">
+                Only stored in your browser. Never stored on servers. Overrides
+                your .env config.
+              </div>
+            </Label>
+
+            <Input
+              id="anthropic-api-key"
+              placeholder="Anthropic API key"
+              value={settings.anthropicApiKey || ""}
+              onChange={(e) =>
+                setSettings((s) => ({
+                  ...s,
+                  anthropicApiKey: e.target.value,
+                }))
+              }
+            />
+          </div>
+
           <Accordion type="single" collapsible className="w-full">
             <AccordionItem value="item-1">
               <AccordionTrigger>Screenshot by URL Config</AccordionTrigger>

frontend/src/components/sidebar/Sidebar.tsx

@@ -0,0 +1,196 @@
+import { Switch } from "@radix-ui/react-switch";
+import classNames from "classnames";
+import { useAppStore } from "../../store/app-store";
+import { useProjectStore } from "../../store/project-store";
+import { AppState } from "../../types";
+import CodePreview from "../preview/CodePreview";
+import Spinner from "../core/Spinner";
+import KeyboardShortcutBadge from "../core/KeyboardShortcutBadge";
+import TipLink from "../messages/TipLink";
+import SelectAndEditModeToggleButton from "../select-and-edit/SelectAndEditModeToggleButton";
+import { Button } from "../ui/button";
+import { Textarea } from "../ui/textarea";
+import { useEffect, useRef } from "react";
+import HistoryDisplay from "../history/HistoryDisplay";
+import Variants from "../variants/Variants";
+
+interface SidebarProps {
+  showSelectAndEditFeature: boolean;
+  doUpdate: (instruction: string) => void;
+  regenerate: () => void;
+  cancelCodeGeneration: () => void;
+}
+
+function Sidebar({
+  showSelectAndEditFeature,
+  doUpdate,
+  regenerate,
+  cancelCodeGeneration,
+}: SidebarProps) {
+  const textareaRef = useRef<HTMLTextAreaElement>(null);
+
+  const {
+    appState,
+    updateInstruction,
+    setUpdateInstruction,
+    shouldIncludeResultImage,
+    setShouldIncludeResultImage,
+  } = useAppStore();
+
+  const { inputMode, referenceImages, executionConsoles, head, commits } =
+    useProjectStore();
+
+  const viewedCode =
+    head && commits[head]
+      ? commits[head].variants[commits[head].selectedVariantIndex].code
+      : "";
+
+  const executionConsole =
+    (head && executionConsoles[commits[head].selectedVariantIndex]) || [];
+
+  // When coding is complete, focus on the update instruction textarea
+  useEffect(() => {
+    if (appState === AppState.CODE_READY && textareaRef.current) {
+      textareaRef.current.focus();
+    }
+  }, [appState]);
+
+  return (
+    <>
+      <Variants />
+
+      {/* Show code preview only when coding */}
+      {appState === AppState.CODING && (
+        <div className="flex flex-col">
+          {/* Speed disclaimer for video mode */}
+          {inputMode === "video" && (
+            <div
+              className="bg-yellow-100 border-l-4 border-yellow-500 text-yellow-700
+            p-2 text-xs mb-4 mt-1"
+            >
+              Code generation from videos can take 3-4 minutes. We do multiple
+              passes to get the best result. Please be patient.
+            </div>
+          )}
+
+          <div className="flex items-center gap-x-1">
+            <Spinner />
+            {executionConsole.slice(-1)[0]}
+          </div>
+
+          <CodePreview code={viewedCode} />
+
+          <div className="flex w-full">
+            <Button
+              onClick={cancelCodeGeneration}
+              className="w-full dark:text-white dark:bg-gray-700"
+            >
+              Cancel
+            </Button>
+          </div>
+        </div>
+      )}
+
+      {appState === AppState.CODE_READY && (
+        <div>
+          <div className="grid w-full gap-2">
+            <Textarea
+              ref={textareaRef}
+              placeholder="Tell the AI what to change..."
+              onChange={(e) => setUpdateInstruction(e.target.value)}
+              onKeyDown={(e) => {
+                if (e.key === "Enter" && !e.shiftKey) {
+                  doUpdate(updateInstruction);
+                }
+              }}
+              value={updateInstruction}
+            />
+            <div className="flex justify-between items-center gap-x-2">
+              <div className="font-500 text-xs text-slate-700 dark:text-white">
+                Include screenshot of current version?
+              </div>
+              <Switch
+                checked={shouldIncludeResultImage}
+                onCheckedChange={setShouldIncludeResultImage}
+                className="dark:bg-gray-700"
+              />
+            </div>
+            <Button
+              onClick={() => doUpdate(updateInstruction)}
+              className="dark:text-white dark:bg-gray-700 update-btn"
+            >
+              Update <KeyboardShortcutBadge letter="enter" />
+            </Button>
+          </div>
+          <div className="flex items-center justify-end gap-x-2 mt-2">
+            <Button
+              onClick={regenerate}
+              className="flex items-center gap-x-2 dark:text-white dark:bg-gray-700 regenerate-btn"
+            >
+              🔄 Regenerate
+            </Button>
+            {showSelectAndEditFeature && <SelectAndEditModeToggleButton />}
+          </div>
+          <div className="flex justify-end items-center mt-2">
+            <TipLink />
+          </div>
+        </div>
+      )}
+
+      {/* Reference image display */}
+      <div className="flex gap-x-2 mt-2">
+        {referenceImages.length > 0 && (
+          <div className="flex flex-col">
+            <div
+              className={classNames({
+                "scanning relative": appState === AppState.CODING,
+              })}
+            >
+              {inputMode === "image" && (
+                <img
+                  className="w-[340px] border border-gray-200 rounded-md"
+                  src={referenceImages[0]}
+                  alt="Reference"
+                />
+              )}
+              {inputMode === "video" && (
+                <video
+                  muted
+                  autoPlay
+                  loop
+                  className="w-[340px] border border-gray-200 rounded-md"
+                  src={referenceImages[0]}
+                />
+              )}
+            </div>
+            <div className="text-gray-400 uppercase text-sm text-center mt-1">
+              {inputMode === "video" ? "Original Video" : "Original Screenshot"}
+            </div>
+          </div>
+        )}
+        <div className="bg-gray-400 px-4 py-2 rounded text-sm hidden">
+          <h2 className="text-lg mb-4 border-b border-gray-800">Console</h2>
+          {Object.entries(executionConsoles).map(([index, lines]) => (
+            <div key={index}>
+              {lines.map((line, lineIndex) => (
+                <div
+                  key={`${index}-${lineIndex}`}
+                  className="border-b border-gray-400 mb-2 text-gray-600 font-mono"
+                >
+                  <span className="font-bold mr-2">{`${index}:${
+                    lineIndex + 1
+                  }`}</span>
+                  {line}
+                </div>
+              ))}
+            </div>
+          ))}
+        </div>
+      </div>
+
+      <HistoryDisplay shouldDisableReverts={appState === AppState.CODING} />
+    </>
+  );
+}
+
+export default Sidebar;

