Reverse Cuthill-McKee and Maximum Bipartite Matching reorderings of sparse graphs #3751

nonhermitian · 2014-06-21T05:12:25Z

Added the symrcm function that takes a symmetric graph in CSR or CSC
sparse format and returns an array of row and column permutations that
reduces the overall bandwidth of the underlying matrix. This is useful for
direct and iterative sparse linear solvers.

Added the symrcm function that takes a symmetric graph in CSR or CSC sparse format and returns an array of row and column permutations that reduces the overall bandwidth of the underlying matrix.

coveralls · 2014-06-21T05:57:50Z

Changes Unknown when pulling 1c2106a on nonhermitian:master into * on scipy:master*.

coveralls · 2014-06-21T06:12:56Z

Changes Unknown when pulling 66960e0 on nonhermitian:master into * on scipy:master*.

coveralls · 2014-06-21T07:02:07Z

Changes Unknown when pulling 66960e0 on nonhermitian:master into * on scipy:master*.

argriffing · 2014-06-23T15:38:18Z

Looks good, as far as I can tell. This would close #3424.

rgommers · 2014-07-05T11:57:55Z

The name isn't very descriptive, needs to be expanded. Also, this algorithm has a fairly large overlap with csgraph.breadh_first_order. Maybe that function should grow a method keyword through which this algorithm can be selected, instead of adding this as a separate function?

rgommers · 2014-07-05T11:58:19Z

@jakevdp want to have a look at this?

rgommers · 2014-07-05T11:59:40Z

scipy/sparse/csgraph/_reordering.pyx

+include 'parameters.pxi'
+
+def symrcm(graph, symmetric_mode=False):
+    """


Docstring format: need a one-line summary, then a blank line. The first line gets picked up in listing in html docs.

argriffing · 2014-07-07T23:19:15Z

I'd be tempted to make the changes rgommers suggested into nonhermitian's PR except that the PR is on master rather than a separate branch.

nonhermitian · 2014-07-07T23:31:24Z

I have time tomorrow to make the changes. I will also add a maximal
bipartite matching method. Although all of these methods rely on BFS,
adding all of them as optional methods would be somewhat messy in my
opinion. One can think of many different orderings based on BFS.
On Jul 8, 2014 8:19 AM, "argriffing" notifications@github.com wrote:

I'd be tempted to make the changes rgommers suggested into nonhermitian's
PR except that the PR is on master rather than a separate branch.

—
Reply to this email directly or view it on GitHub
#3751 (comment).

@rgommers

- Added more descriptive name for RCM function - Made changes based on @rgommers suggestions. - Added non-symmetric maximum bipartite matching function.

coveralls · 2014-07-09T00:06:27Z

Changes Unknown when pulling 7e6d153 on nonhermitian:master into * on scipy:master*.

argriffing · 2014-07-09T00:18:22Z

Regarding the TravisCI test failure, this would be fixed if the indices in the bipartite matching test were explicitly dtype int32? Looking at https://github.com/scipy/scipy/blob/master/scipy/sparse/csgraph/parameters.pxi the dtype of csgraph indices seems to be restricted in this way. I don't know how easy it would be or what changes would be required to allow any integer index types.

coveralls · 2014-07-09T00:33:06Z

Changes Unknown when pulling 7e6d153 on nonhermitian:master into * on scipy:master*.

nonhermitian · 2014-07-09T00:53:17Z

I have explicitly set everything to dtype=int32. However, perhaps a fused type would work here.

argriffing · 2014-07-09T01:05:09Z

Yes fused types might work; these PRs/issues have some related discussion: #243 #2194

Edit: Now I think there is some subtle typing bug in the _reordering.pyx code, so I think it will need some doctoring like in that 243 PR, or rather as in the commit c89abe67ae6748 from some PR that I can't determine.

coveralls · 2014-07-09T09:38:02Z

Changes Unknown when pulling a1461b0 on nonhermitian:master into * on scipy:master*.

coveralls · 2014-07-09T10:00:52Z

Changes Unknown when pulling a1461b0 on nonhermitian:master into * on scipy:master*.

