Update app.py

app.py

@@ -168,24 +168,48 @@ async def get_transcript(youtube_url: str):
                 
                 # Add format-specific parsing
                 if subtitle_file.endswith('.json3'):
-                    import json
-                    subs = json.loads(content)
-                    text = ' '.join([e['segs'][0]['utf8'] for e in subs['events'] if e.get('segs')])
-                elif subtitle_file.endswith('.vtt'):
-                    text = ' '.join(line.strip() for line in content.split('\n') 
-                           if not line.startswith('WEBVTT') 
-                           and '-->' not in line 
-                           and not line.strip().isdigit())
-                else:
-                    text = f"Unsupported format: {subtitle_file}"
-                
-                return {"transcript": text}
-            
-            return {"transcript": f"No subtitle files found for {video_id}"}
-
-    except Exception as e:
-        logger.error(f"Error: {str(e)}", exc_info=True)
-        raise HTTPException(status_code=500, detail=str(e))
+                import json
+                subs = json.loads(content)
+                # Extract text segments and clean duplicates
+                segments = []
+                seen = set()
+                for event in subs['events']:
+                    if 'segs' in event and event['segs']:
+                        text = event['segs'][0]['utf8'].strip()
+                        if text and text not in seen:
+                            segments.append(text)
+                            seen.add(text)
+                transcript = ' '.join(segments)
+    
+            elif subtitle_file.endswith('.vtt'):
+                # Parse VTT format
+                transcript = []
+                current_text = ''
+                for line in content.split('\n'):
+                    if '-->' in line or line.strip().isdigit() or line.startswith('WEBVTT'):
+                        if current_text:
+                            transcript.append(current_text.strip())
+                            current_text = ''
+                        continue
+                    if line.strip() and not line.startswith('NOTE'):
+                        current_text += ' ' + line.strip()
+                transcript = ' '.join(list(dict.fromkeys(transcript)))  # Remove duplicates while preserving order
+    
+            else:
+                transcript = "Unsupported subtitle format"
+    
+            # Post-process formatting
+            cleaned_transcript = (
+                transcript.replace("  ", " ")  # Remove double spaces
+                .replace("hi ", "")  # Remove residual VTT artifacts
+                .replace("Kind: captions Language: en", "")
+                .strip()
+            )
+    
+            return {"transcript": cleaned_transcript}
+    
+        except Exception as e:
+            raise HTTPException(status_code=500, detail=str(e))