Add --frame-start / --frame-end to restrict the frame range, and
--frame-step / --frame-count (mutually exclusive) to decimate within
that range. Kept frames are re-numbered contiguously from 0000 so
LVGL's sequential player never sees index gaps.
Interactive menu gains a matching "是否抽帧" step; CLI and menu both
display a human-readable sampling summary before conversion starts.
Bump version to 0.3.0.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The menu's source step now uses a two-step selection (_select_gifs): pick a
mode — tick several GIFs (questionary.checkbox), all GIFs in the directory at
once, or enter a path manually — then convert them in one batch. Single-GIF
selection and the entire command line stay unchanged.
- add run_conversions(gifs, template) that reuses the unchanged single-GIF
run_conversion() per item, continues on per-GIF errors, and prints an
aggregate "Batch complete: X/Y" summary (non-zero exit if any failed)
- with multiple GIFs the per-GIF stem is forced as the frame prefix so frames
never collide; the prefix prompt is only shown for a single GIF
- update README/AGENTS for the multi/all selection and batch behavior
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Running without a GIF argument (or with -i/--interactive) now launches a
questionary-based menu that selects a source GIF from the current directory
and configures every conversion parameter before running. Passing a GIF on
the command line behaves exactly as before.
- extract a unified run_conversion(ConversionConfig) pipeline shared by the
CLI and the menu, removing the duplicated convert loop in main()
- make the gif positional optional and add -i/--interactive
- import questionary lazily inside the menu so plain CLI usage gains no hard
dependency, with a clear install hint when it is missing
- add __version__ / --version (0.2.0) and document the menu in README/AGENTS
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
