AllMark / app.py
NeeravS's picture
Update app.py
3a3c2be verified
raw
history blame
2.12 kB
import os
import gradio as gr
# Clone the GitHub repository containing app.py if not already cloned
REPO_URL = "https://github.com/NeeravSood/AllMark-MVP" # Replace with your GitHub repo URL
if not os.path.exists("AllMark-MVP"): # Replace with your repo's folder name
os.system(f"git clone {REPO_URL}")
# Import the backend code after cloning the repo
from repository_name.app import DeepfakeAnalyzer # Adjust based on your repo and file structure
# Initialize the analyzer instance
analyzer = DeepfakeAnalyzer()
# Define the function for Gradio to call
def analyze_video(video_file):
results = analyzer.analyze_media(video_file)
combined_probability = results['combined_assessment']
audio_analysis = results["audio_analysis"]
video_probability = results['video_analysis']['probability']
frame_count = len(results['video_analysis']['frame_results'])
# Format the output for display
output = {
"Audio Analysis": audio_analysis,
"Deepfake Probability (Combined Assessment)": f"{combined_probability:.2f}%",
"Video Analysis": {
"Deepfake Probability": f"{video_probability:.4f}",
"Frames Analyzed": frame_count,
"Frame Analysis Summary": [
{
"Frame Number": frame_result["frame_number"],
"Noise Level": frame_result["noise"],
"Edge Density": frame_result["edge_density"],
"Color Consistency": frame_result["color_consistency"],
"Temporal Difference": frame_result["temporal_difference"],
"Probability": frame_result["probability"]
}
for frame_result in results['video_analysis']['frame_results']
]
}
}
return output
# Define the Gradio interface
interface = gr.Interface(
fn=analyze_video,
inputs=gr.Video(label="Upload Video"),
outputs="json",
title="Deepfake Analyzer",
description="Upload a video to analyze for deepfake content."
)
# Launch Gradio app
if __name__ == "__main__":
interface.launch()