You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[doc] Add plus_first and plus_second semirings to docs #497
The plus_first and plus_second semirings are very useful for writing algorithms that only care about the nonzero structure of the matrix because they can work on any user matrix without copy or modification.
Agreed 💯, they're very useful and should be in the online docs.
Related issue #410, which also include a partial implementation.
To explore the API of python-graphblas, using tab-completion in IPython or Jupyter is very nice, but I know everybody doesn't use this workflow, and our docs should be complete regardless.
Btw, we aim to expose 100% of features from SuiteSparse:GraphBLAS. I regard missing functionality as a bug.
Good point, Jupyter tab completion does pick up a lot of this.
My workflow is generally PyCharm's tab completion, which is generally very good but based on static analysis instead of Jupyter's runtime introspection.
python-graphblas generally does not play well with static analysis at all, mostly due to the way the types are moved around at import time. Even gb.Matrix is not found. In my own projects a combination of defining things in __all__ and using type hints goes a long way, but i understand that most python-graphblas members are just a forward of whatever is exposed from C.
Activity
eriknw commentedon Sep 5, 2023
Agreed 💯, they're very useful and should be in the online docs.
Related issue #410, which also include a partial implementation.
To explore the API of
python-graphblas, using tab-completion in IPython or Jupyter is very nice, but I know everybody doesn't use this workflow, and our docs should be complete regardless.Btw, we aim to expose 100% of features from SuiteSparse:GraphBLAS. I regard missing functionality as a bug.
alugowski commentedon Sep 7, 2023
Good point, Jupyter tab completion does pick up a lot of this.
My workflow is generally PyCharm's tab completion, which is generally very good but based on static analysis instead of Jupyter's runtime introspection.
python-graphblas generally does not play well with static analysis at all, mostly due to the way the types are moved around at import time. Even gb.Matrix is not found. In my own projects a combination of defining things in
__all__and using type hints goes a long way, but i understand that most python-graphblas members are just a forward of whatever is exposed from C.Document the plus_first and plus_second semirings (GH #497)
Document the plus_first and plus_second semirings (GH #497)
Docs: fix examples, fill API reference gaps, document gotchas (#602)