Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import whisper
|
| 2 |
+
model = whisper.load_model("base")
|
| 3 |
+
|
| 4 |
+
import gradio as gr
|
| 5 |
+
from moviepy.editor import VideoFileClip
|
| 6 |
+
import os
|
| 7 |
+
|
| 8 |
+
def generate_subtitles(video_path):
|
| 9 |
+
# Extract audio from video
|
| 10 |
+
video = VideoFileClip(video_path)
|
| 11 |
+
audio_path = "temp_audio.wav"
|
| 12 |
+
video.audio.write_audiofile(audio_path)
|
| 13 |
+
|
| 14 |
+
# Transcribe audio using Whisper
|
| 15 |
+
result = model.transcribe(audio_path)
|
| 16 |
+
|
| 17 |
+
# Clean up temporary audio file
|
| 18 |
+
os.remove(audio_path)
|
| 19 |
+
|
| 20 |
+
# Format the subtitles in .srt format
|
| 21 |
+
srt_content = ""
|
| 22 |
+
for index, segment in enumerate(result['segments']):
|
| 23 |
+
start_time = segment['start']
|
| 24 |
+
end_time = segment['end']
|
| 25 |
+
text = segment['text']
|
| 26 |
+
|
| 27 |
+
# Convert seconds to SRT time format (HH:MM:SS,mmm)
|
| 28 |
+
def format_time(seconds):
|
| 29 |
+
hours = int(seconds // 3600)
|
| 30 |
+
minutes = int((seconds % 3600) // 60)
|
| 31 |
+
seconds = seconds % 60
|
| 32 |
+
milliseconds = int((seconds - int(seconds)) * 1000)
|
| 33 |
+
return f"{hours:02}:{minutes:02}:{int(seconds):02},{milliseconds:03}"
|
| 34 |
+
|
| 35 |
+
srt_content += f"{index + 1}\n"
|
| 36 |
+
srt_content += f"{format_time(start_time)} --> {format_time(end_time)}\n"
|
| 37 |
+
srt_content += f"{text}\n\n"
|
| 38 |
+
|
| 39 |
+
# Save the .srt file
|
| 40 |
+
srt_filename = "subtitles.srt"
|
| 41 |
+
with open(srt_filename, "w", encoding="utf-8") as srt_file:
|
| 42 |
+
srt_file.write(srt_content)
|
| 43 |
+
|
| 44 |
+
return srt_content, srt_filename
|