coveralls · 2014-07-09T10:18:03Z

Changes Unknown when pulling 818ef33 on nonhermitian:master into * on scipy:master*.

coveralls · 2014-07-09T11:17:46Z

Changes Unknown when pulling ed23297 on nonhermitian:master into * on scipy:master*.

argriffing · 2014-07-09T11:36:28Z

set type of row_match after multiplication

Good catch, I guess this was the problem all along. I'd suspected it but mistakenly ruled it out when I saw similar code in qutip like https://github.com/qutip/qutip/blob/master/qutip/cy/graph_utils.pyx#L234.

nonhermitian · 2014-07-09T11:39:24Z

QuTiP doesn’t test against NumPy 1.5.1 so I had to do the old trial and error solution method to find this.

On Jul 9, 2014, at 8:36 PM, argriffing notifications@github.com wrote:

set type of row_match after multiplication

Good catch, I guess this was the problem all along. I'd suspected it but mistakenly ruled it out when I saw similar code in qutip like https://github.com/qutip/qutip/blob/master/qutip/cy/graph_utils.pyx#L234.

—
Reply to this email directly or view it on GitHub.

argriffing · 2014-07-10T20:52:18Z

Looks OK, but I'll wait for feedback of more experienced csgraphers.

coveralls · 2014-07-13T05:33:32Z

Changes Unknown when pulling 6aed2e4 on nonhermitian:master into * on scipy:master*.

nonhermitian · 2014-07-15T02:20:41Z

I found a bug, so hold off on pulling.

- Also set length of temp_degrees array to the max node degree of the graph.

Test to verify node reordering works properly.

coveralls · 2014-07-15T14:36:55Z

Changes Unknown when pulling eb85936 on nonhermitian:master into * on scipy:master*.

ev-br · 2014-08-29T10:17:47Z

What is the status of this PR?

pv · 2014-08-29T10:41:21Z

The code looks good to me --- I assume the bug mentioned by @nonhermitian above was fixed in eb85936.

One question is using ITYPE_t which is int32 --- but it seems this is used for values with maximum value of shape[0] rather than nnz.

It would actually be very interesting if this helps with gh-3831

nonhermitian · 2014-08-30T11:35:37Z

Yes, the error in the RCM code was fixed in the last commit. As it turns out, all of the RCM tests had graphs with nodes of the same degree, so sorting wasn't being tested. In the last commit is an example that is more general, and that caught the bug.

ENH: sparse.csgraph: Reverse Cuthill-McKee and Maximum Bipartite Matching reorderings of sparse graphs

pv · 2014-08-31T14:13:32Z

Merging, LGTM.

nonhermitian added 2 commits June 21, 2014 14:07

Reverse Cuthill-McKee reordering of sparse graphs

42d79ae

Added the symrcm function that takes a symmetric graph in CSR or CSC sparse format and returns an array of row and column permutations that reduces the overall bandwidth of the underlying matrix.

Fixed PEP8 issue

1c2106a

Added original source information

66960e0

pv added the PR label Jun 26, 2014

pv mentioned this pull request Jul 2, 2014

add iterative methods and reordering algorithms numpy/numpy#4835

Closed

argriffing added the scipy.sparse.csgraph label Jul 3, 2014

rgommers reviewed Jul 5, 2014
View reviewed changes

reverse_cuthill_mckee & maximum bipartite matching

c4d1af2

- Added more descriptive name for RCM function - Made changes based on @rgommers suggestions. - Added non-symmetric maximum bipartite matching function.

nonhermitian changed the title ~~Reverse Cuthill-McKee reordering of sparse graphs~~ Reverse Cuthill-McKee and Maximum Bipartite Matching reorderings of sparse graphs Jul 8, 2014

Fixed documentation error in reverse_cuthill_mckee

7e6d153

set dtype=int32 in maximum_bipartite_matching test

dcf48ed

nonhermitian added 2 commits July 9, 2014 18:16

Added int64 indices tests

3ba0467

set np.ones explicitly to correct type

a1461b0

set type after multiplication

818ef33

set type of row_match after multiplication

ed23297

nonhermitian added 2 commits July 13, 2014 13:49

Removed unnecessary cimport of Cython

c3cd734

Fixed RCM int64 test

6aed2e4

nonhermitian added 3 commits July 15, 2014 22:16

Fix for unsorted nodes

0956ee6

- Also set length of temp_degrees array to the max node degree of the graph.

