Update app.py

app.py

@@ -1,20 +1,14 @@
-from fastapi import FastAPI
-import edge_tts
-import asyncio
 import os
 import time
-import io
-from fastapi.responses import StreamingResponse
-from tempfile import TemporaryDirectory
-from pydub import AudioSegment
+import asyncio
+from fastapi import FastAPI
+import edge_tts
+from fastapi.responses import FileResponse
 
 app = FastAPI()
 
 def split_text(text, max_chunk_size=500):
-    """Split text into chunks if it exceeds max_chunk_size."""
-    if len(text) <= max_chunk_size:
-        return [text]  
-
+    """Split text into smaller chunks."""
     sentences = text.replace('।', '.').replace('؟', '?').split('.')
     chunks = []
     current_chunk = []
@@ -38,71 +32,40 @@ def split_text(text, max_chunk_size=500):
     return chunks
 
 async def process_chunk(text, voice, temp_dir, chunk_index):
-    """Process a single chunk into an MP3 file."""
-    tmp_path = os.path.join(temp_dir, f"chunk_{chunk_index}_{int(time.time())}.mp3")
+    """Process a single chunk of text."""
+    tmp_path = os.path.join(temp_dir, f"chunk_{chunk_index}.mp3")
+    print(f"🎤 Processing chunk {chunk_index}: {text[:50]}...")  # Logging
     communicate = edge_tts.Communicate(text, voice)
     await communicate.save(tmp_path)
     return tmp_path
 
-async def combine_audio_files(chunk_files):
-    """Combine multiple MP3 files into one final MP3 file."""
+async def combine_audio_files(chunk_files, output_path):
+    """Combine multiple MP3 files into one."""
+    from pydub import AudioSegment
+
     combined = AudioSegment.empty()
-    
     for file in chunk_files:
-        audio_segment = AudioSegment.from_mp3(file)
-        combined += audio_segment
+        print(f"🔹 Adding {file} to final output")  # Logging
+        combined += AudioSegment.from_mp3(file)
 
-    output = io.BytesIO()
-    combined.export(output, format="mp3")
-    output.seek(0)
+    combined.export(output_path, format="mp3")
 
-    # Cleanup chunk files
     for file in chunk_files:
-        try:
-            os.remove(file)
-        except:
-            pass
-
-    return output
-
-@app.get("/")
-def home():
-    return {"message": "EdgeTTS FastAPI is running!"}
+        os.remove(file)
 
 @app.get("/tts")
 async def tts(text: str, voice: str = "en-US-AriaNeural"):
-    if not text.strip():
-        return {"error": "Text cannot be empty."}
-
-    text_chunks = split_text(text)  # Only splits if text > 500 characters
-
-    async def event_stream():
-        """Send real-time status updates to the client while processing."""
-        yield "Processing started...\n"
-        
-        if len(text_chunks) == 1:
-            # Single request processing
-            output_audio = io.BytesIO()
-            communicate = edge_tts.Communicate(text_chunks[0], voice)
-            await communicate.save(output_audio)
-            output_audio.seek(0)
-            yield "Processing completed. Downloading audio...\n"
-            yield output_audio.read()
-            return
-
-        with TemporaryDirectory() as temp_dir:
-            # Process all chunks concurrently
-            tasks = [process_chunk(chunk, voice, temp_dir, i) for i, chunk in enumerate(text_chunks)]
-            chunk_files = await asyncio.gather(*tasks)
-            
-            yield f"Processing {len(text_chunks)} chunks completed. Merging audio...\n"
-            output_audio = await combine_audio_files(chunk_files)
-
-        yield "Merging completed. Downloading final audio...\n"
-        yield output_audio.read()
-
-    return StreamingResponse(event_stream(), media_type="audio/mpeg")
-
-if __name__ == "__main__":
-    import uvicorn
-    uvicorn.run(app, host="0.0.0.0", port=7860)
+    """Main API function to process TTS."""
+    temp_dir = "temp_audio"
+    os.makedirs(temp_dir, exist_ok=True)
+    chunks = split_text(text)
+    
+    if len(chunks) == 1:
+        return await FileResponse(await process_chunk(text, voice, temp_dir, 0), media_type="audio/mpeg", filename="speech.mp3")
+
+    chunk_files = await asyncio.gather(*[process_chunk(ch, voice, temp_dir, i) for i, ch in enumerate(chunks)])
+
+    output_file = "final_output.mp3"
+    await combine_audio_files(chunk_files, output_file)
+
+    return FileResponse(output_file, media_type="audio/mpeg", filename="speech.mp3")