mgil
/
voice2txt
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Use of ogg format

This commit is contained in:
Mario Gonzalez Gil 2023-11-07 06:31:45 +01:00
parent f4e7b55e97
commit b94be5a668
2 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
main.py
View File

@ -35,7 +35,7 @@ def calculate_api_g(response: Response):
import os
os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = "/home/mario/.config/gcloud/application_default_credentials.json"
os.environ['GOOGLE_APPLICATION_CREDENTIALS'] = "token.json"
def transcribe_ogg_audio(audio_file_path, language_code='es-US'):
client = speech_v1.SpeechClient()
@ -45,12 +45,12 @@ def transcribe_ogg_audio(audio_file_path, language_code='es-US'):
audio = speech_v1.RecognitionAudio(content=content)
config = speech_v1.RecognitionConfig(
encoding=types.RecognitionConfig.AudioEncoding.FLAC, # Use OGG encoding
sample_rate_hertz=44100, # Update this to match your audio file
encoding=types.RecognitionConfig.AudioEncoding.OGG_OPUS,#.FLAC, # Use OGG encoding
sample_rate_hertz=48000, # Update this to match your audio file
language_code=language_code,
)
response = client.recognize(config=config, audio=audio)
return list(response.results[0].alternatives)[0].transcript
#results.alternatives[0].transcript
#results.alternatives[0].transcript

2
run.sh
View File

@ -1 +1 @@
uvicorn main:app --reload --port 7860
uvicorn main:app --reload --port 7862