Hdfmap

Add support for detectors with image_data file paths, plus improvements to NeXus defaults and other stuff.
hdf_loader.py
- added standard arguments 'swmr=True' and 'libver=latest' to h5py.File usages.

hdfmap_class.py
- removed several unused 'info_...' funcitons, replaced with 'info_names(...)'
- added first_last_scannables function
- added 'numeric_only' option to get_scannables
- improved doc strings

nexus.py
- improved info_nexus output
- renamed 'nexus_defaults' to 'nexus_default_paths'
- added 'nexus_default_names'
- added collection of 'image_data' filename detector datasets, image_data now may contain paths to non-numeric datasets (docs updated)
- get_plot_data improved using new functions.

test_edge_cases.py
- added additional edge case files and regenerated test_files.json

pypi-publish.yml
- updated to v4 of the artifact actions to fix depreciation error


What's Changed
* fixes for eval and str array by DanPorter in https://github.com/DiamondLightSource/hdfmap/pull/13
* Add image data by DanPorter in https://github.com/DiamondLightSource/hdfmap/pull/14


**Full Changelog**: https://github.com/DiamondLightSource/hdfmap/compare/v0.7.0...v0.8

What's Changed
* Fixes and improvements by DanPorter in https://github.com/DiamondLightSource/hdfmap/pull/12
* Fixed hdfmap.get_scannables_array(), now returns correctly sized array
* Re-factored nexpy.py to default nexus paths is only called once, reducing number of duplicate warnings on certain files.
* minor fixes to docs.

**Full Changelog**: https://github.com/DiamondLightSource/hdfmap/compare/v0.6.3...v0.7.0

What's Changed
* Various improvements by DanPorter in https://github.com/DiamondLightSource/hdfmap/pull/11
* update to the way scannable names are generated if *scan_fields* exists in the NeXus file - all items are now saved by final part of the *local_name* attribute. This allows **nexus2srs** to provide matching column names for scanned data.


**Full Changelog**: https://github.com/DiamondLightSource/hdfmap/compare/v0.6.2...v0.6.3

added extended eval commands allowing default values and multiple names.

changes to eval_functions.py:
- added syntax 'name?(default)'
- added syntax 'name1|name2|name3'

See examples/example_metadata.py for a demonstration of the new evaluation options.

What's Changed
* Extended eval selection by DanPorter in https://github.com/DiamondLightSource/hdfmap/pull/10


**Full Changelog**: https://github.com/DiamondLightSource/hdfmap/compare/v0.6.1...v0.6.2

Minor update with several new features and small improvements

New Features
- dataset2str can now output units
- eval('s_name') will return the string representation of the data, including units
- map.find_groups(*names_or_classes) can be used to find groups based on several keywords
- map.find_dataset(*names_or_classes) can be used to find datasets based on several keywords

**Full Changelog**: https://github.com/DiamondLightSource/hdfmap/compare/v0.6.0...v0.6.1

What's Changed
* Diagnostic hdf features by DanPorter in https://github.com/DiamondLightSource/hdfmap/pull/8
* Auxilliary signals by DanPorter in https://github.com/DiamondLightSource/hdfmap/pull/9

New Features
Diagnostic HDF functions:
- `hdfmap.hdf_tree_string(file.nxs)` Returns multi-line string of HDF tree structure
- `hdfmap.hdf_tree_dict(file.nxs)` Returns dictionary of HDF tree structure
- `hdfmap.hdf_compare(file1.nxs, file2.nxs)` Compare two HDF files

Auxiliary Signal support
- `nexusmap.generate_scannables_from_nxdata()` now uses auxiliary_signals if available
- `nexusmap.get_plot_data` returns dict of useful plot data
- `nexusmap.eval('nameattr')` returns attribute from dataset
- `nexusmap.eval('__signal')`returns name of signal attribute
- `nexusmap.eval(`NXclass_name`) returns dataset /entry/groupNXclass/name
- `nexusmap.get_image_shape()` returns shape of a single image
- `nexusmap.get_scannables_shape()` returns shape of each scannable.

Changed functionality
- when there is no default NXdata, scannables will be generated from all unique dataset names, but will no longer include generated names like 'group_name' and 'NXclass_name'.
- when there are no 'local_name' attributes, metadata will be taken from all size<1 datasets using dataset name, no longer including generated names.

**Full Changelog**: https://github.com/DiamondLightSource/hdfmap/compare/v0.5.1...v0.6.0