openms 2.8

[bernt-matthias]

Significant changes:

• All data parameters with multiple="true" are now in a conditional that allows to switch between true and false (default for true)
  • So far this was solved with a patch for PSMFeatureExctractor
  • But was requested also for PercolatorAdaptor https://github.com/galaxyproteomics/tools-galaxyp/issues/566#issue-813752018
• select parameters only use compatible display attributes (radio / checkboxes)
• Tests are now written directly to the tool's tests node. This should speed up tool loading quite a bit because the large test macro files don't need to be (xml-)parsed for each tool.
• Advanced options are now implemented as (folded) section instead of a conditional (see https://github.com/galaxyproteomics/tools-galaxyp/issues/566#issue-813752018). Advantage: for instance in the OpenSwathWorkflow the out_chrom output can be enabled (out_chrom_FLAG) but the out_chrom_type is not known in cheetah if the advanced options are disabled.
• containerized tests

TODOs/Ideas

• [x] Move additional files (shell scripts etc) to subdir and exclude from shed upload
• [x] add mzqc datatype to Galaxy as subclass of Json https://github.com/galaxyproject/galaxy/pull/13564
  • [x] implies profile 22.01
• [x] some booleans are not rendered as booleans but string parameters https://github.com/OpenMS/OpenMS/issues/5880
  • [x] maybe hardcode as bool for those fixed in the corresponding PR
• [x] use re_match for comparisons .. for this all (input) file names in output data would need to be replaced by .* (or something more specific .. [a-zA-Z0-9._-]+?) and the other file content needs to be regex quoted [re.escape] (https://docs.python.org/3/library/re.html#re.escape)
  • seems too difficult at the moment (since the same file may be used as in and output) ..
• [x] Fix failing tools
  • [x] Decharger test failing because of https://github.com/OpenMS/OpenMS/pull/5920 ?
  • [x] DecoyDatabase (symlink generation seems faulty)
  • [x] 2 FalseDiscoveryRate tests fail if -force is blacklisted https://github.com/OpenMS/OpenMS/issues/5921
  • readd -force
  • [x] FeatureFinderIdentification test failing https://github.com/OpenMS/OpenMS/pull/5917 .. maybe blacklist parameter?
  • [x] ProteinQuantifier .. just needs larger delta_frac test file is small and the file name is included

[bgruening]

restarted the tests

[bernt-matthias]

No idea why the test fails. Can't reproduce this locally.

[bernt-matthias]

All, but linting succeeded. The linter errors should be fixed when the 22.01 galaxy python packages are available. I can wait for this I guess.

Not sure if I should increase the profile, for the mzqc datatype that became available in 22.01...

[bgruening]

I'm fine with merging if you are.

[bernt-matthias]

@bgruening I think its good as long as you agree with keeping the profile (correct would be 22.01 because of the mzqc datatype .. but maybe that is not important enough .. and 22.01 is maybe too recent).

Since linting will fail we will need to push to the TS manually, or?