Added RCM test

7c28dcd

Test to verify node reordering works properly.

call np.max instead of max in rcm

eb85936

pv removed the PR label Aug 13, 2014

nonhermitian mentioned this pull request Aug 29, 2014

SuperLU is too slow #3831

Closed

pv added a commit that referenced this pull request Aug 31, 2014

Merge pull request #3751 from nonhermitian/master

daa7087

ENH: sparse.csgraph: Reverse Cuthill-McKee and Maximum Bipartite Matching reorderings of sparse graphs

pv merged commit daa7087 into scipy:master Aug 31, 2014

pv added this to the 0.15.0 milestone Aug 31, 2014

perimosocordiae mentioned this pull request Oct 12, 2015

reverse Cuthill–McKee algorithm (RCM) for sparse matrices #3424

Closed

jakelishman mentioned this pull request Mar 13, 2020

Consider removing qutip.graph functions in favour of scipy versions qutip/qutip#1203

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Reverse Cuthill-McKee and Maximum Bipartite Matching reorderings of sparse graphs #3751

Reverse Cuthill-McKee and Maximum Bipartite Matching reorderings of sparse graphs #3751

nonhermitian commented Jun 21, 2014

coveralls commented Jun 21, 2014

coveralls commented Jun 21, 2014

coveralls commented Jun 21, 2014

argriffing commented Jun 23, 2014

rgommers commented Jul 5, 2014

rgommers commented Jul 5, 2014

rgommers Jul 5, 2014

argriffing commented Jul 7, 2014

nonhermitian commented Jul 7, 2014

coveralls commented Jul 9, 2014

argriffing commented Jul 9, 2014

coveralls commented Jul 9, 2014

nonhermitian commented Jul 9, 2014

argriffing commented Jul 9, 2014

coveralls commented Jul 9, 2014

coveralls commented Jul 9, 2014

coveralls commented Jul 9, 2014

coveralls commented Jul 9, 2014

argriffing commented Jul 9, 2014

nonhermitian commented Jul 9, 2014

argriffing commented Jul 10, 2014

coveralls commented Jul 13, 2014

nonhermitian commented Jul 15, 2014

coveralls commented Jul 15, 2014

ev-br commented Aug 29, 2014

pv commented Aug 29, 2014

nonhermitian commented Aug 30, 2014

pv commented Aug 31, 2014

Reverse Cuthill-McKee and Maximum Bipartite Matching reorderings of sparse graphs #3751

Reverse Cuthill-McKee and Maximum Bipartite Matching reorderings of sparse graphs #3751

Conversation

nonhermitian commented Jun 21, 2014

coveralls commented Jun 21, 2014

coveralls commented Jun 21, 2014

coveralls commented Jun 21, 2014

argriffing commented Jun 23, 2014

rgommers commented Jul 5, 2014

rgommers commented Jul 5, 2014

rgommers Jul 5, 2014

Choose a reason for hiding this comment

argriffing commented Jul 7, 2014

nonhermitian commented Jul 7, 2014

coveralls commented Jul 9, 2014

argriffing commented Jul 9, 2014

coveralls commented Jul 9, 2014

nonhermitian commented Jul 9, 2014

argriffing commented Jul 9, 2014

coveralls commented Jul 9, 2014

coveralls commented Jul 9, 2014

coveralls commented Jul 9, 2014

coveralls commented Jul 9, 2014

argriffing commented Jul 9, 2014

nonhermitian commented Jul 9, 2014

argriffing commented Jul 10, 2014

coveralls commented Jul 13, 2014

nonhermitian commented Jul 15, 2014

coveralls commented Jul 15, 2014

ev-br commented Aug 29, 2014

pv commented Aug 29, 2014

nonhermitian commented Aug 30, 2014

pv commented Aug 31, 2014