move analysis.x3dna to new analysis.legacy module #906
Comments
|
So how will the namespace work if I want to use analysis.x3dna in the future? Will I have to import from I'm not a huge fan of fragmenting namespace like this - could we instead add a warning on import saying words to the effect of, "This is legacy with no tests, double check your results if you use this code" |
Yes, that was the idea, as discussed. The reason for why at least I favored this solution is because it makes it REALLY clear (semantically) that this package is “use at your own risk” – even before you import it. I like the idea of the user having to go the extra mile to import from legacy – should make them think, and should make clear that these packages are not first class citizens.
Fair enough (“flat is better than nested”) although in this instance I’d consider it justified.
I would ALSO add this warning. But warnings are easily ignored/muted. Anyway, my $0.02. Oliver Beckstein * [email protected] |
- created new MDAnalysis.analysis.legacy package - directory - docs - x3dna is now considered legacy code (no/minimal testing, no maintenance) - see #906 for reasons (mainly because the x3dna license does not allow us to just install it for testing on travis, or rather, licensing and access to x3dna is unknown/too complicated) - closes #906 - see #743 on background for legacy code and also https://github.com/MDAnalysis/mdanalysis/wiki/Style-Guide#tests-for-mdanalysisanalysis - added stub with deprecation warning: until release 1.0, MDAnalysis.analysis.x3dna is still accessible - added test case that the stub is there
* moved analysis.x3dna to analysis.legacy.x3dna - created new MDAnalysis.analysis.legacy package - directory - docs - x3dna is now considered legacy code (no/minimal testing, no maintenance) - see #906 for reasons (mainly because the x3dna license does not allow us to just install it for testing on travis, or rather, licensing and access to x3dna is unknown/too complicated) - closes #906 - see #743 on background for legacy code and also https://github.com/MDAnalysis/mdanalysis/wiki/Style-Guide#tests-for-mdanalysisanalysis - added stub with deprecation warning: until release 1.0, MDAnalysis.analysis.x3dna is still accessible - added test case that the stub is there * updated CHANGELOG - legacy module (#743) - x3dna is now legacy code (#906) - updated upcoming release number to 0.16.0 [skip ci] * Removed legacy submodule from coverage
* moved analysis.x3dna to analysis.legacy.x3dna - created new MDAnalysis.analysis.legacy package - directory - docs - x3dna is now considered legacy code (no/minimal testing, no maintenance) - see MDAnalysis#906 for reasons (mainly because the x3dna license does not allow us to just install it for testing on travis, or rather, licensing and access to x3dna is unknown/too complicated) - closes MDAnalysis#906 - see MDAnalysis#743 on background for legacy code and also https://github.com/MDAnalysis/mdanalysis/wiki/Style-Guide#tests-for-mdanalysisanalysis - added stub with deprecation warning: until release 1.0, MDAnalysis.analysis.x3dna is still accessible - added test case that the stub is there * updated CHANGELOG - legacy module (MDAnalysis#743) - x3dna is now legacy code (MDAnalysis#906) - updated upcoming release number to 0.16.0 [skip ci] * Removed legacy submodule from coverage
As discussed in #898, we cannot easily write full tests for the
MDAnalysis.analysis.x3dnamodule because we cannot easily install x3dna on travis (and I don't even know the terms of the license — impossible to find on the home page!). In accordance with the Guidelines on Testing in MDAnalysis.analysis I propose toMDAnalysis.analysis.legacymodulelegacyfrom testing/coverageMDAnalysis.analysis.x3dnatoMDAnalysis.analysis.legacy.x3dnaThe text was updated successfully, but these errors were encountered: