Missing copy() method in BipartiteGraph

[rhinton]

BipartiteGraph? is missing a copy() method.

sage: bg = BipartiteGraph(graphs.CycleGraph(4))
sage: type(bg)
<class 'sage.graphs.bipartite_graph.BipartiteGraph'>
sage: bg2 = bg.copy()
sage: type(bg2)
<class 'sage.graphs.graph.Graph'>

The result is not horrendous because the base Graph class has a copy() method. But the result is unexpected: copy() a BipartiteGraph? and you get a Graph?

[rhinton]

Oops, bug in doctest, the patch doesn't work at all.

[rhinton]

New patch (replaced old) should work, including fixed doctest.

[rlm]

This is item 9 on trac #1941.

[rhinton]

Right, I had forgotten about this. Obviously I'm fixing methods as I need them. I don't have time right now for the complete overhaul. Is the piecemeal approach OK?

Incidentally, I think I have a decent approach for add_vertex. I was planning on opening another ticket.... :-)

[rlm]

Replying to rhinton:

Right, I had forgotten about this. Obviously I'm fixing methods as I need them. I don't have time right now for the complete overhaul. Is the piecemeal approach OK?

I was just linking the tickets to each other. Adding edges to trac I guess... Any way you want to improve bipartite graphs I'll gladly review once I get the time, I just wanted to make sure you were aware of that ticket.

Incidentally, I think I have a decent approach for add_vertex. I was planning on opening another ticket.... :-)

Go for it!

[rhinton]

Note for someone who is more knowledgeable: the new BipartiteGraph?.copy() method doesn't try to do anything tricky with vertex associations, boundaries, positions, etc. It just tries the Graph constructor. I haven't figured out a way to copy() the Graph part and get out a BipartiteGraph? instance. This way works for me, but I'm not sure if it does all the special stuff that Graph.copy() does.

[rhinton]

I get it ... adding edges. :-)

Replying to rlm:

Replying to rhinton:

Right, I had forgotten about this. Obviously I'm fixing methods as I need them. I don't have time right now for the complete overhaul. Is the piecemeal approach OK?

I was just linking the tickets to each other. Adding edges to trac I guess... Any way you want to improve bipartite graphs I'll gladly review once I get the time, I just wanted to make sure you were aware of that ticket.

[rhinton]

The new patch to GenericGraph?.copy() always creates a new object of the same class as the original object -- no special case code for Graph, DiGraph?, or BipartiteGraph?. Any future subclasses will work as well assuming they can handle the same initializer arguments (by passing them to the superclass initializer).

[rlm]

This patch exposes a bug in another seldom used extension of the Graph class, GraphBundle.

File "/Users/rlmill/sage-4.3.3/devel/sage-main/sage/graphs/graph_bundle.py", line 163:
    sage: B.plot()
Exception raised:
    Traceback (most recent call last):
      File "/Users/rlmill/sage-4.3.3/local/bin/ncadoctest.py", line 1231, in run_one_test
        self.run_one_example(test, example, filename, compileflags)
      File "/Users/rlmill/sage-4.3.3/local/bin/sagedoctest.py", line 38, in run_one_example
        OrigDocTestRunner.run_one_example(self, test, example, filename, compileflags)
      File "/Users/rlmill/sage-4.3.3/local/bin/ncadoctest.py", line 1172, in run_one_example
        compileflags, 1) in test.globs
      File "<doctest __main__.example_5[5]>", line 1, in <module>
        B.plot()###line 163:
    sage: B.plot()
      File "/Users/rlmill/sage-4.3.3/local/lib/python/site-packages/sage/graphs/graph_bundle.py", line 173, in plot
        total_pos = generic_graph_pyx.spring_layout_fast(self, iterations=iters)
      File "generic_graph_pyx.pyx", line 86, in sage.graphs.generic_graph_pyx.spring_layout_fast (sage/graphs/generic_graph_pyx.c:1361)
      File "/Users/rlmill/sage-4.3.3/local/lib/python/site-packages/sage/graphs/graph.py", line 2061, in to_undirected
        return copy(self)
      File "/Users/rlmill/sage-4.3.3/local/lib/python2.6/copy.py", line 79, in copy
        return copier(x)
      File "/Users/rlmill/sage-4.3.3/local/lib/python/site-packages/sage/graphs/generic_graph.py", line 483, in __copy__
        G = self.__class__(self, name=self.name(), pos=copy(self._pos), boundary=copy(self._boundary), implementation=implementation, sparse=sparse)
      File "/Users/rlmill/sage-4.3.3/local/lib/python/site-packages/sage/graphs/graph_bundle.py", line 73, in __init__
        if isinstance(args[0], (list, tuple)):
    IndexError: tuple index out of range

[rlm]

I will offer the suggestion to disable the failing doctest, for discussion. I don't know of anyone who uses graph bundles in Sage. I've emailed Chris Godsil to see if he ever uses it.

[rhinton]

Sounds great to me. The new patch disables the "plot" doctest that was failing and adds a "known bugs" warning to the beginning of the graph_bundle.py.

[rhinton]

modifies copy() code in generic_graph.py and adds test to BipartiteGraph?

[rlm]

Anyone queasy about the GraphBundle doctest, Godsil tells me he has a student who is working on a bundle package, which I am assuming actually does something (unlike bundles in Sage). Look for this to appear in Sage at some point soon!

[rhinton]

Found a failing doctest in combinat/posets/posets.py due to this change. Additional patch coming soon.

[rhinton]

Apply both patches, order does not matter

[rhinton]

New patch fixes posets doctest failure. Apply both patches; order does not matter.

[ncohen]

Good patch ! Applies fine, does its job.. :-)

Nathann

[jhpalmieri]

Merged in 4.4.alpha0:

• trac_8329-bipartite-graph-copy.patch
• trac_8329-posets-repr.patch