frontend/src/components/start-pane/StartPane.tsx

@@ -0,0 +1,27 @@
+import React from "react";
+import ImageUpload from "../ImageUpload";
+import { UrlInputSection } from "../UrlInputSection";
+import ImportCodeSection from "../ImportCodeSection";
+import { Settings } from "../../types";
+import { Stack } from "../../lib/stacks";
+
+interface Props {
+  doCreate: (images: string[], inputMode: "image" | "video") => void;
+  importFromCode: (code: string, stack: Stack) => void;
+  settings: Settings;
+}
+
+const StartPane: React.FC<Props> = ({ doCreate, importFromCode, settings }) => {
+  return (
+    <div className="flex flex-col justify-center items-center gap-y-10">
+      <ImageUpload setReferenceImages={doCreate} />
+      <UrlInputSection
+        doCreate={doCreate}
+        screenshotOneApiKey={settings.screenshotOneApiKey}
+      />
+      <ImportCodeSection importFromCode={importFromCode} />
+    </div>
+  );
+};
+
+export default StartPane;

frontend/src/components/variants/Variants.tsx

@@ -0,0 +1,42 @@
+import { useProjectStore } from "../../store/project-store";
+
+function Variants() {
+  const { inputMode, head, commits, updateSelectedVariantIndex } =
+    useProjectStore();
+
+  // If there is no head, don't show the variants
+  if (head === null) {
+    return null;
+  }
+
+  const commit = commits[head];
+  const variants = commit.variants;
+  const selectedVariantIndex = commit.selectedVariantIndex;
+
+  // If there is only one variant or the commit is already committed, don't show the variants
+  if (variants.length <= 1 || commit.isCommitted || inputMode === "video") {
+    return <div className="mt-2"></div>;
+  }
+
+  return (
+    <div className="mt-4 mb-4">
+      <div className="grid grid-cols-2 gap-2">
+        {variants.map((_, index) => (
+          <div
+            key={index}
+            className={`p-2 border rounded-md cursor-pointer ${
+              index === selectedVariantIndex
+                ? "bg-blue-100 dark:bg-blue-900"
+                : "bg-gray-50 dark:bg-gray-800 hover:bg-gray-100 dark:hover:bg-gray-700"
+            }`}
+            onClick={() => updateSelectedVariantIndex(head, index)}
+          >
+            <h3 className="font-medium mb-1">Option {index + 1}</h3>
+          </div>
+        ))}
+      </div>
+    </div>
+  );
+}
+
+export default Variants;

