CSR/CSC matrices and others can be non-canonical in that they have unsorted or even duplicate indices for a particular row, as well as entries that have an explicit "0" entry in a cell.
While I expect the csgraph methods work for unsorted indices, there should certainly be a test for the effect of duplicate indices (and I don't see one presently). In addition, at least connected_components treats explicit zeros as indicative of an edge (thus connected_components(X) for CSR X is not always equivalent to connected_components(X.A)), which should be documented.
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4