Commit
Β·
fc7755f
1
Parent(s):
beae951
Fix file handling and FFmpeg conversion
Browse files
app.py
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
import gradio as gr
|
2 |
-
from utils import process_video
|
3 |
|
4 |
# Define supported languages
|
5 |
language_map = {
|
@@ -16,33 +16,85 @@ language_map = {
|
|
16 |
"Japanese": "Helsinki-NLP/opus-mt-en-jap"
|
17 |
}
|
18 |
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
25 |
|
26 |
# Define Gradio Interface
|
27 |
-
with gr.Blocks() as demo:
|
28 |
-
gr.Markdown("# AI-Powered Video Subtitling")
|
29 |
-
gr.Markdown("Upload a video and select a language to generate subtitles.")
|
30 |
|
31 |
with gr.Row():
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
40 |
output_srt = gr.File(label="Download Subtitles")
|
41 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
generate_button.click(
|
43 |
generate_subtitles,
|
44 |
inputs=[video_input, language_dropdown],
|
45 |
-
outputs=output_srt
|
46 |
)
|
47 |
|
48 |
demo.launch()
|
|
|
1 |
import gradio as gr
|
2 |
+
from utils import process_video # Ensure this points to the updated utils.py
|
3 |
|
4 |
# Define supported languages
|
5 |
language_map = {
|
|
|
16 |
"Japanese": "Helsinki-NLP/opus-mt-en-jap"
|
17 |
}
|
18 |
|
19 |
+
# Custom CSS for dark mode and animations
|
20 |
+
css = """
|
21 |
+
body {
|
22 |
+
background-color: #1a1a1a;
|
23 |
+
color: #e0e0e0;
|
24 |
+
font-family: 'Arial', sans-serif;
|
25 |
+
}
|
26 |
+
.gradio-container {
|
27 |
+
max-width: 1200px;
|
28 |
+
margin: 0 auto;
|
29 |
+
padding: 20px;
|
30 |
+
border-radius: 10px;
|
31 |
+
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
|
32 |
+
}
|
33 |
+
.file-preview {
|
34 |
+
border: 2px dashed #6c757d;
|
35 |
+
padding: 20px;
|
36 |
+
border-radius: 10px;
|
37 |
+
}
|
38 |
+
.progress-text {
|
39 |
+
font-size: 16px;
|
40 |
+
color: #28a745;
|
41 |
+
animation: blink 1s infinite;
|
42 |
+
}
|
43 |
+
@keyframes blink {
|
44 |
+
50% { opacity: 0.5; }
|
45 |
+
}
|
46 |
+
"""
|
47 |
|
48 |
# Define Gradio Interface
|
49 |
+
with gr.Blocks(theme=gr.themes.Monochrome(), css=css) as demo:
|
50 |
+
gr.Markdown("# π₯ AI-Powered Video Subtitling")
|
51 |
+
gr.Markdown("Upload a video (MP4/MKV/AVI) and select a language to generate subtitles.")
|
52 |
|
53 |
with gr.Row():
|
54 |
+
with gr.Column(scale=2):
|
55 |
+
video_input = gr.File(
|
56 |
+
label="Upload Video File",
|
57 |
+
file_types=["mp4", "mkv", "avi"],
|
58 |
+
elem_classes=["file-preview"]
|
59 |
+
)
|
60 |
+
with gr.Column(scale=1):
|
61 |
+
language_dropdown = gr.Dropdown(
|
62 |
+
choices=list(language_map.keys()),
|
63 |
+
label="Select Subtitle Language",
|
64 |
+
value="English"
|
65 |
+
)
|
66 |
+
|
67 |
+
generate_button = gr.Button("Generate Subtitles π")
|
68 |
+
progress_text = gr.Textbox(
|
69 |
+
label="Progress",
|
70 |
+
interactive=False,
|
71 |
+
elem_classes=["progress-text"]
|
72 |
+
)
|
73 |
output_srt = gr.File(label="Download Subtitles")
|
74 |
|
75 |
+
def generate_subtitles(video_file, language):
|
76 |
+
try:
|
77 |
+
# Validate file type
|
78 |
+
if not video_file.name.lower().endswith(('.mp4', '.mkv', '.avi')):
|
79 |
+
return None, "β Invalid file type. Please upload an MP4, MKV, or AVI file."
