101-0731/api/speech2text.py

import tempfile
from typing import List, Any
from fastapi import Request, APIRouter, UploadFile, File
from fastapi.responses import FileResponse, PlainTextResponse
from fastapi.exceptions import HTTPException
from fastapi.encoders import jsonable_encoder
from urllib.parse import urlparse, urljoin
from pathlib import Path
from icecream import ic
from google.oauth2 import service_account
from google.cloud import speech

client_file = Path(__file__).parent.parent/'keys/pure-lodge-426406-e4-af94156a748a.json'
credentials = service_account.Credentials.from_service_account_file(client_file)
client = speech.SpeechClient(credentials=credentials)

router = APIRouter()

@router.post('/gcp')
def gcp(language_code: str=None, file: UploadFile = File()):
    extension = file.filename.split(".")[-1] 
    if extension not in ("mp3", "wav", "webm"):
        return HTTPException(status_code=400, detail="Audio must be mp3, wav, webm or webm format!")
    content = file.file.read()
    audio = speech.RecognitionAudio(content=content)
    if extension=='mp3': encoding=speech.RecognitionConfig.AudioEncoding.MP3
    elif extension=='wav': encoding=speech.RecognitionConfig.AudioEncoding.LINEAR16
    elif extension=='webm': encoding=speech.RecognitionConfig.AudioEncoding.WEBM_OPUS
    else: return HTTPException(status_code=400, detail="no such encoding.")
    config = speech.RecognitionConfig(
        encoding=encoding,
        sample_rate_hertz=48000,
        language_code=language_code
    )
    response = client.recognize(config=config, audio=audio)
    results = [results.alternatives[0].transcript for results in response.results]
    ic(results)
    
    return results