Update app.py

app.py

@@ -26,6 +26,8 @@ from PIL import Image
 import base64
 from googleapiclient.discovery import build
 from googleapiclient.errors import HttpError
+import assemblyai as aai
+
 
 # LangChain & LangGraph
 from langgraph.graph.message import add_messages
@@ -682,93 +684,27 @@ class YoutubeInput(BaseModel):
 @tool(args_schema=YoutubeInput)
 def get_youtube_transcript(video_url: str) -> str:
     """
-    Fetches YouTube video transcript/captions using YouTube Data API v3.
-    Much more reliable than yt-dlp on cloud environments.
+    Fetches YouTube video transcript using AssemblyAI.
+    Works reliably on Hugging Face Spaces.
     """
-    if not video_url:
-        return "Error: Invalid URL."
-    
-    print(f"📺 YouTube transcript (API v3): {video_url}")
-    
-    # Get API key
-    YOUTUBE_API_KEY = os.getenv("YOUTUBE_API_KEY")
-    if not YOUTUBE_API_KEY:
-        return "Error: YOUTUBE_API_KEY not set in Space secrets."
-    
     try:
-        # Extract video ID
-        video_id = None
-        if "watch?v=" in video_url:
-            video_id = video_url.split("v=")[1].split("&")[0]
-        elif "youtu.be/" in video_url:
-            video_id = video_url.split("youtu.be/")[1].split("?")[0]
-        
-        if not video_id:
-            return "Error: Could not extract video ID from URL."
-        
-        print(f"   Video ID: {video_id}")
-        
-        # Initialize YouTube API
-        youtube = build('youtube', 'v3', developerKey=YOUTUBE_API_KEY)
-        
-        # Get caption tracks
-        captions_response = youtube.captions().list(
-            part='snippet',
-            videoId=video_id
-        ).execute()
-        
-        if not captions_response.get('items'):
-            return "N/A - No captions available for this video."
-        
-        # Find English caption track
-        caption_id = None
-        for caption in captions_response['items']:
-            lang = caption['snippet'].get('language', '')
-            if lang.startswith('en'):
-                caption_id = caption['id']
-                print(f"   Found English captions: {lang}")
-                break
+        # Set API key (store in HF Spaces secrets)
+        aai.settings.api_key = os.getenv("ASSEMBLYAI_API_KEY")
+        
+        print(f"📺 Transcribing: {video_url}")
+        
+        # Transcribe directly from YouTube URL
+        transcriber = aai.Transcriber()
+        transcript = transcriber.transcribe(video_url)
+        
+        # Wait for transcription
+        if transcript.status == aai.TranscriptStatus.error:
+            return f"Error: {transcript.error}"
+        
+        print(f"✓ Transcribed {len(transcript.text)} chars")
+        return f"Transcript:\n{transcript.text}"
         
-        if not caption_id:
-            # Try first available caption
-            caption_id = captions_response['items'][0]['id']
-            print(f"   Using first available caption track")
-        
-        # Download caption content
-        caption_content = youtube.captions().download(
-            id=caption_id,
-            tfmt='srt'  # or 'vtt'
-        ).execute()
-        
-        # Parse SRT format (remove timestamps and numbers)
-        lines = caption_content.decode('utf-8').split('\n')
-        transcript_parts = []
-        
-        for line in lines:
-            line = line.strip()
-            # Skip line numbers, timestamps, and empty lines
-            if (line and 
-                not line.isdigit() and 
-                '-->' not in line):
-                transcript_parts.append(line)
-        
-        full_transcript = ' '.join(transcript_parts)
-        
-        if not full_transcript:
-            return "Error: Transcript was empty."
-        
-        print(f"✓ Transcript retrieved: {len(full_transcript)} chars")
-        return f"Transcript:\n{truncate_if_needed(full_transcript)}"
-        
-    except HttpError as e:
-        if e.resp.status == 403:
-            return "Error: YouTube API quota exceeded or captions are disabled for this video."
-        elif e.resp.status == 404:
-            return "Error: Video not found or captions not available."
-        else:
-            return f"YouTube API error: {str(e)}"
     except Exception as e:
-        print(f"❌ Error: {str(e)}")
         return f"Error: {str(e)}"