v2.4.0

What's Changed

• Introduces Pydantic Dataclasses for the different levels of data analysed and passed around between processing, TopoStats, GrainCrops, DisorderedTrace, Node, OrderedTrace, MatchedBranch, UnmatchedBranch, giving strong typing of attributes to facilitate cleaner development going forward.
• Revises and updates the .topostats HDF5 format to reflect the internal structure introduced by the above dataclasses. These are compressed format HDF5 and can easily be read by AFMReader which returns a dictionary representation. TopoStats contains functions to take these dictionaries and recreate the TopoStats objects with the same nested structure of attributes they were created from.
• All tests have been updated to reflect the new Dataclasses
• Switches to using Syrupy for regression tests rather than pytest-regtest
• Adds advanced documentation on the classes structure and notes on extending TopoStats.
• Revised some of the internal nomenclature to make it clearer what the required parameters are at different stages (mainly distinguishing between the different forms of skeletons required at different stages).
• Removes ImageGrainCrop and GrainCropsDirection classes. Instead the threshold and "direction" of a grain is stored as an attribute of the GrainCrop class.
• Filenames used in logging and for the creation of per-image output now include the file extension. This is so that older Bruker files which have the same stem (the part before the file extension) do not have data over-written. As a consequence the .topostats files that are written at the end of processing now reside under <output_dir>/processed/topostats/*.topostats as having a file and directory of the same name in the same location isn't possible.

New Contributions

This release see's new contributions from...

• @tobyallwood
• @ahobbs7

Thank you 🙏

Full Changelog: v2.3.2...v2.4.0