doc/release/release_2.1.rst
Release date: 22 January 2018
Supports Python 2.7, 3.4, 3.5, and 3.6.
NetworkX is a Python package for the creation, manipulation, and study of the structure, dynamics, and functions of complex networks.
For more information, please visit our website <https://networkx.org/>_
and our gallery of examples <https://networkx.org/documentation/latest/auto_examples/index.html>.
Please send comments and questions to the networkx-discuss mailing list <http://groups.google.com/group/networkx-discuss>.
This release is the result of four months of work with 75 pull requests by 37 contributors. Highlights include:
Arrows for drawing DiGraph edges are vastly improved! And an example to show them.
More than 12 new functions for graph generation, manipulation and/or new graph algorithms.
G.name is no longer processed by graph operators. It remains as a
property mechanism to access G.graph['name'] but the user is in
charge of updating or changing it for copies, subgraphs, unions and
other graph operations.
**kwds
to avoid name collisions -- especially add_node[#2498 <https://github.com/networkx/networkx/pull/2498>_]
In geographical_threshold_graph, starting in NetworkX 2.1 the parameter
alpha is deprecated and replaced with the customizable p_dist
function parameter, which defaults to r^-2
if p_dist is not supplied. To reproduce networks of earlier NetworkX
versions, a custom function needs to be defined and passed as the p_dist
parameter. For example, if the parameter alpha = 2 was used in NetworkX 2.0,
the custom function def custom_dist(r): r**-2 can be passed in versions >=2.1
as the parameter p_dist = custom_dist to produce an equivalent network.
Note the change in sign from +2 to -2 in this parameter change.
[#2554 <https://github.com/networkx/networkx/issues/2554>_]
New algorithms for finding k-edge-connected components and k-edge-connected
subgraphs in directed and undirected graphs. Efficient implementations are
provided for the special case of k=1 and k=2. The new functionality is
provided by:
:func:k_edge_components()
:func:k_edge_subgraphs()
[#2572 <https://github.com/networkx/networkx/issues/2572>_]
New algorithm finding for finding k-edge-augmentations in undirected graphs.
Efficient implementations are provided for the special case of k=1 and k=2.
New functionality is provided by:
k_edge_augmentation()[#2812 <https://github.com/networkx/networkx/pull/2812>_]
Removed bellman_ford, please use
bellman_ford_predecessor_and_distance.
[#2811 <https://github.com/networkx/networkx/pull/2811>_]
Removed to_pandas_dataframe and from_pandas_dataframe, please use
to_pandas_adjacency, from_pandas_adjacency, to_pandas_edgelist,
or from_pandas_edgelist.
[#2766 <https://github.com/networkx/networkx/pull/2766>_]
Add seed keyword argument to random_layout and spring_layout
[#2776 <https://github.com/networkx/networkx/pull/2776>_]
Add threshold option to spring layout
[#2774 <https://github.com/networkx/networkx/pull/2774>_]
max_weight_matching returns set of edges
[#2753 <https://github.com/networkx/networkx/pull/2753>_]
Add directed graphs support for jit_graph reading
[#2788 <https://github.com/networkx/networkx/pull/2788>_]
Control node-border color in draw_networkx_nodes
#2819 <https://github.com/networkx/networkx/pull/2819>_]
Deprecate connected_component_subgraphs, biconnected_component_subgraphs,
attracting_component_subgraphs, strongly_connected_component_subgraphs,
weakly_connected_component_subgraphs.
Instead use: [G.subgraph(c) for c in *_components]to_pandas_dataframe and from_pandas_dataframe (#2811)bellman_ford (#2812)furether in networkx tutorial documentation (#2764)doc.txt in requirements/README.md (#2699)strict from function networkx.drawing.nx_pydot.to_pydot (#2672)