frontend/src/constants.ts

@@ -1 +1,3 @@
+//  WebSocket protocol (RFC 6455) allows for the use of custom close codes in the range 4000-4999
+export const APP_ERROR_WEB_SOCKET_CODE = 4332;
 export const USER_CLOSE_WEB_SOCKET_CODE = 4333;

frontend/src/generateCode.ts

@@ -1,6 +1,9 @@
 import toast from "react-hot-toast";
 import { WS_BACKEND_URL } from "./config";
-import { USER_CLOSE_WEB_SOCKET_CODE } from "./constants";
+import {
+  APP_ERROR_WEB_SOCKET_CODE,
+  USER_CLOSE_WEB_SOCKET_CODE,
+} from "./constants";
 import { FullGenerationSettings } from "./types";
 
 const ERROR_MESSAGE =
@@ -8,12 +11,18 @@ const ERROR_MESSAGE =
 
 const CANCEL_MESSAGE = "Code generation cancelled";
 
+type WebSocketResponse = {
+  type: "chunk" | "status" | "setCode" | "error";
+  value: string;
+  variantIndex: number;
+};
+
 export function generateCode(
   wsRef: React.MutableRefObject<WebSocket | null>,
   params: FullGenerationSettings,
-  onChange: (chunk: string) => void,
-  onSetCode: (code: string) => void,
-  onStatusUpdate: (status: string) => void,
+  onChange: (chunk: string, variantIndex: number) => void,
+  onSetCode: (code: string, variantIndex: number) => void,
+  onStatusUpdate: (status: string, variantIndex: number) => void,
   onCancel: () => void,
   onComplete: () => void
 ) {
@@ -28,13 +37,13 @@ export function generateCode(
   });
 
   ws.addEventListener("message", async (event: MessageEvent) => {
-    const response = JSON.parse(event.data);
+    const response = JSON.parse(event.data) as WebSocketResponse;
     if (response.type === "chunk") {
-      onChange(response.value);
+      onChange(response.value, response.variantIndex);
     } else if (response.type === "status") {
-      onStatusUpdate(response.value);
+      onStatusUpdate(response.value, response.variantIndex);
     } else if (response.type === "setCode") {
-      onSetCode(response.value);
+      onSetCode(response.value, response.variantIndex);
     } else if (response.type === "error") {
       console.error("Error generating code", response.value);
       toast.error(response.value);
@@ -46,9 +55,13 @@ export function generateCode(
     if (event.code === USER_CLOSE_WEB_SOCKET_CODE) {
       toast.success(CANCEL_MESSAGE);
       onCancel();
+    } else if (event.code === APP_ERROR_WEB_SOCKET_CODE) {
+      console.error("Known server error", event);
+      onCancel();
     } else if (event.code !== 1000) {
-      console.error("WebSocket error code", event);
+      console.error("Unknown server or connection error", event);
       toast.error(ERROR_MESSAGE);
+      onCancel();
     } else {
       onComplete();
     }

frontend/src/hooks/useBrowserTabIndicator.ts

@@ -0,0 +1,29 @@
+import { useEffect } from "react";
+
+const CODING_SETTINGS = {
+  title: "Coding...",
+  favicon: "/favicon/coding.png",
+};
+const DEFAULT_SETTINGS = {
+  title: "Screenshot to Code",
+  favicon: "/favicon/main.png",
+};
+
+const useBrowserTabIndicator = (isCoding: boolean) => {
+  useEffect(() => {
+    const settings = isCoding ? CODING_SETTINGS : DEFAULT_SETTINGS;
+
+    // Set favicon
+    const faviconEl = document.querySelector(
+      "link[rel='icon']"
+    ) as HTMLLinkElement | null;
+    if (faviconEl) {
+      faviconEl.href = settings.favicon;
+    }
+
+    // Set title
+    document.title = settings.title;
+  }, [isCoding]);
+};
+
+export default useBrowserTabIndicator;

frontend/src/lib/models.ts

@@ -1,12 +1,20 @@
 // Keep in sync with backend (llm.py)
+// Order here matches dropdown order
 export enum CodeGenerationModel {
+  CLAUDE_3_5_SONNET_2024_06_20 = "claude-3-5-sonnet-20240620",
+  GPT_4O_2024_05_13 = "gpt-4o-2024-05-13",
+  GPT_4_TURBO_2024_04_09 = "gpt-4-turbo-2024-04-09",
   GPT_4_VISION = "gpt_4_vision",
   CLAUDE_3_SONNET = "claude_3_sonnet",
 }
 
