Update app.py

app.py

@@ -2,6 +2,7 @@
 import gradio as gr
 from transformers import pipeline
 import torch
+import ffmpeg
 
 # --- Model Loading ---
 # We load the model once when the app starts, not on every function call.
@@ -21,6 +22,33 @@ transcriber = pipeline(
 )
 
 # --- Transcription Function ---
+def extract_audio_from_video(video_path, output_audio_path="temp_extracted_audio.wav"):
+    """
+    Extracts audio from a video file using python-ffmpeg.
+    """
+    print(f"\n[STEP 1/9] Extracting audio from video: {video_path}")
+    
+    try:
+        (
+            ffmpeg
+            .input(video_path)
+            .output(
+                output_audio_path,
+                vn=None,           # Disable video
+                acodec='mp3',      # Audio codec
+                ab='192k',         # Audio bitrate
+                ar='44100',        # Sample rate
+                ac=2,              # Audio channels
+                f='wav'            # Output format
+            )
+            .run(overwrite_output=True, quiet=True)
+        )
+        print(f"✅ Audio extracted successfully to: {output_audio_path}")
+        return output_audio_path
+    except ffmpeg.Error as e:
+        print(f"Error: Failed to extract audio from video. stderr: {e.stderr.decode('utf8')}")
+        return None
+
 def transcribe_audio(audio_path):
     """
     This function takes an audio file path, transcribes it using the Whisper model,
@@ -41,6 +69,11 @@ def transcribe_audio(audio_path):
         print(f"An error occurred during transcription: {e}")
         return f"Sorry, an error occurred. Please try again. Details: {str(e)}"
 
+def main_run(video_path):
+    original_audio_file = extract_audio_from_video(video_path)
+    original_text = transcribe_audio(original_audio_file)
+    return original_text
+    
 # --- Gradio Interface Definition ---
 # Title and description for the new Space
 title = "Custom Whisper Transcription App"
@@ -55,11 +88,9 @@ article = "<p style='text-align: center'><a href='https://huggingface.co/openai/
 # Create the Gradio interface with our custom function
 # We define the input as an Audio component and the output as a Textbox
 app_interface = gr.Interface(
-    fn=transcribe_audio,
-    inputs=gr.Audio(
-        sources=["microphone", "upload"], 
-        type="filepath", 
-        label="Upload Audio or Record"
+    fn=main_run,
+    inputs=gr.video(
+        label="Upload Video"
     ),
     outputs=gr.Textbox(label="Transcription Result"),
     title=title,