Remove deprecated record_by metadata

[pietsjoh]

Description of the change

This PR solves #79, removing the deprecated record_by metadata attribute for the affected file readers.

record_by was used in these cases to infer the signal_dimension, as the navigate attribute wasn't set for the axes.
Now navigate is set for all readers so that record_by can be removed.

Progress of the PR

• Remove record_by metadata attribute
  • blockfile
  • bruker
  • digitalmicrograph
  • edax
  • empad
  • image
  • jeol
  • mrc
  • msa
  • netcdf
  • ripple
  • tia
  • tiff
• add an changelog entry in the upcoming_changes folder
• add tests
  • bruker
  • digitalmicrograph
  • ripple
  • tiff
• ready for review.

[codecov]

Codecov Report

Attention: Patch coverage is 82.60870% with 4 lines in your changes missing coverage. Please review.

Project coverage is 85.01%. Comparing base (ff86cce) to head (2be342c).
Report is 548 commits behind head on main.

Files with missing lines	Patch %	Lines
rsciio/netcdf/_api.py	0.00%	2 Missing ⚠️
rsciio/mrc/_api.py	0.00%	1 Missing ⚠️
rsciio/ripple/_api.py	75.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #102      +/-   ##
==========================================
+ Coverage   85.00%   85.01%   +0.01%     
==========================================
  Files          73       73              
  Lines        9054     9062       +8     
  Branches     2049     2049              
==========================================
+ Hits         7696     7704       +8     
- Misses        892      893       +1     
+ Partials      466      465       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jlaehne]

Could you merge in main @pietsjoh to see whether #106 fixed the currently failing tests?

[ericpre]

This looks good.
Can please you add tests to check that the signal and navigation are correct for the following reader:

• bruker
• digitalmicrograph
• ripple
• tiff

There are no tests for mrc... but that should be sorted in a separate PR.

[pietsjoh]

Can please you add tests to check that the signal and navigation are correct for the following reader

Yes. However, this depends on the existing testfiles. For example tia has no tests for images/maps according to the reader. I will have a look at the tests/testfiles for these readers and I am going to try to cover the different axes sizes.

[ericpre]

@pietsjoh, do you want me to help with the rebase? The conflicts should come from #123 and the change to the path of the test files, etc..

[pietsjoh]

@ericpre yes, I am currently working on it

[pietsjoh]

I missread your comment. I want to try it by myself first and I will ask you when I cannot make it work

[pietsjoh]

@ericpre While rebasing it came to my attention that occasionally constructs such as

def test_example():
    with tempfile.TemporaryDirectory() as tmpdir:
        fname = os.path.join(tmpdir, "tiff_files", "filename.tif")
        s.save(fname, overwrite=True, export_scale=True)``

were replaced by

def test_example(tmp_path):
    s = hs.load(tmp_path / "test_rgba16.tif")

Why is this done only occasionally? Should this be replaced in all places instead?

[ericpre]

Indeed, this hasn't done everywhere because it is not possible to use the same pattern, see for example: #123 (comment) and #123 (comment).

There were already a lot of changes in that PR, so I have left this part as it was.

[pietsjoh]

ok, thanks for the clarification

[ericpre]

Thank you, that's great!