Devved11/RestAPI_AI_VOICE_DETECTION

1from fastapi import FastAPI, Header, HTTPException
2from pydantic import BaseModel
3import base64
4import numpy as np
5import pydub
6import tempfile
7import os
8from dotenv import load_dotenv
9load_dotenv()
10
11from models import detect_audio
12
13app = FastAPI(title="AI-Generated Voice Detection API - GUVI Hackathon")
14
15API_KEY = os.getenv("API_KEY")
16
17
18class AudioInput(BaseModel):
19    language: str
20    audioFormat: str
21    audioBase64: str
22
23
24@app.post("/api/voice-detection")
25def detect_voice(
26    input_data: AudioInput,
27    x_api_key: str = Header(None, alias="x-api-key")
28):
29    if x_api_key != API_KEY:
30        raise HTTPException(status_code=401, detail="Invalid API key")
31
32    if input_data.audioFormat.lower() != "mp3":
33        raise HTTPException(status_code=400, detail="Only mp3 supported")
34
35    try:
36        audio_bytes = base64.b64decode(input_data.audioBase64)
37
38        with tempfile.NamedTemporaryFile(suffix=".mp3", delete=False) as tmp:
39            tmp.write(audio_bytes)
40            tmp_path = tmp.name
41
42        try:
43            segment = pydub.AudioSegment.from_file(tmp_path, format="mp3")
44        finally:
45            os.remove(tmp_path)
46
47        segment = segment.set_channels(1).set_frame_rate(16000)
48        samples = segment.get_array_of_samples()
49        y = np.array(samples, dtype=np.float32) / 32768.0
50
51        classification, confidence, explanation = detect_audio(y)
52
53        return {
54            "status": "success",
55            "languageProvided": input_data.language,
56            "classification": classification,
57            "confidenceScore": confidence,
58            "explanation": explanation
59        }
60
61    except Exception as e:
62        raise HTTPException(status_code=400, detail=str(e))
63
64
65@app.get("/health")
66def health():
67    return {"status": "healthy"}
68
69
70@app.get("/")
71def root():
72    return {"message": "AI Voice Detection API - Ready"}
73