avoid using .vertices() in comparability, hyperbolicity and distances_all_pairs

[dcoudert]

We expose the new parameter vertex_list of init_short_digraph in methods of distances_all_pairs.pyx and so

• add the same parameter to methods of distances_all_pairs.pyx
• remove the use of .vertices() in hyperbolicity.pyx and comparability.pyx

We could do the same in bandwidth.pyx but this might be in conflict with #26520

Several use of .vertices() remain in distances_all_pairs.pyx that are harder to remove.

[dcoudert]

Hope you like it ;)

---

New commits:

​90c83bc

trac #26528: reduce use of .vertices() in distances_all_pairs

[tscrim]

I am not convinced that the test for is_transitive is always going to give the same answer:

         sage: digraphs.DeBruijn(5,2).is_transitive(certificate=True)
-        ('00', '10')
+        ('22', '02')

I think a more robust test would be to verify this by:

sage: G = digraphs.DeBruijn(5,2)
sage: cert = G.is_transitive(certificate=True)
sage: G.has_edge(*cert)
False
sage: G.shortest_path(*cert) != []
True

[tscrim]

Also, I think it would be better to do `0..n-1` -> `(0, \ldots, n-1)`. Moreover, you need to change `i`th -> `i`-th because Sphinx would not treat the closing backtick like it should.

Otherwise LGTM.

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​564cb70

trac #26528: reviewer's comments

[dcoudert]

Thanks for the comments.

[tscrim]

Thanks. LGTM.