delta/python-packages/networkx.git/networkx/classes/digraph.py, branch docdraft

"Warning" is not a numpydoc section

2015-06-13T20:22:22+00:00

Expand data keyword in G.edges and add default data

2015-01-18T02:42:00+00:00

When data=True, the whole data dictionary is returned for each edge.
When data=False, no data is returned.
Otherwise data is used as the attribute name of the edge data to return.
>>> [ (u,v,edata['weight']) for u,v,edata in G.edges(data=True) ]
becomes
>>> G.edges(data='weight')

>>> for n,nbrs in G.adjacency_iter():
...     for nbr,ddict in nbrs.items():
...         print (n,nbr,ddict.get('weight',1))
becomes
>>> G.edges(data='weight', default=1)

Address ironpython issues from #1327

2015-01-10T02:36:54+00:00

Meld special classes with base classes.

2015-01-02T05:18:11+00:00

This provides the ability to use OrderedGraph or ThinGraph by
simple subclassing of Graph.

Make add_nodes_from compatible with IronPython

2014-07-04T00:46:12+00:00

clarified edge attribute update precedence in add_edges_from()

2014-04-12T16:07:10+00:00

Fix/adjust/remove/modify tests that depend on dictionary ordering.

2012-07-08T16:09:17+00:00

Addresses #741

There may still be more that need fixing since

1) NumPy and some other optional packages do not yet work with Python3.3 so those tests weren't checked

2) Finding these errors by running the tests is nondeterministic itself (e.g. passes 4 times and fails on the 5th).  And inspecting all of the tests by hand is error-prone too.

Adjust graph/digraph/multigraph/multidgraph node attribute example so it doesn't depend on data ordering.

2012-06-23T17:52:02+00:00

Add comment about edges() v out_edges().

2012-04-08T19:34:41+00:00

Fixes #686

Change node lookups to G.node instead of G.adj

2012-02-06T04:32:34+00:00

This should help a small amount when playing with alternate
data structures.

Addreses:  #206, #583