Speedup in GenericGraph.relabel() and two new options

[nathanncohen]

Because of some weird graphs built by ennemies of The Good Graph People, it seems that most of the work done by GenericGraph.relabel() is totally useless.
This patch adds two optional arguments to this method, which let the users enable/disable tests.

---

Apply:

1. attachment: trac_14000.patch
2. attachment: trac_14000-doctests.patch

CC: @stumpc5 @fchapoton

Component: graph theory

Keywords: relabelling

Author: Nathann Cohen

Reviewer: Anne Schilling

Merged: sage-5.8.beta2

Issue created by migration from https://trac.sagemath.org/ticket/14000

[nathanncohen]

Attachment: trac_14000.patch.gz

[anneschilling]

comment:2

Nathann, you need to put in some tests that show how the code works with the new inputs set to False. Also, please post a timing comparison before and after the patch.

Thanks,

Anne

[nathanncohen]

comment:3

O_o

Well... The code works the same way. It just does not check some things internally... The output does not change O_o

Nathann

[anneschilling]

comment:4

I know, but it is still good to put in a test that checks that it works the same ;-)

Replying to @nathanncohen:

O_o

Well... The code works the same way. It just does not check some things internally... The output does not change O_o

Nathann

[nathanncohen]

comment:5

Yep yep I'm doing it right now, but the point is that I am not totally sure that it does not produce a memory leak somewhere O_o

Nathann

[nathanncohen]

Attachment: trac_14000-doctests.patch.gz

[nathanncohen]

comment:6

A timing, just for the show :

Before :

sage: g = graphs.KneserGraph(11,4)
sage: %time g.relabel()           
CPU times: user 0.37 s, sys: 0.00 s, total: 0.37 s
Wall time: 0.37 s

After :

sage: g = graphs.KneserGraph(11,4)
sage: %time g.relabel()           
CPU times: user 0.01 s, sys: 0.00 s, total: 0.01 s
Wall time: 0.01 s

Ready for review again ! ;-)

Nathann

[anneschilling]

Author: Nathann Cohen

[anneschilling]

Changed keywords from none to relabelling

[anneschilling]

comment:8

Replying to @nathanncohen:

Yep yep I'm doing it right now, but the point is that I am not totally sure that it does not produce a memory leak somewhere O_o

What does this mean? Do you mean disabling the option is at the user's own risk?

[nathanncohen]

comment:9

What does this mean? Do you mean disabling the option is at the user's own risk?

Precisely ! I wrote that in the doctest itself :

"Checking that all vertices have an image can require some time, and this feature can be disabled (at your own risk)"

If you "rename" the vertices of a graph in such a way that two vertices end up having the same name at the end of the procedure, I can totally imagine that some data will be lost on the way :-P

Nathann

[anneschilling]

comment:10

Ok, then I am happy with your patch!

Anne

[anneschilling]

Reviewer: Anne Schilling

[nathanncohen]

comment:13

Ok, then I am happy with your patch!

Cool. Thanks for reviewing it ! :-D

Nathann

[jdemeyer]

Merged: sage-5.8.beta2