Github/screenshot-to-code
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

support best of n evals

Browse Source
This commit is contained in:
Abi Raja 2024-04-24 14:54:03 -04:00
parent f9c4dd9c7c
commit a5fe0960d8
3 changed files with 47 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
backend/routes/evals.py
View File

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

14
backend/run_evals.py
View File

@ -14,7 +14,8 @@ from evals.core import generate_code_core
from evals.utils import image_to_data_url from evals.utils import image_to_data_url
STACK = "html_tailwind" STACK = "html_tailwind"
MODEL = Llm.CLAUDE_3_SONNET MODEL = Llm.GPT_4_TURBO_2024_04_09
N = 1 # Number of outputs to generate
async def main(): async def main():
@ -28,6 +29,7 @@ async def main():
for filename in evals: for filename in evals:
filepath = os.path.join(INPUT_DIR, filename) filepath = os.path.join(INPUT_DIR, filename)
data_url = await image_to_data_url(filepath) data_url = await image_to_data_url(filepath)
for _ in range(N): # Generate N tasks for each input
task = generate_code_core(image_url=data_url, stack=STACK, model=MODEL) task = generate_code_core(image_url=data_url, stack=STACK, model=MODEL)
tasks.append(task) tasks.append(task)
@ -35,9 +37,13 @@ async def main():
os.makedirs(OUTPUT_DIR, exist_ok=True) os.makedirs(OUTPUT_DIR, exist_ok=True)
for filename, content in zip(evals, results): for i, content in enumerate(results):
# File name is derived from the original filename in evals # Calculate index for filename and output number
output_filename = f"{os.path.splitext(filename)[0]}.html" 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) output_filepath = os.path.join(OUTPUT_DIR, output_filename)
with open(output_filepath, "w") as file: with open(output_filepath, "w") as file:
file.write(content) file.write(content)

14
frontend/src/components/evals/EvalsPage.tsx
View File

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