v0.4.0rc4

What's Changed

• PP-11: convert, named-axis OME-TIFF copies of images with axis errors by @guijacquemet in #124
• PP-11c: conversion integrity (copy SHA-256, metadata epochs, exact CLI sizes) by @guijacquemet in #126
• PP-11b: the notebook's Make named-axis copies card by @guijacquemet in #127
• OME-Zarr axes and calibration: nothing guessed, nothing silent (#72, #73, #94) by @guijacquemet in #129
• Experiment design identity and counts: excluded folders, duplicate columns, output folder (#82, #83, #98) by @guijacquemet in #130
• Links stay inside the experiment and masks folders; --log is checked before it opens (#81, #99) by @guijacquemet in #128
• Micrometre conversions are always positive and finite; check and design exit 1 on errors (#93) by @guijacquemet in #133
• same_objects is a one-to-one matching; size and plane-count warnings follow the confirmed roles (#67, #74) by @guijacquemet in #131
• PP-12: staging and resume, crash-consistent publication, OME-Zarr inventory fingerprint (#86, #90) by @guijacquemet in #132
• PP-13: migration guide, CHANGELOG, Colab release check, docs pass; version 0.4.0rc4 by @guijacquemet in #134
• Audit B1, I1: keep the experiment, copies and output out of the masks folder by @guijacquemet in #135
• Audit I4: notebook inputs start from the earlier run's recorded inputs by @guijacquemet in #137
• Fix audit findings I2, I3, I5, m4, m5, m6 (what RunInputs gives the bricks) by @guijacquemet in #136
• Compare output and copy folders with the experiment by device and inode by @guijacquemet in #138

Full Changelog: v0.4.0rc3...v0.4.0rc4