Cleanup and tests

[israelmcmc]

I'm reorganizing the cosipy module in preparation for the unit tests. Each module will have at least a test.

Work done so far:

• Removed cosipy/coordinates. That code is now in mhealpy or scoords
• Removed cosipy/config. That code is now in yayc
  • Added yayc to the dependencies, and changed the import where appropriate (DataIO and imaging modules)
• Moved cosipy_makeplots to DataIO (only module using it)
• Fix existing unit tests
  • The response module test was failing because the config file now uses the astromodels format for source specification
  • (IMC: actually still failing, but for another reason. Investigating.) The Data IO module was failing because it was missing the corresponding test data. I added a subset (621 event) of GalacticScan.inc1.id1.crab2hr.extracted(.testsample.)tra.gz and inputs_crab.yaml

Fixes #170

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 27.46%. Comparing base (d7924b8) to head (1222229). Report is 14 commits behind head on develop.

Files	Coverage Δ
cosipy/__init__.py	100.00% <ø> (+100.00%)	:arrow_up:
cosipy/_version.py	100.00% <ø> (ø)
cosipy/data_io/BinnedData.py	13.55% <100.00%> (+13.55%)	:arrow_up:
cosipy/data_io/DataIO.py	100.00% <100.00%> (+100.00%)	:arrow_up:
cosipy/data_io/ReadTraTest.py	16.04% <ø> (+16.04%)	:arrow_up:
cosipy/data_io/make_plots.py	13.63% <ø> (ø)
...mage_deconvolution/deconvolution_algorithm_base.py	19.81% <ø> (+19.81%)	:arrow_up:
cosipy/image_deconvolution/image_deconvolution.py	28.16% <100.00%> (+28.16%)	:arrow_up:
cosipy/image_deconvolution/modelmap.py	37.50% <ø> (+37.50%)	:arrow_up:
cosipy/response/FullDetectorResponse.py	39.61% <100.00%> (+39.61%)	:arrow_up:
... and 1 more

... and 22 files with indirect coverage changes

[ckarwin]

@israelmcmc I started adding more unit tests to dataIO, and I'm working from this branch. I noticed one small thing: the unit test for data fetching (test_data_fetching.py) writes the test file (test_file.txt) to the root directory (cosipy). Probably we want to remove the file at the end of the test script.

[israelmcmc]

@ckarwin Yes, good catch. I fixed it by using tempfile to create a system temporary folder that is automatically removed at the end of the tests. Since I was at it, I added a few test cases to have full coverage for that file.

However, due to what it seems an unrelated issue, the automatic pipelines are not failing. I think it has to do with new releases of some dependencies. I'll look into it, but they are passing on my machine.

[ckarwin]

@israelmcmc Yeah, everything is still passing on my machine as well, after pulling your latest commit.

[ckarwin]

@Yong2Sheng @israelmcmc Thanks, Yong, for approving the PR. So can it be merged soon, or still waiting on something? I see there is still a problem with some of the unit tests.

[israelmcmc]

Thanks @Yong2Sheng and @ckarwin. I'd like to have the test passing before merging. I think it's a problem with the dependencies, maybe I just need to specify the versions. I'll work on it.

[ckarwin]

Thanks @israelmcmc! Yeah, it would be helpful to have it merged whenever you can, to make it easier for adding more tests.

[israelmcmc]

In the meantime, people can branch off my branch (israelmcmc:cleanup_and_tests) instead of develop, and everything else is the same. But yet, I agree it's easier and better to just merge it, I'll get to it

[Yong2Sheng]

Hi @israelmcmc, do you mean to fork and make PR to your branch (israelmcmc:cleanup_and_test)? Thanks!

[israelmcmc]

@Yong2Sheng Kind off. Yes to making a new branch from my branch, but you don't need to fork my repo and you can open the PR directly to develop on this repo.

[ckarwin]

@israelmcmc I tested the error in the checks as follows:

1. I Copied your cleanup and tests branch into my forked repo.
2. I Made a PR #185, to replicate the error that you get.
3. Indeed, I get the same errors related to matplotlib.cm, relating to the discussion in issue #180.
4. I then updated install_requires in setup.py file with 'matplotlib<=3.8.3'. After doing this all checks passed.

So I think if you do the same in your setup file it should fix the issue. Can you try?

[ckarwin]

@israelmcmc Actually, I just made a PR into your branch with the update, if that's easier.

[israelmcmc]

Thanks, @ckarwin. That worked! Please go ahead and approve and merge this.

[ckarwin]

Excellent.