Basic class mapping

The primary function of the inheritance_explorer is to trace out the class inheritance starting from a single parent class and recursively traversing all child classes. To get started, initialize a ClassGraphTree with the starting class:

from matplotlib.axes import Axes
from inheritance_explorer import ClassGraphTree

base_class = Axes # the starting class to map from
cgt = ClassGraphTree(base_class) # traces an inheritance tree

From here, there are a number of ways to visualize the inheritance graph.

From a jupyter notebook, you can construct a pyvis interactive network graph

with:

graph = cgt.build_interactive_graph(width="1200px",
                                    height="1200px",
                                    bgcolor='#222222',
                                    font_color='white')
graph.show('_tmp.html') # render the pyvis interactive graph

If you have graphviz installed, you can also render a static graph either to

display in a jupyter notebook or save to file. To display in a notebook:

cgt.show_graph()

To save a file, you can access the underlying graph object directly:

cgt.graph.write('matplotlib_Axes.png')

Function tracking

The ClassGraphTree can also track a selected function of a class during traversal. When a child class overrides the function, the source code is stored. To use this functionality, use the funcname keyword argument to provide the function name as a string:

from matplotlib.axes import Axes
from inheritance_explorer import ClassGraphTree

base_class = Axes # the starting class to map from
cgt = ClassGraphTree(base_class, funcname='clear')

By default, after the traversal completes, the different versions of the function that is tracked will be run through a code-similarity calculation. Cases where the similarity fraction is above a cutoff value (default of .75), the two nodes will be connected on the resulting graph:

Additionally, you can inspect the source code itself for any of the classes that define it using cgt.get_source_code(class_name). If in a jupyter notebook, you can view it with syntax highlighting with

from IPython.display import Code
Code(cgt.get_source_code('PolarAxes'), language="python")

For a complete description of the code similarity calculation and how to modify it, check out the full documentation.

Additional Features

The following additional features are described in more detail in the full documentation:

• command line usage

• code comparison widget for Jupyter