+// Will generate a static error if a model in the enum above is not in the descriptions
 export const CODE_GENERATION_MODEL_DESCRIPTIONS: {
   [key in CodeGenerationModel]: { name: string; inBeta: boolean };
 } = {
-  gpt_4_vision: { name: "GPT-4 Vision", inBeta: false },
-  claude_3_sonnet: { name: "Claude 3 Sonnet", inBeta: true },
+  "gpt-4o-2024-05-13": { name: "GPT-4o", inBeta: false },
+  "claude-3-5-sonnet-20240620": { name: "Claude 3.5 Sonnet", inBeta: false },
+  "gpt-4-turbo-2024-04-09": { name: "GPT-4 Turbo (deprecated)", inBeta: false },
+  gpt_4_vision: { name: "GPT-4 Vision (deprecated)", inBeta: false },
+  claude_3_sonnet: { name: "Claude 3 (deprecated)", inBeta: false },
 };

frontend/src/lib/stacks.ts

@@ -1,6 +1,8 @@
 // Keep in sync with backend (prompts/types.py)
+// Order here determines order in dropdown
 export enum Stack {
   HTML_TAILWIND = "html_tailwind",
+  HTML_CSS = "html_css",
   REACT_TAILWIND = "react_tailwind",
   BOOTSTRAP = "bootstrap",
   VUE_TAILWIND = "vue_tailwind",
@@ -11,6 +13,7 @@ export enum Stack {
 export const STACK_DESCRIPTIONS: {
   [key in Stack]: { components: string[]; inBeta: boolean };
 } = {
+  html_css: { components: ["HTML", "CSS"], inBeta: false },
   html_tailwind: { components: ["HTML", "Tailwind"], inBeta: false },
   react_tailwind: { components: ["React", "Tailwind"], inBeta: false },
   bootstrap: { components: ["Bootstrap"], inBeta: false },

frontend/src/lib/takeScreenshot.ts

@@ -0,0 +1,14 @@
+import html2canvas from "html2canvas";
+
+export const takeScreenshot = async (): Promise<string> => {
+  const iframeElement = document.querySelector(
+    "#preview-desktop"
+  ) as HTMLIFrameElement;
+  if (!iframeElement?.contentWindow?.document.body) {
+    return "";
+  }
+
+  const canvas = await html2canvas(iframeElement.contentWindow.document.body);
+  const png = canvas.toDataURL("image/png");
+  return png;
+};

frontend/src/setupTests.ts

@@ -0,0 +1,3 @@
+// So jest test runner can read env vars from .env file
+import { config } from "dotenv";
+config({ path: ".env.jest" });

frontend/src/store/app-store.ts

@@ -0,0 +1,36 @@
+import { create } from "zustand";
+import { AppState } from "../types";
+
+// Store for app-wide state
+interface AppStore {
+  appState: AppState;
+  setAppState: (state: AppState) => void;
+
+  // UI state
+  updateInstruction: string;
+  setUpdateInstruction: (instruction: string) => void;
+  shouldIncludeResultImage: boolean;
+  setShouldIncludeResultImage: (shouldInclude: boolean) => void;
+
+  inSelectAndEditMode: boolean;
+  toggleInSelectAndEditMode: () => void;
+  disableInSelectAndEditMode: () => void;
+}
+
+export const useAppStore = create<AppStore>((set) => ({
+  appState: AppState.INITIAL,
+  setAppState: (state: AppState) => set({ appState: state }),
+
+  // UI state
+  updateInstruction: "",
+  setUpdateInstruction: (instruction: string) =>
+    set({ updateInstruction: instruction }),
+  shouldIncludeResultImage: true,
+  setShouldIncludeResultImage: (shouldInclude: boolean) =>
+    set({ shouldIncludeResultImage: shouldInclude }),
+
+  inSelectAndEditMode: false,
+  toggleInSelectAndEditMode: () =>
+    set((state) => ({ inSelectAndEditMode: !state.inSelectAndEditMode })),
+  disableInSelectAndEditMode: () => set({ inSelectAndEditMode: false }),
+}));

