Spaces:
Sleeping
Sleeping
Afrinetwork7
commited on
Commit
•
3cf82c2
1
Parent(s):
4d56027
Update app.py
Browse files
app.py
CHANGED
@@ -1,15 +1,20 @@
|
|
1 |
-
from fastapi import FastAPI,
|
2 |
from fastapi.responses import JSONResponse, FileResponse
|
3 |
-
import uvicorn
|
4 |
from pydantic import BaseModel
|
5 |
import numpy as np
|
6 |
import io
|
7 |
import soundfile as sf
|
8 |
import base64
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
from asr import transcribe, ASR_LANGUAGES
|
10 |
from tts import synthesize, TTS_LANGUAGES
|
11 |
from lid import identify
|
12 |
-
import logging
|
13 |
|
14 |
# Configure logging
|
15 |
logging.basicConfig(level=logging.INFO)
|
@@ -17,21 +22,26 @@ logger = logging.getLogger(__name__)
|
|
17 |
|
18 |
app = FastAPI(title="MMS: Scaling Speech Technology to 1000+ languages")
|
19 |
|
|
|
|
|
|
|
|
|
|
|
20 |
class TTSRequest(BaseModel):
|
21 |
text: str
|
22 |
language: str
|
23 |
speed: float
|
24 |
|
25 |
-
class AudioRequest(BaseModel):
|
26 |
-
audio: str # Base64 encoded audio data
|
27 |
-
language: str
|
28 |
-
|
29 |
@app.post("/transcribe")
|
30 |
async def transcribe_audio(request: AudioRequest):
|
31 |
try:
|
32 |
audio_bytes = base64.b64decode(request.audio)
|
33 |
audio_array, sample_rate = sf.read(io.BytesIO(audio_bytes))
|
34 |
|
|
|
|
|
|
|
|
|
35 |
result = transcribe(audio_array, request.language)
|
36 |
return JSONResponse(content={"transcription": result})
|
37 |
except Exception as e:
|
@@ -83,4 +93,4 @@ async def get_tts_languages():
|
|
83 |
return JSONResponse(content=TTS_LANGUAGES)
|
84 |
except Exception as e:
|
85 |
logger.error(f"Error in get_tts_languages: {str(e)}")
|
86 |
-
raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")
|
|
|
1 |
+
from fastapi import FastAPI, HTTPException
|
2 |
from fastapi.responses import JSONResponse, FileResponse
|
|
|
3 |
from pydantic import BaseModel
|
4 |
import numpy as np
|
5 |
import io
|
6 |
import soundfile as sf
|
7 |
import base64
|
8 |
+
import logging
|
9 |
+
import torch
|
10 |
+
import librosa
|
11 |
+
from transformers import Wav2Vec2ForCTC, AutoProcessor
|
12 |
+
from pathlib import Path
|
13 |
+
|
14 |
+
# Import functions from other modules
|
15 |
from asr import transcribe, ASR_LANGUAGES
|
16 |
from tts import synthesize, TTS_LANGUAGES
|
17 |
from lid import identify
|
|
|
18 |
|
19 |
# Configure logging
|
20 |
logging.basicConfig(level=logging.INFO)
|
|
|
22 |
|
23 |
app = FastAPI(title="MMS: Scaling Speech Technology to 1000+ languages")
|
24 |
|
25 |
+
# Define request models
|
26 |
+
class AudioRequest(BaseModel):
|
27 |
+
audio: str # Base64 encoded audio data
|
28 |
+
language: str
|
29 |
+
|
30 |
class TTSRequest(BaseModel):
|
31 |
text: str
|
32 |
language: str
|
33 |
speed: float
|
34 |
|
|
|
|
|
|
|
|
|
35 |
@app.post("/transcribe")
|
36 |
async def transcribe_audio(request: AudioRequest):
|
37 |
try:
|
38 |
audio_bytes = base64.b64decode(request.audio)
|
39 |
audio_array, sample_rate = sf.read(io.BytesIO(audio_bytes))
|
40 |
|
41 |
+
# Convert to mono if stereo
|
42 |
+
if len(audio_array.shape) > 1:
|
43 |
+
audio_array = audio_array.mean(axis=1)
|
44 |
+
|
45 |
result = transcribe(audio_array, request.language)
|
46 |
return JSONResponse(content={"transcription": result})
|
47 |
except Exception as e:
|
|
|
93 |
return JSONResponse(content=TTS_LANGUAGES)
|
94 |
except Exception as e:
|
95 |
logger.error(f"Error in get_tts_languages: {str(e)}")
|
96 |
+
raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")
|