Release history =============== 0.12 ---- - Added ``ObjectGraph.edgeData`` to retrieve the edge data from a specific edge. - Added ``AltGraph.update_edge_data`` and ``ObjectGraph.updateEdgeData`` to update the data associated with a graph edge. 0.11 ---- - Stabilize the order of elements in dot file exports, patch from bitbucket user 'pombredanne'. - Tweak setup.py file to remove dependency on distribute (but keep the dependency on setuptools) 0.10.2 ------ - There where no classifiers in the package metadata due to a bug in setup.py 0.10.1 ------ This is a bugfix release Bug fixes: - Issue #3: The source archive contains a README.txt while the setup file refers to ReadMe.txt. This is caused by a misfeature in distutils, as a workaround I've renamed ReadMe.txt to README.txt in the source tree and setup file. 0.10 ----- This is a minor feature release Features: - Do not use "2to3" to support Python 3. As a side effect of this altgraph now supports Python 2.6 and later, and no longer supports earlier releases of Python. - The order of attributes in the Dot output is now always alphabetical. With this change the output will be consistent between runs and Python versions. 0.9 --- This is a minor bugfix release Features: - Added ``altgraph.ObjectGraph.ObjectGraph.nodes``, a method yielding all nodes in an object graph. Bugfixes: - The 0.8 release didn't work with py2app when using python 3.x. 0.8 ----- This is a minor feature release. The major new feature is a extensive set of unittests, which explains almost all other changes in this release. Bugfixes: - Installing failed with Python 2.5 due to using a distutils class that isn't available in that version of Python (issue #1 on the issue tracker) - ``altgraph.GraphStat.degree_dist`` now actually works - ``altgraph.Graph.add_edge(a, b, create_nodes=False)`` will no longer create the edge when one of the nodes doesn't exist. - ``altgraph.Graph.forw_topo_sort`` failed for some sparse graphs. - ``altgraph.Graph.back_topo_sort`` was completely broken in previous releases. - ``altgraph.Graph.forw_bfs_subgraph`` now actually works. - ``altgraph.Graph.back_bfs_subgraph`` now actually works. - ``altgraph.Graph.iterdfs`` now returns the correct result when the ``forward`` argument is ``False``. - ``altgraph.Graph.iterdata`` now returns the correct result when the ``forward`` argument is ``False``. Features: - The ``altgraph.Graph`` constructor now accepts an argument that contains 2- and 3-tuples instead of requireing that all items have the same size. The (optional) argument can now also be any iterator. - ``altgraph.Graph.Graph.add_node`` has no effect when you add a hidden node. - The private method ``altgraph.Graph._bfs`` is no longer present. - The private method ``altgraph.Graph._dfs`` is no longer present. - ``altgraph.ObjectGraph`` now has a ``__contains__`` methods, which means you can use the ``in`` operator to check if a node is part of a graph. - ``altgraph.GraphUtil.generate_random_graph`` will raise ``GraphError`` instead of looping forever when it is impossible to create the requested graph. - ``altgraph.Dot.edge_style`` raises ``GraphError`` when one of the nodes is not present in the graph. The method silently added the tail in the past, but without ensuring a consistent graph state. - ``altgraph.Dot.save_img`` now works when the mode is ``"neato"``. 0.7.2 ----- This is a minor bugfix release Bugfixes: - distutils didn't include the documentation subtree 0.7.1 ----- This is a minor feature release Features: - Documentation is now generated using `sphinx <http://pypi.python.org/pypi/sphinx>`_ and can be viewed at <http://packages.python.org/altgraph>. - The repository has moved to bitbucket - ``altgraph.GraphStat.avg_hops`` is no longer present, the function had no implementation and no specified behaviour. - the module ``altgraph.compat`` is gone, which means altgraph will no longer work with Python 2.3. 0.7.0 ----- This is a minor feature release. Features: - Support for Python 3 - It is now possible to run tests using 'python setup.py test' (The actual testsuite is still very minimal though)