Spaces:
Sleeping
Sleeping
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import subprocess
|
3 |
+
import os
|
4 |
+
import shutil
|
5 |
+
import uuid
|
6 |
+
from transformers import pipeline
|
7 |
+
from gtts import gTTS
|
8 |
+
|
9 |
+
|
10 |
+
def translate_video(file_path):
|
11 |
+
|
12 |
+
try:
|
13 |
+
audio_path = os.path.join(file_path, "audio.wav")
|
14 |
+
|
15 |
+
if not os.path.exists(audio_path):
|
16 |
+
raise FileNotFoundError("Audio extraction failed. yt-dlp did not produce a .wav file.")
|
17 |
+
|
18 |
+
# 3. Translate the audio using the whisper-tiny model
|
19 |
+
translator = pipeline(
|
20 |
+
"automatic-speech-recognition",
|
21 |
+
model="openai/whisper-tiny",
|
22 |
+
device="cpu"
|
23 |
+
)
|
24 |
+
|
25 |
+
translation = translator(audio_path, return_timestamps=True, generate_kwargs={"task": "translate"})
|
26 |
+
|
27 |
+
translated_text = translation["text"]
|
28 |
+
|
29 |
+
if not translated_text:
|
30 |
+
return "No speech was detected in the video.", None, video_path
|
31 |
+
|
32 |
+
# 4. Convert translated text to speech using gTTS
|
33 |
+
tts = gTTS(translated_text.strip(), lang='en')
|
34 |
+
translated_audio_path = os.path.join(temp_dir, "translated_audio.mp3")
|
35 |
+
tts.save(translated_audio_path)
|
36 |
+
|
37 |
+
return translated_text, translated_audio_path, video_path
|
38 |
+
|
39 |
+
except Exception as e:
|
40 |
+
gr.Warning(f"An unexpected error occurred: {str(e)}")
|
41 |
+
return f"An error occurred: {str(e)}", None, None
|
42 |
+
|
43 |
+
# Create the Gradio interface
|
44 |
+
iface = gr.Interface(
|
45 |
+
fn=translate_video,
|
46 |
+
inputs=gr.Video(label="Upload your video to translate"),
|
47 |
+
outputs=[
|
48 |
+
gr.Textbox(label="Translated Text", interactive=False),
|
49 |
+
gr.Audio(label="Translated Audio"),
|
50 |
+
gr.Video(label="Original Video"),
|
51 |
+
],
|
52 |
+
title="Twitter/X Video Translator",
|
53 |
+
description="Enter a link to a Twitter/X video to translate its audio to English. Handles videos longer than 30 seconds.",
|
54 |
+
allow_flagging="never",
|
55 |
+
examples=[["https://x.com/OpenAI/status/1790145460393734444"]]
|
56 |
+
)
|
57 |
+
|
58 |
+
if __name__ == "__main__":
|
59 |
+
if not os.path.exists("downloads"):
|
60 |
+
os.makedirs("downloads")
|
61 |
+
iface.launch()
|