Decision trees in python using recursion and eval.

decision_tree.py

# Decision trees in python using recursion and eval.

TREE_SIMPLE = {
    "expression": 'state["x"] > 1',
    True: {"result": "b"},
    False: {"result": "a"},
}

def decide(state, node):
    return (
        node["result"]
        if "result" in node
        else decide(state, node[eval(node["expression"])])
    )


print(decide(state={"x": 1}, node=TREE_SIMPLE))  # a
print(decide(state={"x": 2}, node=TREE_SIMPLE))  # b


TREE = {
    "expression": 'state["x"] > 0',
    True: {
        "expression": 'state["x"] > 1',
        True: {"result": "c"},
        False: {"result": "b"},
    },
    False: {"result": "a"},
}


print(decide(state={"x": 0}, node=TREE))  # a
print(decide(state={"x": 1}, node=TREE))  # b
print(decide(state={"x": 2}, node=TREE))  # c