frontend/src/store/project-store.ts

@@ -0,0 +1,149 @@
+import { create } from "zustand";
+import { Commit, CommitHash } from "../components/commits/types";
+
+// Store for app-wide state
+interface ProjectStore {
+  // Inputs
+  inputMode: "image" | "video";
+  setInputMode: (mode: "image" | "video") => void;
+  isImportedFromCode: boolean;
+  setIsImportedFromCode: (imported: boolean) => void;
+  referenceImages: string[];
+  setReferenceImages: (images: string[]) => void;
+
+  // Outputs
+  commits: Record<string, Commit>;
+  head: CommitHash | null;
+
+  addCommit: (commit: Commit) => void;
+  removeCommit: (hash: CommitHash) => void;
+  resetCommits: () => void;
+
+  appendCommitCode: (
+    hash: CommitHash,
+    numVariant: number,
+    code: string
+  ) => void;
+  setCommitCode: (hash: CommitHash, numVariant: number, code: string) => void;
+  updateSelectedVariantIndex: (hash: CommitHash, index: number) => void;
+
+  setHead: (hash: CommitHash) => void;
+  resetHead: () => void;
+
+  executionConsoles: { [key: number]: string[] };
+  appendExecutionConsole: (variantIndex: number, line: string) => void;
+  resetExecutionConsoles: () => void;
+}
+
+export const useProjectStore = create<ProjectStore>((set) => ({
+  // Inputs and their setters
+  inputMode: "image",
+  setInputMode: (mode) => set({ inputMode: mode }),
+  isImportedFromCode: false,
+  setIsImportedFromCode: (imported) => set({ isImportedFromCode: imported }),
+  referenceImages: [],
+  setReferenceImages: (images) => set({ referenceImages: images }),
+
+  // Outputs
+  commits: {},
+  head: null,
+
+  addCommit: (commit: Commit) => {
+    // When adding a new commit, make sure all existing commits are marked as committed
+    set((state) => ({
+      commits: {
+        ...Object.fromEntries(
+          Object.entries(state.commits).map(([hash, existingCommit]) => [
+            hash,
+            { ...existingCommit, isCommitted: true },
+          ])
+        ),
+        [commit.hash]: commit,
+      },
+    }));
+  },
+  removeCommit: (hash: CommitHash) => {
+    set((state) => {
+      const newCommits = { ...state.commits };
+      delete newCommits[hash];
+      return { commits: newCommits };
+    });
+  },
+  resetCommits: () => set({ commits: {} }),
+
+  appendCommitCode: (hash: CommitHash, numVariant: number, code: string) =>
+    set((state) => {
+      const commit = state.commits[hash];
+      // Don't update if the commit is already committed
+      if (commit.isCommitted) {
+        throw new Error("Attempted to append code to a committed commit");
+      }
+      return {
+        commits: {
+          ...state.commits,
+          [hash]: {
+            ...commit,
+            variants: commit.variants.map((variant, index) =>
+              index === numVariant
+                ? { ...variant, code: variant.code + code }
+                : variant
+            ),
+          },
+        },
+      };
+    }),
+  setCommitCode: (hash: CommitHash, numVariant: number, code: string) =>
+    set((state) => {
+      const commit = state.commits[hash];
+      // Don't update if the commit is already committed
+      if (commit.isCommitted) {
+        throw new Error("Attempted to set code of a committed commit");
+      }
+      return {
+        commits: {
+          ...state.commits,
+          [hash]: {
+            ...commit,
+            variants: commit.variants.map((variant, index) =>
+              index === numVariant ? { ...variant, code } : variant
+            ),
+          },
+        },
+      };
+    }),
+  updateSelectedVariantIndex: (hash: CommitHash, index: number) =>
+    set((state) => {
+      const commit = state.commits[hash];
+      // Don't update if the commit is already committed
+      if (commit.isCommitted) {
+        throw new Error(
+          "Attempted to update selected variant index of a committed commit"
+        );
+      }
+      return {
+        commits: {
+          ...state.commits,
+          [hash]: {
+            ...commit,
+            selectedVariantIndex: index,
+          },
+        },
+      };
+    }),
+
+  setHead: (hash: CommitHash) => set({ head: hash }),
+  resetHead: () => set({ head: null }),
+
+  executionConsoles: {},
+  appendExecutionConsole: (variantIndex: number, line: string) =>
+    set((state) => ({
+      executionConsoles: {
+        ...state.executionConsoles,
+        [variantIndex]: [
+          ...(state.executionConsoles[variantIndex] || []),
+          line,
+        ],
+      },
+    })),
+  resetExecutionConsoles: () => set({ executionConsoles: {} }),
+}));