Release 1.0

[richardjgowers]

I think we're due another release. I think the only ongoing thing is @lilyminium 's work on docs, so ideally this release captures all changes necessary to the package for their doc work. (ie if this is an awful idea tell me and we can postpone).

Looking at #2303 (esp #2303 (comment) ) it looks like the plan is that this release is called 1.0 and is the last python 2.7 release. Work then continues (supporting only py 3) as 1.x until we finish the annoying API change stuff and we call the result 2.0 (and drop py 2)

following feedback, todo:

• flags @richardjgowers
• ncdf @IAlibay
• tpr jazz @jbarnoud
• Modifications to ITPParser #2408 @lilyminium
• added TopologyObject API #2382 @lilyminium
• instant selectors removal @richardjgowers

[IAlibay]

As previously discussed with @richardjgowers I'd like to at least fix the NCDF reader portion of #2323 and #2327 before release 1.0.

I'll hopefully get this done within the next couple of weekends if that's ok with your schedule?

[richardjgowers]

[jbarnoud]

With gromacs 2020 released, I'd like to see #2431 merged before 1.0. Especially as, in the current state there will be unreadable errors as the TPR parser expects a TPR from gromacs 2020-beta that are very different.

[lilyminium]

Sounds good! Will you be going through and looking at currently open PRs before release? There's some GSOC stuff and I would prefer to have #2408 in the release, as the current ITPParser is very flawed. #2382 would be nice too.

[orbeckst]

In an ideal world we would have everything ready for 1.0.

But the things mentioned above should be in 1.0 and we also wanted to remove the flags #782 . It would be easier to make this 0.21 and then follow up with 1.0. A release is overdue, and it might as well be 0.21 and then we follow up with 1.0.

Btw, @richardjgowers – you looked at the wrong comment, the last one was #2303 (comment) and the official roadmap is the (ambitious) https://www.mdanalysis.org/2019/11/06/roadmap/.

1.0 will be similar to upcoming 0.21 (i.e., no major annoying API breaks but clean-up and deprecations)

[richardjgowers]

Ok thanks @orbeckst so we'll see what we can get finished in a week or two then make a release, maybe it's 1.0 maybe it's 0.21.

[orbeckst]

Ok, sounds fair.

[xiki-tempula]

Sorry for trying to rush this in the last minute #2480. I have replaced the distance search with capped_distance, so it should provide a decent speed boost.

I will do the bond search in a later date after I get familiar with #2382 as I don't expect it provide much speed boost.

[IAlibay]

@richardjgowers

Re: netcdf it looks like if we aren't going for an API complete version 1.0, the only thing needed here is to remove the degrees keyword on the reader. I'll deal with adding scale_factor support to the writer in a future version.

Additionally, as mentioned in #2479 there's a few things that need to be removed on time for v1.0 (currently making my way through this and will make appropriate PRs over the next few days):

Whole file removal (#2481)

• core/distances.py
• analysis/x3dna.py
• core/log.py
• core/parallel.py
• core/qcprot.py
• core/transformations.py
• core/units.py
• core/utils.py
• core/KDTree.py
• at least one test from test_transformation.py

Analysis changes

• save() and store() analysis methods and certain instances of filename remove save() methods from analysis classes #1745 (PR Remove save() methods for v1.0 - Part 1 #2487 and Remove save part2 #2494)
• extra kwags in analysis/base.py
• waterdynamics t0, tf, and dtmax (question for this, it was set for deprecation but no removal version was set, is v1.0 reasonable?)
• heuristic test in test_hbonds.py (PR Remove save part2 #2494)

Reader changes (#2481)

• degrees in the NCDFReader (Removes stubs files for v1.0 #2481)
• skip and format keywords to :meth:DCDReader.timeseries. (Removes skip and format from DCD reader #2501)

AtomGroup stubs (#2562)

• Remove stubs from core/AtomGroup.py.

Memory reader

• Remove format keyword from :meth:MemoryReader.timeseries.

Others

• test_generated_residueselection in test_segment.py
• FullSelgroupSelection in core/selection.py (PR FullGroup selection keyword removal #2500)
• instance selector in __getitem__ in :class:AtomGroup core/groups.py
• test_chainid_quick_select in core/test_universe.py
• test_segments in core/test_atomgroup.py

Since it doesn't seem clear if we are going for v1.0, I'll start the PRs for these and if they don't lead to anything, we could either keep them around for when we do go v1.0, or just drop em and re-open new PRs next time around?