WIP

app.py

@@ -2,12 +2,32 @@ import gradio as gr
 import requests
 import subprocess
 from loguru import logger
+import datetime
 
 # Configure loguru
 logger.add("app.log", rotation="500 MB", level="DEBUG")
 
 API_URL = "https://skdpcqcdd929o4k3.us-east-1.aws.endpoints.huggingface.cloud"
 
+def format_time(seconds):
+    """Convert seconds to SRT time format (HH:MM:SS,mmm)"""
+    td = datetime.timedelta(seconds=float(seconds))
+    hours = td.seconds // 3600
+    minutes = (td.seconds % 3600) // 60
+    seconds = td.seconds % 60
+    milliseconds = td.microseconds // 1000
+    return f"{hours:02d}:{minutes:02d}:{seconds:02d},{milliseconds:03d}"
+
+def generate_srt(chunks):
+    """Generate SRT format subtitles from chunks"""
+    srt_content = []
+    for i, chunk in enumerate(chunks, 1):
+        start_time = format_time(chunk["timestamp"][0])
+        end_time = format_time(chunk["timestamp"][1])
+        text = chunk["text"].strip()
+        srt_content.append(f"{i}\n{start_time} --> {end_time}\n{text}\n\n")
+    return "".join(srt_content)
+
 # Check if ffmpeg is installed
 def check_ffmpeg():
     try:
@@ -20,7 +40,7 @@ def check_ffmpeg():
 # Initialize ffmpeg check
 check_ffmpeg()
 
-def transcribe(inputs, return_timestamps):
+def transcribe(inputs, return_timestamps, generate_subs):
     if inputs is None:
         logger.warning("No audio file submitted")
         raise gr.Error("No audio file submitted! Please upload or record an audio file before submitting your request.")
@@ -63,9 +83,9 @@ def transcribe(inputs, return_timestamps):
             "text": result["text"]
         }
         
+        chunks = []
         if return_timestamps and "chunks" in result:
             logger.info(f"Processing {len(result['chunks'])} chunks")
-            formatted_result["chunks"] = []
             for i, chunk in enumerate(result["chunks"]):
                 logger.debug(f"Processing chunk {i}: {chunk}")
                 try:
@@ -74,18 +94,27 @@ def transcribe(inputs, return_timestamps):
                     text = chunk.get("text", "").strip()
                     
                     if start_time is not None and end_time is not None:
-                        formatted_result["chunks"].append({
+                        chunk_data = {
                             "text": text,
                             "timestamp": [start_time, end_time]
-                        })
+                        }
+                        formatted_result["chunks"] = chunks
+                        chunks.append(chunk_data)
                     else:
                         logger.warning(f"Invalid timestamp in chunk {i}: {chunk}")
                 except Exception as chunk_error:
                     logger.error(f"Error processing chunk {i}: {str(chunk_error)}")
                     continue
-            logger.info(f"Successfully processed transcription with {len(formatted_result['chunks'])} chunks")
+            logger.info(f"Successfully processed transcription with {len(chunks)} chunks")
+        
+        # Generate subtitles if requested
+        srt_content = None
+        if generate_subs and chunks:
+            logger.info("Generating SRT subtitles")
+            srt_content = generate_srt(chunks)
+            logger.info("SRT subtitles generated successfully")
         
-        return formatted_result
+        return formatted_result, srt_content
     except Exception as e:
         logger.exception(f"Error during transcription: {str(e)}")
         raise gr.Error(f"Failed to transcribe audio: {str(e)}")
@@ -97,13 +126,16 @@ mf_transcribe = gr.Interface(
     inputs=[
         gr.Audio(sources="microphone", type="filepath"),
         gr.Checkbox(label="Include timestamps", value=True),
+        gr.Checkbox(label="Generate subtitles", value=True),
     ],
     outputs=[
         gr.JSON(label="Transcription", open=True),
+        gr.File(label="Subtitles (SRT)", visible=True),
     ],
     title="Whisper Large V3 Turbo: Transcribe Audio",
     description=(
         "Transcribe long-form microphone or audio inputs with the click of a button! "
+        "Generate subtitles for your videos in SRT format."
     ),
     flagging_mode="manual",
     flagging_options=[
@@ -119,13 +151,16 @@ file_transcribe = gr.Interface(
     inputs=[
         gr.Audio(sources="upload", type="filepath", label="Audio file"),
         gr.Checkbox(label="Include timestamps", value=True),
+        gr.Checkbox(label="Generate subtitles", value=True),
     ],
     outputs=[
         gr.JSON(label="Transcription", open=True),
+        gr.File(label="Subtitles (SRT)", visible=True),
     ],
     title="Whisper Large V3: Transcribe Audio",
     description=(
         "Transcribe long-form microphone or audio inputs with the click of a button! "
+        "Generate subtitles for your videos in SRT format."
     ),
     flagging_mode="manual",
     flagging_options=[