|
80 |
+
|
81 |
+
# Update progress
|
82 |
+
progress = "π Processing video..."
|
83 |
+
yield None, progress # Initial progress update
|
84 |
+
|
85 |
+
# Process video
|
86 |
+
srt_path = process_video(video_file.name, language)
|
87 |
+
if srt_path:
|
88 |
+
yield gr.File(srt_path), "β
Subtitles generated successfully!"
|
89 |
+
else:
|
90 |
+
yield None, "β Error during processing. Check logs."
|
91 |
+
except Exception as e:
|
92 |
+
yield None, f"β Error: {str(e)}"
|
93 |
+
|
94 |
generate_button.click(
|
95 |
generate_subtitles,
|
96 |
inputs=[video_input, language_dropdown],
|
97 |
+
outputs=[output_srt, progress_text]
|
98 |
)
|
99 |
|
100 |
demo.launch()
|
utils.py
CHANGED
@@ -7,16 +7,18 @@ import subprocess
|
|
7 |
# Load Whisper model
|
8 |
model = whisper.load_model("base")
|
9 |
|
10 |
-
def process_video(video_path, language):
|
11 |
-
|
12 |
-
|
|
|
|
|
13 |
|
14 |
try:
|
15 |
# Convert video to MP4 using ffmpeg
|
16 |
-
print("Converting video to MP4...")
|
17 |
subprocess.run(
|
18 |
["ffmpeg", "-i", video_path, "-c:v", "libx264", "-preset", "fast", output_video_path],
|
19 |
-
check=True,
|
20 |
stdout=subprocess.PIPE,
|
21 |
stderr=subprocess.PIPE
|
22 |
)
|
@@ -48,6 +50,7 @@ def process_video(video_path, language): # Accept file path, not file object
|
|
48 |
if not model_name:
|
49 |
return f"Unsupported language: {language}"
|
50 |
|
|
|
51 |
print(f"Loading translation model: {model_name}")
|
52 |
if language == "Telugu":
|
53 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
@@ -76,7 +79,7 @@ def process_video(video_path, language): # Accept file path, not file object
|
|
76 |
end = f"{segment['end']:.3f}".replace(".", ",")
|
77 |
text = segment["text"].strip()
|
78 |
f.write(f"{i}\n00:00:{start} --> 00:00:{end}\n{text}\n\n")
|
79 |
-
print("SRT file
|
80 |
return srt_path
|
81 |
|
82 |
except subprocess.CalledProcessError as e:
|
|
|
7 |
# Load Whisper model
|
8 |
model = whisper.load_model("base")
|
9 |
|
10 |
+
def process_video(video_path, language):
|
11 |
+
# Create a temporary directory
|
12 |
+
temp_dir = tempfile.gettempdir()
|
13 |
+
output_video_path = os.path.join(temp_dir, "converted_video.mp4")
|
14 |
+
srt_path = os.path.join(temp_dir, "subtitles.srt")
|
15 |
|
16 |
try:
|
17 |
# Convert video to MP4 using ffmpeg
|
18 |
+
print(f"Converting video: {video_path} to MP4...")
|
19 |
subprocess.run(
|
20 |
["ffmpeg", "-i", video_path, "-c:v", "libx264", "-preset", "fast", output_video_path],
|
21 |
+
check=True,
|
22 |
stdout=subprocess.PIPE,
|
23 |
stderr=subprocess.PIPE
|
24 |
)
|
|
|
50 |
if not model_name:
|
51 |
return f"Unsupported language: {language}"
|
52 |
|
53 |
+
# Load translation model
|
54 |
print(f"Loading translation model: {model_name}")
|
55 |
if language == "Telugu":
|
56 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
|
|
79 |
end = f"{segment['end']:.3f}".replace(".", ",")
|
80 |
text = segment["text"].strip()
|
81 |
f.write(f"{i}\n00:00:{start} --> 00:00:{end}\n{text}\n\n")
|
82 |
+
print(f"SRT file saved to {srt_path}")
|
83 |
return srt_path
|
84 |
|
85 |
except subprocess.CalledProcessError as e:
|