clean generic_graph.py (part 6)

[dcoudert]

Done here:

• correct method eulerian_orientation for graphs without edge (was not using the right set of vertices)

  • Should we move this method to orientation.py ?

• some care in min_spanning_tree, but we have remaining issues:

  • algorithm "Prim_fringe" compare vertex labels. I added hash(..) around, but this is certainly not enough.
  • the output of min spanning tree methods is sorted. As said in clean spanning_tree.pyx #26547, there is no reason for that.

• avoid a call to .edges in cycle_basis

CC: @tscrim @fchapoton @seblabbe

Component: graph theory

Author: David Coudert

Branch/Commit: 7766d8e

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/26637

[dcoudert]

Branch: public/26637_generic_graph_part_6

[dcoudert]

New commits:

7766d8e

trac #26637: part 6

[dcoudert]

Commit: 7766d8e

[tscrim]

comment:2

LGTM.

[tscrim]

Reviewer: Travis Scrimshaw

[jhpalmieri]

comment:3

Is there a metaticket for files in graphs? I'm particularly interested in more on generic_graph.py, especially is_isomorphic, and also base/sparse_graph.pyx, especially iterator_edges. Both of these try to sort vertices, which cause Python 3 doctest failures in homology/simplicial_complexes.py and homology/examples.py.

[dcoudert]

comment:4

Let us use #26640 as meta ticket. Should have been opened earlier. Feel free to modify it as I don't know how to organize it.

[vbraun]

Changed branch from public/26637_generic_graph_part_6 to 7766d8e