-
Notifications
You must be signed in to change notification settings - Fork 8
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Sectional curvature #184
Sectional curvature #184
Conversation
This looks very nice! Thanks for adapting that. If the test coverage is fine, I am ok with this. |
Question: should |
Why is that harder for the Riemannian tensor? |
The formula |
I think the clearer way would be “we do not use linearly dependent vectors”? |
That's how it currently works. |
Ah, yes. I see. And it is also well-documented that way. I think it is ok to keep it that way. |
OK, then we need a test for linear independence in |
Or we can just leave that not implemented for now. |
I would thoroughly document that we assume they are linearly independent – and for now not implement a check. |
It doesn't work for product manifolds though. You may have two vectors that are linearly independent, but their projections on one or more of the component manifolds are not linearly independent, so that needs to be handled somewhere since sectional curvature for them is well-defined. |
But then one could check that on every “manifold factor” and only add that to the min if they are linearly independent and ignore them otherwise? |
Yes, that's what would be needed. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Overall this looks fine to me. Just two small remarks.
Co-authored-by: Ronny Bergmann <git@ronnybergmann.net>
This looks good so far. I am just missing the good old code coverage – did we miss to update the GitHub action? I remember something like this in a recent PR either on Manopt or Manifolds as well that I had to specifically add the secret (again) to the action. |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #184 +/- ##
=======================================
Coverage 99.96% 99.96%
=======================================
Files 27 27
Lines 3107 3156 +49
=======================================
+ Hits 3106 3155 +49
Misses 1 1 ☔ View full report in Codecov by Sentry. |
Right, I somehow didn't notice the lack of coverage report. It works now. I think we can merge this now? |
Adapted from Manopt.jl, with added functions for lower and upper bounds.