Introduce has_components for Euclidean

[kellertuer]

Since we now have available that

distance(Euclidean(3), p, q, 1)

for the -norm, this introduces it also for Euclidean. I do not directly see other manifolds that would be available for that.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 96.39%. Comparing base (5d28783) to head (69d1704).
Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #759   +/-   ##
=======================================
  Coverage   96.38%   96.39%           
=======================================
  Files         123      123           
  Lines       11438    11441    +3     
=======================================
+ Hits        11025    11028    +3     
  Misses        413      413           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[kellertuer]

Currently Has_components does not “pass transparently” through MetricManifold. Should we add that to ManifoldsBase?

To be precise, we currently have that

M = Manifolds.MetricManifold(E, Manifolds.EuclideanMetric())

yields has_components(M) to be false. I think if we fix that we have to use traits – and this would have to be done in ManifoldsBase.jl I think.

[mateuszbaran]

Nice. Could you also add tests that distance and norm work with r?

Currently Has_components does not “pass transparently” through MetricManifold. Should we add that to ManifoldsBase?

Right, yes, has_components should be a trait function with transparent passing.

[kellertuer]

sure I can add those lines.

For the transparency, we should probably do that in ManifoldsBase?

[mateuszbaran]

Yes, transparency should be in ManifoldsBase.jl.

[kellertuer]

Sure, will do.

[kellertuer]

Added the test for distance and norm. Also started the ManifoldsTestSuite. Similar to Lie groups I will try to now lower-level add functions therein and slowly transfer to using the test suite then. Yes that means I tweak that into other PRs a bit every now and then, but that is easier for me, I think, than the current open PR for that.

[mateuszbaran]

Sure, that's fine.

[mateuszbaran]

BTW, I haven't tagged 0.10.4 yet so this could be a part of it together with Oliver's PR.

[kellertuer]

Just saw that 0.10.4 is not yet registered, then this can also be part of that. But I will do the transparency first in ManifoldsBase maybe.

[kellertuer]

Already merged master here, but since we just merged a branch in 3 packages, all CIs might take a time ;) afterward I will merge this as well. We could register those two as 0.10.4 for sure.