WIP

app.py

@@ -10,9 +10,6 @@ import os
 import json
 from pathlib import Path
 
-# Configure loguru
-logger.add("app.log", rotation="500 MB", level="DEBUG")
-
 MODEL_NAME = "muhtasham/whisper-tg"
 
 def format_time(seconds):
@@ -129,53 +126,9 @@ def transcribe(inputs, return_timestamps, generate_subs, batch_size, chunk_lengt
         logger.exception(f"Error during transcription: {str(e)}")
         raise gr.Error(f"Failed to transcribe audio: {str(e)}")
 
-# Create a custom flagging callback
-class TranscriptionFlaggingCallback(gr.FlaggingCallback):
-    def __init__(self, flagging_dir):
-        self.flagging_dir = Path(flagging_dir)
-        self.flagging_dir.mkdir(exist_ok=True)
-        self.log_file = self.flagging_dir / "flagged_data.jsonl"
-        
-    def setup(self, components, flagging_dir):
-        pass
-        
-    def flag(self, components, flag_data, flag_option, username):
-        try:
-            # Create a unique filename for the audio file
-            audio_file = components[0]  # First component is the audio input
-            if audio_file:
-                audio_filename = os.path.basename(audio_file)
-                # Copy audio file to flagged directory
-                audio_dir = self.flagging_dir / "audio"
-                audio_dir.mkdir(exist_ok=True)
-                import shutil
-                shutil.copy2(audio_file, audio_dir / audio_filename)
-            else:
-                audio_filename = None
-                
-            # Prepare the data to save
-            data = {
-                "timestamp": datetime.datetime.now().isoformat(),
-                "audio_file": audio_filename,
-                "transcription": components[1],  # JSON output
-                "correction": components[2] if len(components) > 2 else None,  # Correction text if provided
-                "username": username
-            }
-            
-            # Append to JSONL file
-            with open(self.log_file, "a", encoding="utf-8") as f:
-                f.write(json.dumps(data) + "\n")
-                
-            logger.info(f"Saved flagged data: {data}")
-        except Exception as e:
-            logger.error(f"Error while flagging: {str(e)}")
-            raise gr.Error(f"Failed to save feedback: {str(e)}")
 
 demo = gr.Blocks(theme=gr.themes.Ocean())
 
-# Create flagging callback
-flagging_callback = TranscriptionFlaggingCallback("flagged_data")
-
 # Define interfaces first
 mf_transcribe = gr.Interface(
     fn=transcribe,
@@ -195,9 +148,7 @@ mf_transcribe = gr.Interface(
         "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the"
         f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
         " of arbitrary length."
-    ),
-    flagging_mode="manual",
-    flagging_dir="flagged_data"
+    )
 )
 
 file_transcribe = gr.Interface(
@@ -218,63 +169,12 @@ file_transcribe = gr.Interface(
         "Transcribe long-form microphone or audio inputs with the click of a button! Demo uses the"
         f" checkpoint [{MODEL_NAME}](https://huggingface.co/{MODEL_NAME}) and 🤗 Transformers to transcribe audio files"
         " of arbitrary length."
-    ),
-    flagging_mode="manual",
-    flagging_dir="flagged_data"
+    )
 )
 
 # Then set up the demo with the interfaces
 with demo:
-    with gr.Tabs() as tabs:
-        with gr.Tab("Audio file"):
-            file_transcribe.render()
-        with gr.Tab("Microphone"):
-            mf_transcribe.render()
-        with gr.Tab("Feedback"):
-            with gr.Row():
-                with gr.Column():
-                    gr.Markdown("### Provide Feedback")
-                    gr.Markdown("If you notice any issues with the transcription, please provide the correct text below.")
-                    feedback_text = gr.Textbox(
-                        label="Correct transcription",
-                        placeholder="Enter the correct transcription here...",
-                        lines=5
-                    )
-                    submit_btn = gr.Button("Submit Feedback")
-                with gr.Column():
-                    gr.Markdown("### Instructions")
-                    gr.Markdown("""
-                    1. Transcribe your audio in the Audio file or Microphone tab
-                    2. If you notice any issues, copy the transcription here
-                    3. Edit the text to provide the correct version
-                    4. Click Submit Feedback
-                    """)
-            
-            def submit_feedback(text):
-                if not text.strip():
-                    raise gr.Error("Please provide the correct transcription.")
-                try:
-                    # Get the current outputs from either interface
-                    file_outputs = file_transcribe.output_components
-                    mic_outputs = mf_transcribe.output_components
-                    
-                    # Save the feedback
-                    flagging_callback.flag(
-                        components=[None, file_outputs[0], text],  # No audio file, just transcription and correction
-                        flag_data=None,
-                        flag_option=None,
-                        username=None
-                    )
-                    return "Thank you for your feedback!"
-                except Exception as e:
-                    logger.error(f"Error submitting feedback: {str(e)}")
-                    raise gr.Error(f"Failed to save feedback: {str(e)}")
-            
-            submit_btn.click(
-                submit_feedback,
-                inputs=[feedback_text],
-                outputs=[gr.Textbox(label="Status")]
-            )
+    gr.TabbedInterface([file_transcribe, mf_transcribe], ["Audio file", "Microphone"])
 
 logger.info("Starting Gradio interface")
 demo.queue().launch(ssr_mode=False)