framestoryx is a successor to FrameStory for better compatibility with modern Python tooling. Leveraging state-of-the-art machine learning models, it can provide detailed descriptions of video content, making it a powerful tool for content analysis, accessibility, and summarization.

To install framestoryx, you can use pip:

pip install git+https://github.com/fossiaorg/framestoryx

Using framestoryx is straightforward. Below are examples demonstrating how to extract and describe significant frames from videos with various parameters.

from framestoryx.video_describer import VideoDescriber

video_url = "https://example.com/video.mp4"
describer = VideoDescriber(show_progress=True)
descriptions = describer.get_video_descriptions(video_url=video_url)
print(descriptions)

video_path = "/path/to/your/video.mp4"
describer = VideoDescriber(show_progress=True, max_tokens=50)
descriptions = describer.get_video_descriptions(video_path=video_path)
print(descriptions)

The extract_significant_frames method allows you to customize the threshold for what constitutes a "significant" change between frames.

video_url = "https://example.com/video.mp4"
describer = VideoDescriber(threshold=25000)
descriptions = describer.get_video_descriptions(video_url=video_url)
print(descriptions)

These examples demonstrate the versatility of frame_story in processing videos from different sources and with various levels of detail in descriptions.

• Extraction of significant frames from videos for detailed analysis.
• Generation of descriptive text for each significant frame using state-of-the-art image captioning models.
• Support for videos from URLs or local file paths.
• Customizable settings for progress display, description length, and frame extraction threshold.
• Easy to integrate into Python projects for content analysis, summarization, and accessibility applications.

Contributions, issues, and feature requests are welcome! Feel free to check the issues page.

This project is licensed under the MIT License.