main.py

from fastapi import FastAPI, HTTPException
from fastapi.middleware.cors import CORSMiddleware
from youtube_transcript_api import YouTubeTranscriptApi, TranscriptsDisabled, NoTranscriptFound, VideoUnavailable
import re

app = FastAPI()

# Enable CORS so your Vercel frontend can call this API
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"], # In production, you can change "*" to your Vercel URL
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

def extract_video_id(url: str) -> str:
    pattern = r'(?:youtube\.com\/(?:[^\/]+\/.+\/|(?:v|e(?:mbed)?)\/|.*[?&]v=)|youtu\.be\/)([^"&?\/\s]{11})'
    match = re.match(pattern, url)
    if match:
        return match.group(1)
    return None

@app.post("/api/transcript")
async def get_transcript(request: dict):
    url = request.get("url")
    if not url:
        raise HTTPException(status_code=400, detail="Missing URL")

    video_id = extract_video_id(url)
    if not video_id:
        raise HTTPException(status_code=400, detail="Invalid YouTube URL")

    try:
        # Fetch available transcripts
        transcript_list = YouTubeTranscriptApi.list_transcripts(video_id)
        
        # Find the best transcript (Prefer manual English, then auto English, then any other)
        transcript = None
        try:
            transcript = transcript_list.find_manually_created_transcript(['en'])
        except NoTranscriptFound:
            try:
                transcript = transcript_list.find_generated_transcript(['en'])
            except NoTranscriptFound:
                # If no English, get the first available language
                for t in transcript_list:
                    transcript = t
                    break
        
        if not transcript:
            raise HTTPException(status_code=404, detail="No captions found for this video.")

        # Fetch the actual text data
        transcript_data = transcript.fetch()
        
        # Combine all text lines
        full_text = " ".join([line['text'] for line in transcript_data])
        
        # Clean up HTML entities
        full_text = full_text.replace("'", "'").replace(""", '"').replace("&", "&").replace("'", "'")
        
        return {"text": full_text}

    except TranscriptsDisabled:
        raise HTTPException(status_code=403, detail="Captions are disabled for this video.")
    except VideoUnavailable:
        raise HTTPException(status_code=404, detail="Video is unavailable.")
    except Exception as e:
        raise HTTPException(status_code=500, detail=f"Failed to fetch transcript: {str(e)}")

# ========================================================
# KEEP-ALIVE ENDPOINT
# ========================================================
@app.get("/api/health")
async def health_check():
    return {"status": "ok", "message": "Transcript API is running!"}