Update app.py

app.py

@@ -1,37 +1,41 @@
 from fastapi import FastAPI, UploadFile, File
-from fastapi.responses import JSONResponse
+from fastapi.middleware.cors import CORSMiddleware
 import torchaudio
 import torch
-from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+import uvicorn
 import io
 
-# Load model and processor
-processor = Wav2Vec2Processor.from_pretrained("Mustafaa4a/ASR-Somali")
-model = Wav2Vec2ForCTC.from_pretrained("Mustafaa4a/ASR-Somali")
-model.eval()
+app = FastAPI()
 
-# Initialize FastAPI
-app = FastAPI(
-    title="Somali Speech-to-Text API",
-    description="Upload a Somali audio file (.wav) and receive text transcription using ASR model.",
-    version="1.0",
+# Allow requests from Flutter (localhost or any domain)
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
 )
 
+# Load model and processor once at startup
+processor = Wav2Vec2Processor.from_pretrained("Mustafaa4a/ASR-Somali")
+model = Wav2Vec2ForCTC.from_pretrained("Mustafaa4a/ASR-Somali")
+
 @app.post("/transcribe")
-async def transcribe(audio: UploadFile = File(...)):
-    # Read audio bytes
-    audio_bytes = await audio.read()
-    waveform, sample_rate = torchaudio.load(io.BytesIO(audio_bytes))
+async def transcribe_audio(file: UploadFile = File(...)):
+    contents = await file.read()
+    audio_bytes = io.BytesIO(contents)
+
+    waveform, sample_rate = torchaudio.load(audio_bytes)
 
-    # Ensure 16kHz sample rate
     if sample_rate != 16000:
-        waveform = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)(waveform)
+        resampler = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)
+        waveform = resampler(waveform)
 
-    # Process input
     inputs = processor(waveform.squeeze(), sampling_rate=16000, return_tensors="pt")
     with torch.no_grad():
         logits = model(**inputs).logits
 
     predicted_ids = torch.argmax(logits, dim=-1)
     transcription = processor.decode(predicted_ids[0])
-    return JSONResponse(content={"transcription": transcription})
+    return {"transcription": transcription}