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
A list of open tasks, both more to less important ones, before the release of the package.
Testing
Linalg
Tests to be added for hoinetx.linalg
tests for incidence_matrix
tests for incidence_matrix_by_order
tests for incidence_matrices_all_orders
tests for laplacian_matrix_by_order
tests for laplacian_matrix_all_orders
tests for compute_multiorder_laplacian
tests for are_commuting
Measures
Tests to be added for hoinetx.measures.
there are some basic tests for sub_hypergraph_centrality. We still need to test the correctness of the final calculations. It would be ideal to construct some small hypergraphs, compute their analytical eigenvector and eigenvalues, and check that for these the function returns the correct sub-hypergraph centrality.
Features
Linalg
For all the Laplacian-related functions in hoinetx.linalg: do we need to add the return_mapping similar to other functions in linalg?
Similar question for function for all function split by order (e.g. incidence_matrix_by_order etc.). What is the correct mapping?
Documentation
Docstring and other stuff
in the documentation of Hypergraph maybe add the following observation? Nodes can be anything, and are given inside the edge list. However, they need to respect two conditions (as far as I can tell from the code):
- first, they need to be comparable inside a sorting operation inside every hyperedge (since hyperedges are sorted during the hypergraph creation)
- second, they need to be hashable, since they are keys of the Hypergraph._neighbours dictionary.
In practice, most immutable native Python types work, as well as any custom object that implements __hash__ (for hasing) and __ge__, __le__, etc. (for comparisons)
Open tasks
A list of open tasks, both more to less important ones, before the release of the package.
Testing
Linalg
Tests to be added for
hoinetx.linalg
incidence_matrix
incidence_matrix_by_order
incidence_matrices_all_orders
laplacian_matrix_by_order
laplacian_matrix_all_orders
compute_multiorder_laplacian
are_commuting
Measures
Tests to be added for
hoinetx.measures
.sub_hypergraph_centrality
. We still need to test the correctness of the final calculations. It would be ideal to construct some small hypergraphs, compute their analytical eigenvector and eigenvalues, and check that for these the function returns the correct sub-hypergraph centrality.Features
Linalg
hoinetx.linalg
: do we need to add thereturn_mapping
similar to other functions inlinalg
?incidence_matrix_by_order
etc.). What is the correct mapping?Documentation
Docstring and other stuff
Hypergraph
maybe add the following observation? Nodes can be anything, and are given inside the edge list. However, they need to respect two conditions (as far as I can tell from the code):- first, they need to be comparable inside a sorting operation inside every hyperedge (since hyperedges are sorted during the hypergraph creation)
- second, they need to be hashable, since they are keys of the
Hypergraph._neighbours
dictionary.In practice, most immutable native Python types work, as well as any custom object that implements
__hash__
(for hasing) and__ge__
,__le__
, etc. (for comparisons)The text was updated successfully, but these errors were encountered: