LieAlgebras: Introduce WeightLattice#4344
Conversation
d55c9cd to
5d48fee
Compare
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #4344 +/- ##
==========================================
- Coverage 84.32% 84.31% -0.02%
==========================================
Files 649 651 +2
Lines 86430 86483 +53
==========================================
+ Hits 72880 72915 +35
- Misses 13550 13568 +18
|
|
can we get this merged? @fingolfin or someone else? |
|
Re-drafting until conflicts with #4339 are resolved. |
5d19aab to
faf6df4
Compare
|
CI is happy with my merge resolution. So this is again ready for merge |
| !!! warning | ||
| Weyl groups in OSCAR afford both left and right actions on roots and weights. | ||
| Note however, that **the left action** is the default (to align with the literature), | ||
| and all more complex functionality is defined with respect to the left action, e.g. | ||
| [`conjugate_dominant_weight_with_elem(::WeightLatticeElem)`](@ref). |
There was a problem hiding this comment.
This is none of my business, but I spent too much time transposing-matrices-until-the-result-looks-correct to not comment here:
So far, my understanding was that all group actions are from the right in OSCAR (at least by default) and we happily break textbook conventions at least in the linear algebra and invariant theory to have this consistent. Why not here?
There was a problem hiding this comment.
Short answer: this is still in discussion with Max (before this migrates to src/). This is the current state of things, and I wanted to document how it is right now (but subject to change). The removal of some *_left_* and *_right_* functions in this PR is to clean things up that partially made it possible to do both action directions, but this didn't seem like a healthy way forward (and it didn't think of all places where this was an issue). So I would like to clean this up again, further discuss with Max and afterwards tackle this more systematically.
joschmitt
left a comment
There was a problem hiding this comment.
The Aachen people seem to be happy with it.
Changes:
WeightLattice, a formal parent forWeightLatticeElemcharacterand friends now return aDict{WeightLatticeElem, T}(with defaultT = Int)preview: https://docs.oscar-system.org/previews/PR4344/Experimental/LieAlgebras/