Spaces:
Running
Running
File size: 4,430 Bytes
ece3d5d 6d4d8fb ece3d5d a8c4113 ece3d5d 6d4d8fb a8c4113 ece3d5d a8c4113 ece3d5d a8c4113 ece3d5d a8c4113 ece3d5d a8c4113 ece3d5d a8c4113 6d4d8fb a8c4113 6d4d8fb a8c4113 |
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 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 |
import streamlit as st
from TTS.api import TTS
import torch
import subprocess
import locale
import assemblyai as aai
import requests
# Load TTS model
tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2", gpu=True).to("cuda")
# Function to translate video
class Translation:
def __init__(self, video_path, original_language, target_language):
self.video_path = video_path
self.original_language = original_language
self.target_language = target_language
def org_language_parameters(self, original_language):
if original_language == 'English':
self.lan_code = 'en'
elif original_language == 'German':
self.lan_code = 'de'
elif original_language == 'French':
self.lan_code = 'fr'
elif original_language == 'Spanish':
self.lan_code = 'es'
else:
self.lan_code = ''
def target_language_parameters(self, target_language):
if target_language == 'English':
self.tran_code = 'en'
elif target_language == 'German':
self.tran_code = 'de'
elif target_language == 'French':
self.tran_code = 'fr'
elif target_language == 'Spanish':
self.tran_code = 'es'
else:
self.tran_code = ''
def extract_audio(self):
audio_path = "output_audio.wav"
subprocess.run(['ffmpeg', '-i', self.video_path, '-vn', '-acodec', 'pcm_s16le', '-ar', '44100', '-ac', '2', audio_path])
st.success("Audio extracted successfully!")
return audio_path
def transcribe_audio(self, audio_path):
aai.settings.api_key = "c29eb650444a4ae4bea01d82dd861cbb"
config = aai.TranscriptionConfig(language_code=self.lan_code)
transcriber = aai.Transcriber(config=config)
transcript = transcriber.transcribe(audio_path)
transcript_text = transcript.text
return transcript_text
def translate_text(self, transcript_text):
base_url = "https://api.cognitive.microsofttranslator.com"
endpoint = "/translate"
headers = {
"Ocp-Apim-Subscription-Key": "cd226bb1f3644276bea01d82dd861cbb",
"Content-Type": "application/json",
"Ocp-Apim-Subscription-Region": "southeastasia"
}
params = {
"api-version": "3.0",
"from": self.lan_code,
"to": self.tran_code
}
body = [{"text": transcript_text}]
response = requests.post(base_url + endpoint, headers=headers, params=params, json=body)
response.raise_for_status()
translation = response.json()[0]["translations"][0]["text"]
return translation
def generate_audio(self, translated_text):
tts.tts_to_file(translated_text,
speaker_wav='output_audio.wav',
file_path="output_synth.wav",
language=self.tran_code
)
return "output_synth.wav"
def translate_video(self):
audio_path = self.extract_audio()
self.org_language_parameters(self.original_language)
self.target_language_parameters(self.target_language)
transcript_text = self.transcribe_audio(audio_path)
translated_text = self.translate_text(transcript_text)
translated_audio_path = self.generate_audio(translated_text)
# Convert translated audio to video
subprocess.run(['ffmpeg', '-i', self.video_path, '-i', translated_audio_path, '-c:v', 'copy', '-map', '0:v:0', '-map', '1:a:0', 'output_video.mp4'])
st.title("Translate Your Video")
st.write("Upload your video and select the original and target languages.")
# Upload video
video_file = st.file_uploader("Upload Video", type=["mp4"])
if video_file is not None:
# Get original and target languages
original_language = st.selectbox("Select Original Language", ['English', 'German', 'French', 'Spanish'])
target_language = st.selectbox("Select Target Language", ['English', 'German', 'French', 'Spanish'])
translation = Translation(video_path=video_file.name,
original_language=original_language,
target_language=target_language)
if st.button("Translate"):
translation.translate_video()
st.success("Video translation complete! You can download the translated video.")
|