Polish the use of iterators in C graphs

[rlmill]

http://groups.google.com/group/sage-devel/browse_thread/thread/2bdfa75d401575f2

CC: @kliem @tscrim

Component: graph theory

Author: David Coudert

Branch/Commit: c7dc1b1

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/6604

[rlmill]

comment:3

I am right now watching a talk on closures in Cython, which are all but finished. Then, yield statements in Cython will soon follow! So don't work on this ticket for now...

[dcoudert]

Branch: public/graphs/6604_iterators

[dcoudert]

comment:9

It seems there is little remaining to do for iterators in backends now, so let's do it.

Before

sage: a = DiGraph(graphs.CompleteGraph(100))
sage: b = a._backend
sage: %timeit L = list(b.iterator_nbrs(30))
28.3 µs ± 1.44 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)

After

sage: a = DiGraph(graphs.CompleteGraph(100))
sage: b = a._backend
sage: %timeit L = list(b.iterator_nbrs(30))
18.6 µs ± 698 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)

---

New commits:

07d52d9

trac #6604: polish some iterators in backends

[dcoudert]

Commit: 07d52d9

[dcoudert]

Author: David Coudert

[tscrim]

Reviewer: Travis Scrimshaw

[tscrim]

comment:10

That is quite a decent speedup and will likely have effects in other methods. Looks like some trivial failures due to slight differences in how the iteration is done:

sage -t --long --random-seed=210304412354021215508062896360889570175 src/sage/graphs/generic_graph.py
**********************************************************************
File "src/sage/graphs/generic_graph.py", line 10654, in sage.graphs.generic_graph.GenericGraph.neighbor_iterator
Failed example:
    list(D.neighbor_iterator(0))
Expected:
    [1, 2, 3]
Got:
    [3, 1, 2]
**********************************************************************
File "src/sage/graphs/generic_graph.py", line 17591, in sage.graphs.generic_graph.GenericGraph.?
Failed example:
    list(D.depth_first_search(1, ignore_direction=True, edges=True))
Expected:
    [(1, 4), (4, 5), (5, 6), (5, 2), (4, 3)]
Got:
    [(1, 3), (3, 4), (4, 5), (5, 6), (5, 2)]

This one I am not sure if it is because of the random seed or this ticket:

sage -t --long --random-seed=210304412354021215508062896360889570175 src/sage/schemes/toric/sheaf/klyachko.py
**********************************************************************
File "src/sage/schemes/toric/sheaf/klyachko.py", line 950, in sage.schemes.toric.sheaf.klyachko.KlyachkoBundle_class.random_deformation
Failed example:
    Vtilde.cohomology(dim=True, weight=(0,))
Expected:
    (1, 0)
Got:
    (0, 0)

in either case, it should be a trivial fix.

[sagetrac-git]

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

c7dc1b1

trac #6604: fix doctests in genenric_graph.py

[sagetrac-git]

Changed commit from 07d52d9 to c7dc1b1

[dcoudert]

comment:12

I fixed the doctests in generic_graph.py.

I cannot reproduce the error in klyachko.py. It occurs in method random_deformation. Some help might be helpful here.

[dcoudert]

comment:13

Actually, this random error in klyachko.py has already been reported in #32773.

[tscrim]

comment:14

Thank you. LGTM.

[vbraun]

Changed branch from public/graphs/6604_iterators to c7dc1b1