Configurable partitioning algorithm in mapping tester

[fsimonis]

Main changes of this PR

• Add configurable partitioning algorithm
• Make tests repeatable with topology-based partitioning

Author's checklist

• I used the pre-commit hook and used pre-commit run --all to apply all available hooks.
• I added a test to cover the proposed changes in our test suite.
• I updated the documentation in docs/README.md.
• I added a changelog entry in ./changelog-entries/ (create if necessary).
• I updated potential breaking changes in the tutorial precice/tutorials/aste-turbine.

[fsimonis]

I kept the default ´"meshfree"´ to be backwards-compatible. It is an open question if this is desirable though.

[davidscn]

That's ok.
precice-aste-partition --help lists available options. I would suggest to put a note for the kmeans variant, stating that the result is non-deterministic and not reproducible.

[fsimonis]

A comment in the code or a message in the script output?

[davidscn]

I would say "user visible", i.e., in the script output.

[davidscn]

I was more thinking of a comment here

aste/src/precice-aste-partition

Lines 163 to 167 in d2c6fae

	help="""Change the algorithm used for determining a partition.
	A meshfree algorithm works on arbitrary meshes without needing topological information.
	A topology-based algorithm needs topology information
	and is therefore useless on point clouds.
	A uniform algorithm will assume a uniform 2d mesh laid out somehow in 3d and partition accordingly.""",

In the end, the behavior is not specific to the mapping-tester, but to the partitioning, right?

[fsimonis]

Ah I see what you mean.

I think we need both as the help text needs to be read.

A warning in the partition script will fill the terminal when testing many rank combinations in the mapping tester.