Version 2.0.0

Prebuilt package version 2.0.0.

• Refactored the core extractor into a proper library API (new SharpFrameExtractor class) and updated the CLI entrypoint so it's easier to embed in other tools
• Reworked the CLI around simpler sampling modes (e.g., target frame count vs. fixed time interval) and introduced a two-pass pipeline for more memory-efficient analysis and faster processing
• Improved sharpness scoring (Tenengrad/Sobel-based focus measure) to pick better “best frame per block” results, especially in challenging motion/blur segments
• Replaced hardcoded PNG file writes with pluggable output handlers so consumers can customize how/where frames are exported
• Added async file writing in the default file output handler to increase throughput when exporting large frame sets
• Reduced memory use for multi-worker extraction via shared-memory-backed NumPy arrays and a per-job shared memory pool (less copying, lower peak RAM)
• Further lowered memory footprint by decoupling stream block sizing from interval length and adding psutil for memory-aware behavior
• Fixed concurrency and resource-retention issues (e.g., worker future handling and a CLI memory leak) to improve long-run stability
• Improved CLI UX with more informative, colorful progress reporting via Rich (plus minor cleanup/formatting fixes)

Version 2.0.0-rc.2

Prebuilt package version 2.0.0-rc.2.

• Refactored the extractor so it can be used cleanly as a library (new SharpFrameExtractor class plus updated CLI entrypoint), making it easier to embed in other tools
• Added pluggable output handlers instead of hardcoded PNG writes, allowing consumers to customize how and where frames are exported
• Implemented async file writing support in the default file output handler, improving throughput when exporting large numbers of frames
• Introduced shared memory backing for NumPy arrays and a per-job shared memory pool, reducing copying and lowering overall memory usage for multi-worker extraction
• Reduced memory footprint by decoupling stream block size from interval length, and added psutil to support memory-aware behavior
• Fixed a concurrency issue in worker future handling and ensured completed futures are cleared to avoid retaining references
• Fixed a CLI memory leak by @dna-topoisomerase in #3
• Improved CLI UX with colorful progress bars using a custom Rich progress column
• Minor CLI import formatting cleanup

New Contributors

• @dna-topoisomerase made their first contribution in #3

Version 2.0.0-rc.1

Prebuilt package version 2.0.0-rc.1.

Version 1.6.6

• Added command line alias sfextract
• Limited CPU usage by default to 1/4 of the available CPUs

To install use the following command:

pip install git+https://github.com/cansik/sharp-frame-extractor.git@1.6.6

Version 1.6.5

Add option to extract all frames from the video --all.

To install use the following command:

pip install git+https://github.com/cansik/sharp-frame-extractor.git@1.6.5

Version 1.6.4

Fixed missing requirements.

To install use the following command:

pip install git+https://github.com/cansik/sharp-frame-extractor.git@1.6.4

Version 1.6.3

Fixes None extraction bug.

To install use the following command:

pip install git+https://github.com/cansik/sharp-frame-extractor.git@1.6.3

Version 1.6.2

This release removes the CPU estimation with the RAM size and adds a good looking progress bar.

To install use the following command:

pip install git+https://github.com/cansik/sharp-frame-extractor.git@1.6.2

Version 1.6.1

Small bugfixes.

To install use the following command:

pip install git+https://github.com/cansik/sharp-frame-extractor.git@1.6.1

Version 1.6.0

Finally it's possible to use sharp-frame-extractor as a runnable module python -m sharp_frame_extractor --help.

To install use the following command:

pip install git+https://github.com/cansik/sharp-frame-extractor.git@1.6.0