File size: 3,265 Bytes
3cf82c2
09ab406
 
 
 
 
06e4c74
3cf82c2
 
 
 
 
 
 
09ab406
 
 
4d56027
 
 
 
09ab406
 
 
3cf82c2
 
 
 
 
09ab406
 
 
 
 
 
06e4c74
4d56027
 
 
 
3cf82c2
 
 
 
4d56027
 
 
 
 
09ab406
 
 
4d56027
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7bcf8d7
09ab406
06e4c74
4d56027
 
 
 
 
 
 
 
 
09ab406
 
 
4d56027
 
 
 
 
09ab406
 
 
4d56027
 
 
 
3cf82c2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
from fastapi import FastAPI, HTTPException
from fastapi.responses import JSONResponse, FileResponse
from pydantic import BaseModel
import numpy as np
import io
import soundfile as sf
import base64
import logging
import torch
import librosa
from transformers import Wav2Vec2ForCTC, AutoProcessor
from pathlib import Path

# Import functions from other modules
from asr import transcribe, ASR_LANGUAGES
from tts import synthesize, TTS_LANGUAGES
from lid import identify

# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

app = FastAPI(title="MMS: Scaling Speech Technology to 1000+ languages")

# Define request models
class AudioRequest(BaseModel):
    audio: str  # Base64 encoded audio data
    language: str

class TTSRequest(BaseModel):
    text: str
    language: str
    speed: float

@app.post("/transcribe")
async def transcribe_audio(request: AudioRequest):
    try:
        audio_bytes = base64.b64decode(request.audio)
        audio_array, sample_rate = sf.read(io.BytesIO(audio_bytes))
        
        # Convert to mono if stereo
        if len(audio_array.shape) > 1:
            audio_array = audio_array.mean(axis=1)
        
        result = transcribe(audio_array, request.language)
        return JSONResponse(content={"transcription": result})
    except Exception as e:
        logger.error(f"Error in transcribe_audio: {str(e)}")
        raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")

@app.post("/synthesize")
async def synthesize_speech(request: TTSRequest):
    try:
        audio, filtered_text = synthesize(request.text, request.language, request.speed)
        
        # Convert numpy array to bytes
        buffer = io.BytesIO()
        sf.write(buffer, audio, 22050, format='wav')
        buffer.seek(0)
        
        return FileResponse(
            buffer, 
            media_type="audio/wav", 
            headers={"Content-Disposition": "attachment; filename=synthesized_audio.wav"}
        )
    except Exception as e:
        logger.error(f"Error in synthesize_speech: {str(e)}")
        raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")

@app.post("/identify")
async def identify_language(request: AudioRequest):
    try:
        audio_bytes = base64.b64decode(request.audio)
        audio_array, sample_rate = sf.read(io.BytesIO(audio_bytes))
        
        result = identify(audio_array)
        return JSONResponse(content={"language_identification": result})
    except Exception as e:
        logger.error(f"Error in identify_language: {str(e)}")
        raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")

@app.get("/asr_languages")
async def get_asr_languages():
    try:
        return JSONResponse(content=ASR_LANGUAGES)
    except Exception as e:
        logger.error(f"Error in get_asr_languages: {str(e)}")
        raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")

@app.get("/tts_languages")
async def get_tts_languages():
    try:
        return JSONResponse(content=TTS_LANGUAGES)
    except Exception as e:
        logger.error(f"Error in get_tts_languages: {str(e)}")
        raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")