Update app.py

app.py

@@ -2,6 +2,8 @@ from transformers import pipeline
 import gradio as gr
 import os
 import torch
+import srt
+from datetime import timedelta
 
 
 
@@ -15,13 +17,33 @@ if not auth_token:
 pipe = pipeline(model="fahadqazi/whisper-small-sindhi", device=device, token=auth_token)  # change to "your-username/the-name-you-picked"
 
 def transcribe(audio):
-    text = pipe(audio)["text"]
-    return text
+    # Perform transcription
+    result = pipe(audio)
+    transcription = result["text"]
+    
+    # Generate timestamps for transcription (You might need to tweak this to match your desired chunks)
+    segments = result.get("chunks", [])  # Assuming the model returns chunks (this depends on model and pipeline)
+    
+    # Create an SRT object
+    subtitle_generator = []
+    start_time = timedelta(seconds=0)
+    
+    for i, segment in enumerate(segments):
+        end_time = start_time + timedelta(seconds=segment["end"])  # Using segment['end'] to create time intervals
+        subtitle_generator.append(srt.Subtitle(index=i+1, start=start_time, end=end_time, content=segment["text"]))
+        start_time = end_time  # Update start_time for next subtitle
+    
+    # Write subtitles to .srt file
+    srt_file = "output.srt"
+    with open(srt_file, "w") as f:
+        f.write(srt.compose(subtitle_generator))
+    
+    return transcription, srt_file
 
 iface = gr.Interface(
     fn=transcribe,
     inputs=gr.Audio(type="filepath"),
-    outputs="text",
+    outputs=["text", "file"],
     title="Whisper Small Sindhi",
     description="Realtime demo for Sindhi speech recognition using a fine-tuned Whisper small model.",
 )