Update app.py

app.py

@@ -63,22 +63,60 @@
 
 import gradio as gr
 import google.generativeai as genai
-import whisper
+import speech_recognition as sr
+from pydub import AudioSegment
 import os
 import tempfile
+import moviepy.editor as mp
 
-
+# Initialize Gemini API
 genai.configure(api_key="AIzaSyBPQF0g5EfEPzEiGRzA3iNzJZK4jDukMvE")
 
+# Initialize the Gemini model
 model = genai.GenerativeModel('gemini-pro')
 
-whisper_model = whisper.load_model("base")
-
 def transcribe_video(video_path):
     """Transcribe the audio from a video file."""
     try:
-        result = whisper_model.transcribe(video_path)
-        return result["text"]
+        # Extract audio from video
+        video = mp.VideoFileClip(video_path)
+        audio_path = tempfile.mktemp(suffix=".wav")
+        video.audio.write_audiofile(audio_path, codec='pcm_s16le')
+        
+        # Load audio file
+        audio = AudioSegment.from_wav(audio_path)
+        
+        # Initialize recognizer
+        r = sr.Recognizer()
+        
+        # Split audio into chunks to handle long audio
+        chunk_length_ms = 30000  # 30 seconds
+        chunks = [audio[i:i+chunk_length_ms] for i in range(0, len(audio), chunk_length_ms)]
+        
+        transcript = ""
+        for i, chunk in enumerate(chunks):
+            # Export chunk to a temporary file
+            chunk_path = tempfile.mktemp(suffix=".wav")
+            chunk.export(chunk_path, format="wav")
+            
+            # Recognize speech in the chunk
+            with sr.AudioFile(chunk_path) as source:
+                audio_listened = r.record(source)
+                try:
+                    text = r.recognize_google(audio_listened)
+                    transcript += text + " "
+                except sr.UnknownValueError:
+                    print(f"Could not understand audio in chunk {i+1}")
+                except sr.RequestError:
+                    print(f"Could not request results from Google Speech Recognition service for chunk {i+1}")
+            
+            # Clean up temporary chunk file
+            os.remove(chunk_path)
+        
+        # Clean up temporary audio file
+        os.remove(audio_path)
+        
+        return transcript.strip()
     except Exception as e:
         return f"Error transcribing video: {str(e)}"
 
@@ -119,20 +157,17 @@ def process_video(video, num_questions):
     with gr.Row():
         gr.Markdown("Processing video and generating summary and quiz...")
     
-    
     with tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") as temp_video:
         video.save(temp_video.name)
         video_path = temp_video.name
-   
-    transcript = transcribe_video(video_path)
 
+    transcript = transcribe_video(video_path)
     result = generate_summary_and_quiz(transcript, num_questions)
 
     os.unlink(video_path)
 
     return transcript, result
 
-
 with gr.Blocks() as demo:
     gr.Markdown("# Video Lecture Summarizer and Quiz Generator")