File size: 1,364 Bytes
66dff0f 887d074 087b372 f531a85 887d074 f531a85 887d074 b11f488 087b372 887d074 1571261 887d074 24f7854 29b5120 087b372 0fb7c05 29b5120 |
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 |
import whisper
import numpy as np
from pydub import AudioSegment
from deep_translator import GoogleTranslator, detection
import os
available_languages = GoogleTranslator().get_supported_languages(as_dict=True)
formatted_languages = {key.title(): value for key, value in available_languages.items()}
formatted_codes = {value: key.title() for key, value in available_languages.items()}
lang_detect_key = os.getenv("detect_language_api_key")
def audio_to_numpy(audio_file_input):
audio = AudioSegment.from_file(audio_file_input)
audio = audio.set_channels(1).set_frame_rate(16000)
samples = np.array(audio.get_array_of_samples(), dtype=np.float32)
return samples / np.iinfo(audio.array_type).max
def src_audio_to_eng_translator(audio_file_input, model_size = "turbo", target_lang = "English"):
audio_data = audio_to_numpy(audio_file_input)
model = whisper.load_model(model_size)
result = model.transcribe(audio_data)
input_text = result["text"]
src_lang_code = detection.single_detection(input_text, api_key = lang_detect_key)
src_lang = formatted_codes.get(src_lang_code, 'Source language not detected')
target_lang_code = formatted_languages.get(target_lang, 'en')
translated_text = GoogleTranslator(source='auto', target=target_lang_code).translate(input_text)
return input_text, translated_text, src_lang |