[Lie algebras] misc bugfixes, root system in Lie alg interface, dim_of_simple_module implementation in julia
#4030
Conversation
lgoettgens
added
topic: LieAlgebras
experimental
lgoettgens
force-pushed
the
lg/Lie-algebra-misc
branch
from
3eecc49
to
746ac41
Compare
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #4030 +/- ##
==========================================
+ Coverage 84.56% 84.58% +0.02%
==========================================
Files 597 597
Lines 82197 82237 +40
==========================================
+ Hits 69508 69559 +51
+ Misses 12689 12678 -11
|
This reverts commit 0932db8 from oscar-system#3913.
| return L.root_system | ||
| end | ||
|
|
||
| function chevalley_basis(L::AbstractLieAlgebra) |
There was a problem hiding this comment.
It seems this function was only moved, why was the doc string deleted?
There was a problem hiding this comment.
The docstring is now attached to chevalley_basis(::LieAlgebra), as it contained nothing specific for the Lie algebra subtype
lgoettgens
force-pushed
the
lg/Lie-algebra-misc
branch
from
6c16446
to
2d1c537
Compare
|
|
||
| Return whether a root system for `L` is known. | ||
|
|
||
| This function should be implemented by subtypes that support root systems. |
There was a problem hiding this comment.
Another off-topic comment, but: perhaps we should agree to whom docstrings of exported (documented) functions are addressed. I'd say it should be users. Of course sometimes there is also a point in adding some remarks for implementors. But I think those then should be marked explicitly as such, e.g. via a callout (so the !!! syntax). That said, in this case I think one could just rephrase it slightly:
| This function should be implemented by subtypes that support root systems. | |
| This is only implemented for Lie algebras that support root systems. |
or so?
There was a problem hiding this comment.
Thanks for the remark. I collected the information about which functions need to be implemented by subtypes that support root systems to the general list about subtype-implemented functions to the top. Furthermore, the docstrings are adapted to reflect what users need to know. (Note that I don't have to mention subtypes with no root system support there, as in these cases has_root_system should always return false, i.e. the NotImplemented() should never trigger.). Changes in 7ea927d
…of_simple_module` implementation in julia (oscar-system#4030) * Add `!` to `set_root_system_type` * Add `has_root_system` to Lie algebra interface * Enhance some root system tests * Add `root_system` to Lie algebra interface * Add missing `@attr` * Implement `dim_of_simple_module` in julia (in some cases) * Revert "Skip one excessive test" This reverts commit 0932db8 from oscar-system#3913. * Test some more code paths in `dim_of_simple_module` * Move `dim_of_simple_module` to the root system * Remove `is_integral` * Copy fix for `id \in W` * Cache `cartan_symmetrizer` * Make root system interface a bit clearer * Update experimental/LieAlgebras/src/AbstractLieAlgebra.jl
contains:
@attr, typos)This contains all parts of #3980 that are not serialization related to allow for a quicker merge of the bugfixes also included there.