Skip to content

Inklet 4.0.0.dev5

Pre-release
Pre-release

Choose a tag to compare

@Mapika Mapika released this 12 Sep 14:13
· 134 commits to master since this release

This preview improves the shared plotting and composition engine: combine reusable plot layers, make independently styled variants, and resize mixed-content figures without losing authored plot height.

  • PlotSpec.copy() separates recorded instructions and nested plots while retaining explicit live data dependencies. extend() combines recipes with optional key prefixes and atomic collision checks. style() changes named instructions without replacing their positional data.
  • Panel.axes() and PlotSpec.axes() accept independent x/y options. Preset grids use matching label formatting and typography. Shared Series definitions support local colour/name overrides without duplicate keyword errors.
  • Automatic document rows grow when legends wrap, preserving data-region height even without globally shared margins. Plot dimension updates validate together before applying changes.
  • A five-panel example combines response variants, category bars, distribution overlays and a responsive diagram. It exports at 180 and 150 mm, retains old snapshots after live data edits, and includes an illustrated guide and native HTML viewer.

Also includes the previously completed bounded mixed GeoJSON adapter. Further geographic expansion is deferred; shared composition and reproducible authoring remain the priorities. General project templates and browser label/layout editing are still open.

Install explicitly:

python -m pip install "inklet==4.0.0.dev5"

Reusable plot guide · 4.0 roadmap

Validation includes 857 plotting/document tests, 39 targeted recipe/documentation checks, unchanged reviewed visual baselines, passing performance budgets, strict docs and package metadata checks, and an isolated core-only wheel installation. Full release CI checks the tagged commit across supported platforms and optional integrations.

This is a development prerelease. 3.1.0 remains stable. Experimental APIs and saved-state schemas may change. The attached wheel and source archive are the exact files uploaded to PyPI; SHA-256 hashes are attached in SHA256SUMS.