Update app.py

app.py

@@ -383,7 +383,6 @@ def process_youtube_video(url="", keywords=""):
         summary = ""
         sentiment_label = "N/A"
         recommendations = ""
-        text = ""
 
         if not url.strip():
             return thumbnail, "Please enter a YouTube URL", sentiment_label, recommendations
@@ -394,45 +393,41 @@ def process_youtube_video(url="", keywords=""):
 
         thumbnail = f"https://img.youtube.com/vi/{video_id}/maxresdefault.jpg"
 
-        # Method 1: Direct transcript fetch
+        # Multiple methods to get transcript
+        text = ""
+        error_messages = []
+
+        # Method 1: YouTube Transcript API
         try:
-            transcript = YouTubeTranscriptApi.get_transcript(video_id, languages=['en'])
+            transcript = YouTubeTranscriptApi.get_transcript(video_id)
             text = " ".join([t['text'] for t in transcript])
-        except:
-            # Method 2: List available transcripts
+        except Exception as e:
+            error_messages.append(str(e))
+
+        # Method 2: PyTube if first method fails
+        if not text:
             try:
-                transcript_list = YouTubeTranscriptApi.list_transcripts(video_id)
-                transcript = None
-
-                # Try English variants
-                for lang_code in ['en', 'en-US', 'en-GB', 'a.en']:
-                    try:
-                        transcript = transcript_list.find_transcript([lang_code])
-                        if transcript:
-                            text = " ".join([t['text'] for t in transcript.fetch()])
-                            break
-                    except:
-                        continue
-
-                # Try auto-generated if no manual transcript
-                if not text:
-                    try:
-                        transcript = transcript_list.find_generated_transcript(['en'])
-                        text = " ".join([t['text'] for t in transcript.fetch()])
-                    except:
-                        # Try translation as last resort
-                        try:
-                            manual_transcript = transcript_list.find_manually_created_transcript()
-                            translated = manual_transcript.translate('en')
-                            text = " ".join([t['text'] for t in translated.fetch()])
-                        except:
-                            raise Exception("No available transcripts found")
+                yt = YouTube(url)
+                captions = yt.captions
+                if 'en' in captions:
+                    text = captions['en'].generate_srt_captions()
+                elif 'a.en' in captions:
+                    text = captions['a.en'].generate_srt_captions()
+            except Exception as e:
+                error_messages.append(str(e))
 
+        # Method 3: Try auto-generated captions
+        if not text:
+            try:
+                transcript_list = YouTubeTranscriptApi.list_transcripts(video_id)
+                auto_transcript = transcript_list.find_generated_transcript(['en'])
+                text = " ".join([t['text'] for t in auto_transcript.fetch()])
             except Exception as e:
-                return thumbnail, f"⚠️ No transcripts available: {str(e)}", sentiment_label, recommendations
+                error_messages.append(str(e))
 
         if not text:
-            return thumbnail, "⚠️ Could not extract transcript text", sentiment_label, recommendations
+            error_msg = "\n".join(error_messages)
+            return thumbnail, f"⚠️ Could not access video content. Please try another video with English subtitles.\nDetails: {error_msg}", sentiment_label, recommendations
 
         # Process valid transcript
         try:
@@ -451,9 +446,16 @@ def process_youtube_video(url="", keywords=""):
                 f"Subjectivity: {subjectivity:.2f}"
             )
 
-            # Generate summary
+            # Generate summary using Gemini
             model = genai.GenerativeModel("gemini-pro")
-            summary = model.generate_content(f"Summarize this content: {cleaned_text[:4000]}").text
+            prompt = f"""Provide a comprehensive summary of this content in clear points:
+            {cleaned_text[:4000]}
+            Include:
+            1. Main topics
+            2. Key points
+            3. Important takeaways"""
+            
+            summary = model.generate_content(prompt).text
 
         except Exception as e:
             return thumbnail, f"⚠️ Error processing content: {str(e)}", sentiment_label, recommendations
@@ -468,15 +470,6 @@ def process_youtube_video(url="", keywords=""):
         print(f"Debug - Main Error: {str(e)}")
         return None, f"Error: {str(e)}", "N/A", ""
 
-
-
-
-
-
-
-
-
-
 def get_recommendations(keywords, max_results=5):
     if not keywords:
         return "Please provide search keywords"