pyminimax

Simple Python 3 example of the minimax algorithm using a NetworkX graph

This example assumes you have NetworkX installed as part of your Python 3 distribution.

Using conda to install NetworkX:

conda install -c anaconda networkx

or if you prefer pip:

pip install networkx

run the example from the src folder:

python src/minimax.py 

Setting up a minimax graph

NetworkX is a great Python library for managing and traversing graph based structures.

Use G.add_node() to add a node to the graph. Set terminal to None if the node is not a terminal value, otherwise set to a valid integer or float score:

G.add_node("nodeName", terminal=int_or_float_value)

Use G.add_edge() to add an edge from a source node to a destination node:

G.add_edge("sourceNode", "destinationNode")

Visualizing the graph

Check the correctness of your graph by visualizing it. You will require pydot and Graphviz. If running on Ubuntu/Linux you require Graphviz:

sudo apt install graphviz

And pydot Python installation using pip:

pip install pydot

(conda does not seem to support pydot and Python 3)

And the Python code to accomplish this:

    p = nx.drawing.nx_pydot.to_pydot(G)
    p.write_png('output_filename.png')

And below is a graph example from the complex_tree() function:

For sake of clarity and to highlight terminal nodes in the graph visualization, I have annotated the add_node() methods like this:

G.add_node("nodeName=terminal", terminal=int_or_float_value)

That is add the terminal value to the name of the node. Note, add_edge() should also correspond to the annotated node name.