Obtención de la transcripción de archivos de audio corto (.flac)

transcription_short_audios.py

import audioread
import glob
import json
import os
from google.cloud import speech
from google.cloud.speech import enums
from google.cloud.speech import types

x = os.path.dirname(os.path.realpath(__file__))
x = os.path.join(x, "*.flac")
audios = glob.glob(x)
transcripts = {}

config = {
    "encoding": enums.RecognitionConfig.AudioEncoding.FLAC,
    "language_code": "es-MX"
}

output = {}

client = speech.SpeechClient()

for af in audios:
    with open(af, "rb") as audio_file:
        audio_channels = 1
        with audioread.audio_open(af) as afc:
            audio_channels = int(afc.channels)

        content = audio_file.read()
        af_name = os.path.basename(audio_file.name)

        config.update({
            "audio_channel_count": audio_channels,
            "enable_separate_recognition_per_channel": True
        })

        rec_config = types.RecognitionConfig(**config)

        audio = types.RecognitionAudio(content=content)
        response = client.recognize(rec_config, audio)

        output.update({af_name: {"transcripciones": [], "confianza": []}})
        print("\n\nArchivo: {}\n".format(af_name))
        for result in response.results:
            for r in result.alternatives:
                output[af_name]["transcripciones"].append(r.transcript)
                output[af_name]["confianza"].append(r.transcript)
                print(r.transcript)