Changelog¶
All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
Unreleased¶
v0.2.1 - 2026-11-19¶
Added¶
- An Acknowledgements section in the README: the open tools the pipeline uses, and the CC BY-NC 4.0
licence of the MMS-LID weights that the
lidextra downloads.
Changed¶
- The contact address is contact@burkimbia.com: in the PyPI metadata, the README and
__email__.
v0.2.0 - 2026-11-19¶
First release on PyPI.
Added¶
VLMOCRruns on Apple's GPU (MPS) when there is no CUDA, and loads Qwen3-VL in bfloat16 there: on a Mac M5 it reads as float32 does, 30 times faster. It fell back to the CPU before.strip_overlaysandprocess_batch(overlay_frames=60)take logos and banners out of the frames before merging;mark_non_subtitlesmarks credits, the video's title and scrolling text in the merged segments.windowsjoins subtitles less than 2 s apart into windows of 5 to 30 s, as segmentsAudioGeneratorcuts.LanguageID(MMS-LID, the newlidextra) says which language each window is spoken in;written_insays whether a subtitle is written in French or English.TextIndex(the newcopiesextra) finds which video already read a subtitle line comes from, to tell a copy published by another channel.- A pipeline diagram drawn in draw.io (
docs/diagrams/pipeline.drawio) in the README and the guide, and a README example that runs the whole pipeline. scripts/figures.pymakes the guide's figures fromsandbox/sample_short.mp4with the library itself.
Security¶
- certifi 2026.10.23 in
uv.lock, and uv pinned to 0.13.4 in CI and in the pre-commit hook: earlier versions trust the compromised emSign ECC Root CA - G3 root.
Changed¶
- The project lives at
BurkimbIA/Frame2Text4LLM; the documentation moved to https://burkimbia.github.io/Frame2Text4LLM/. - The version is set in
pyproject.toml;HISTORY.mdis nowCHANGELOG.md.
[0.1.0] - 2026-10-06¶
Not published on PyPI.
Added¶
pyproject.toml, built with hatchling, anduv.lockfor development. The OCR engines are extras:easyocr,paddle,vlm,openai,all.VLMOCRreads with any image-text-to-text model of transformers, Qwen3-VL-2B by default: on 48 subtitle frames it made 0.53 % character error, against 37.49 % fortool="easyocr".merge_segments: readings of one subtitle are merged when they share most of their words, one empty frame does not end a subtitle, and each segment hasbest_time, the frame read best.OCRManager(region=...): the subtitle band, detected once per video when it is not given.YoutubeConsumer(client=...): the pytubefix client,MWEBby default; files are namedyoutube_<id>.mp4.- A documentation site: https://sawallesalfo.github.io/Frame2Text4LLM/.
Fixed¶
- A segment ends when its subtitle leaves the screen; a subtitle seen on one frame no longer lasts 0 s.
AudioGeneratorreads the frames'MM:SS:mmmtimes and seeks before reading: audio was cut in the wrong place.SubtitleRegionDetectortakes the full width of the image; its margin cut the first letters.OCRManagerloads a model once, not once per image.SRTExporterwrites valid times past an hour, and each subtitle's own end.
[0.0.1]¶
- First version, not published.