WIP

app.py

@@ -5,6 +5,8 @@ from transformers import pipeline
 import subprocess
 from loguru import logger
 import datetime
+import tempfile
+import os
 
 # Configure loguru
 logger.add("app.log", rotation="500 MB", level="DEBUG")
@@ -30,6 +32,17 @@ def generate_srt(chunks):
         srt_content.append(f"{i}\n{start_time} --> {end_time}\n{text}\n\n")
     return "".join(srt_content)
 
+def save_srt_to_file(srt_content):
+    """Save SRT content to a temporary file and return the file path"""
+    if not srt_content:
+        return None
+    
+    # Create a temporary file with .srt extension
+    temp_file = tempfile.NamedTemporaryFile(suffix='.srt', delete=False)
+    temp_file.write(srt_content.encode('utf-8'))
+    temp_file.close()
+    return temp_file.name
+
 # Check if ffmpeg is installed
 def check_ffmpeg():
     try:
@@ -102,22 +115,20 @@ def transcribe(inputs, return_timestamps, generate_subs, batch_size, chunk_lengt
             logger.info(f"Successfully processed transcription with {len(chunks)} chunks")
         
         # Generate subtitles if requested
-        srt_content = None
+        srt_file = None
         if generate_subs and chunks:
             logger.info("Generating SRT subtitles")
             srt_content = generate_srt(chunks)
+            srt_file = save_srt_to_file(srt_content)
             logger.info("SRT subtitles generated successfully")
         
-        return formatted_result, srt_content
+        return formatted_result, srt_file
     except Exception as e:
         logger.exception(f"Error during transcription: {str(e)}")
         raise gr.Error(f"Failed to transcribe audio: {str(e)}")
 
 demo = gr.Blocks(theme=gr.themes.Ocean())
 
-# Create flagging callback with custom options
-flagging_callback = gr.CSVLogger()
-
 # Define interfaces first
 mf_transcribe = gr.Interface(
     fn=transcribe,
@@ -139,12 +150,8 @@ mf_transcribe = gr.Interface(
         " of arbitrary length."
     ),
     flagging_mode="manual",
-    flagging_options=[
-        "Text Issue",
-        "Timestamp Issue",
-        "Missing Content",
-        "Other Issue"
-    ]
+    flagging_options=["👍 Good", "👎 Bad"],
+    flagging_dir="flagged_data"
 )
 
 file_transcribe = gr.Interface(
@@ -167,12 +174,8 @@ file_transcribe = gr.Interface(
         " of arbitrary length."
     ),
     flagging_mode="manual",
-    flagging_options=[
-        "Text Issue",
-        "Timestamp Issue",
-        "Missing Content",
-        "Other Issue"
-    ]
+    flagging_options=["👍 Good", "👎 Bad"],
+    flagging_dir="flagged_data"
 )
 
 # Then set up the demo with the interfaces