Preparation for aiida 2.x compatibility

[JPchico]

In the develop branch most of the things should be compatible with aiida-2.x. But it needs to be thoroughly tested and make sure that no outstanding deprecation warnings are left. The idea of this issue is to act as a placeholder to ensure that we can discuss if anything needs to be done for this.

[JPchico]

@chrisjsewell How is the progress with this? I know that you did some changes regarding this and in my tests it seems to work in 2.x but I'm unsure on whether there is something else that needs to be done.

[chrisjsewell]

If you turned on the (off the top of my head) AIIDA_V3=1 when running pytest, I imagine you would see a number of deprecation warnings, for method renaming. But obviously these are not critical for compatibility

[JPchico]

cool cool, will explore this!

[chrisjsewell]

Actually it was AIIDA_WARN_v3 = 1

[chrisjsewell]

https://github.com/aiidateam/aiida-core/blob/3e4c8837aa6a4472482465177a13ed00c99d6dd0/pyproject.toml#L399

[chrisjsewell]

See also https://github.com/aiidateam/aiida-core/wiki/AiiDA-2.0-plugin-migration-guide (and the change log)

[JPchico]

I had found ti in the wiki when talking about migration to version 2 so no worries =) currently checking the deprecation warnings

[JPchico]

@chrisjsewell I think I have managed to get rid of most of the warnings. The only one that I have not managed to figure out what it trying to tell me is the following. Any idea?

orm/entities.py:175: AiidaDeprecationWarning: This property is deprecated, use `.collection` instead. (this will be removed in v3)
    warn_deprecation('This property is deprecated, use `.collection` instead.', version=3, stacklevel=2)

[JPchico]

Never mind I think I figured it out it seems to be that in the test utilities it was calling for Computer.objects.get(label=name) and it should have been Computer.collection.get(label=name)

[JPchico]

@chrisjsewell I have managed to get rid of all the deprecation warnings. However, there is an issue with the test for the md_multi

It seems to be trying to add a file to the retrieved?

        for stage, (step, rpath) in restart_map.items():
            with open(rpath, "rb") as handle:
                self.retrieved.base.repository.put_object_from_filelike(
                    handle, os.path.basename(rpath)
                )

Which is causing an Exception to be raised

 REPORT   aiida.orm.nodes.process.calculation.calcjob.CalcJobNode:process.py:564 [185|MdMultiCalculation|on_except]: Traceback (most recent call last):
  File "/usr/share/miniconda/envs/__setup_conda/lib/python3.8/site-packages/plumpy/process_states.py", line 228, in execute
    result = self.run_fn(*self.args, **self.kwargs)
  File "/usr/share/miniconda/envs/__setup_conda/lib/python3.8/site-packages/aiida/engine/processes/calcjobs/calcjob.py", line 473, in parse
    exit_code_retrieved = self.parse_retrieved_output(retrieved_temporary_folder)
  File "/usr/share/miniconda/envs/__setup_conda/lib/python3.8/site-packages/aiida/engine/processes/calcjobs/calcjob.py", line 564, in parse_retrieved_output
    exit_code = parser.parse(**parse_kwargs)
  File "/home/runner/work/aiida-lammps/aiida-lammps/aiida_lammps/parsers/lammps/md_multi.py", line 112, in parse
    self.retrieved.base.repository.put_object_from_filelike(
  File "/usr/share/miniconda/envs/__setup_conda/lib/python3.8/site-packages/aiida/orm/nodes/repository.py", line 234, in put_object_from_filelike
    self._check_mutability()
  File "/usr/share/miniconda/envs/__setup_conda/lib/python3.8/site-packages/aiida/orm/nodes/repository.py", line 68, in _check_mutability
    raise exceptions.ModificationNotAllowed('the node is stored and therefore the repository is immutable.')
aiida.common.exceptions.ModificationNotAllowed: the node is stored and therefore the repository is immutable.

Which seems to just be indicating that you cannot modify a node after it has been stored (which makes sense). However, I have no idea how this was working before if this was the case. As Nodes have always been immutable, there seems that there was a Force=True parameter which does not exist now, I wonder if that is the difference.

Any idea on how to solve this? Or should we just skip this? After all this calculation is going to be deprecated and replaced with the new calculation methods.

You can check my changes in my fork

[JPchico]

@chrisjsewell if you agree I will remove the tests for the md_multi (since this will be deprecated in the next release) and if everything passes then I think this should be ready to merge. Do you agree?

[JPchico]

Closed via #48 . If any other issues are detected they will be solved in individual issues.