pure_radix

Radix tree library but in pure clean Python. Run it on PyPy or IronPython or wherever.

Dependencies

We try to keep dependencies light:

• attrs dataclass library (2881 LoC)

To run the tests, you’ll need pytest and optionally hypothesis.

Examples

>>> from pure_radix import SetNode
>>>
>>> # Create a tree where each node holds a set. You are free to
>>> # customize what information is held inside each node.
>>> t = SetNode()
>>>
>>> node1 = t.node_get((10, 11, 12), create=True)
>>> node2 = t.node_force((10, 11, 30))  # equivalent
>>>
>>> node1.add("cat")
>>> node2.add("dog")
>>>
>>> print(t.node_debug_string())
[]
  [10, 11]
    [30] {'dog'}
    [12] {'cat'}
>>>
>>> # This node exists because it's a common prefix between node1 and node2.
>>> t.node_get((10, 11))
SetNode(raw_data=None)
>>>
>>> # This node doesn't exist.
>>> print(t.node_get((10,)))
None

The library works with any kind of sequence, including strings!

>>> # It also works with strings!
>>> t = SetNode()
>>>
>>> t.node_force("catalogue").add(1)
>>> t.node_force("category").add(2)
>>> t.node_force("categorization").add(3)
>>> t.node_force("cart").add(4)
>>> t.node_force("carpet")  # leaf node without actual data!
SetNode(raw_data=None)
>>> t.node_force("car").add(5)
>>> t.node_force("star").add(6)
>>>
>>> print(t.node_debug_string())
[]
  ['c', 'a']
    ['t']
      ['a', 'l', 'o', 'g', 'u', 'e'] {1}
      ['e', 'g', 'o', 'r']
        ['i', 'z', 'a', 't', 'i', 'o', 'n'] {3}
        ['y'] {2}
    ['r'] {5}
      ['p', 'e', 't']
      ['t'] {4}
  ['s', 't', 'a', 'r'] {6}

Are you looking for the longest common prefix? Here is how you can get the nodes that have the longest common prefix with a sequence:

>>> for n, node in t.node_find_closest_nodes("catenary"):
...     print(n, node)
... 
4 SetNode(raw_data=None)
3 SetNode(raw_data={1})
2 SetNode(raw_data={5})
0 SetNode(raw_data={6})

Do you need to recursively get all non-empty nodes under a node?

>>> list(t.node_get("cat").node_find())
[SetNode(raw_data={2}), SetNode(raw_data={3}), SetNode(raw_data={1})]

Let’s empty some nodes now!

>>> t.node_get("category").clear()
>>> t.node_get("catalogue").clear()
>>> t.node_get("star").clear()
>>> print(t.node_debug_string())
[]
  ['c', 'a']
    ['t']
      ['a', 'l', 'o', 'g', 'u', 'e']
      ['e', 'g', 'o', 'r']
        ['i', 'z', 'a', 't', 'i', 'o', 'n'] {3}
        ['y']
    ['r'] {5}
      ['p', 'e', 't']
      ['t'] {4}
  ['s', 't', 'a', 'r']

Empty nodes do not get removed automatically! You must remove them yourself:

>>> t.node_prune()
>>> print(t.node_debug_string())
[]
  ['c', 'a']
    ['r'] {5}
      ['t'] {4}
    ['t', 'e', 'g', 'o', 'r', 'i', 'z', 'a', 't', 'i', 'o', 'n'] {3}

Use a powerful method of traversal:

>>> def _enter(stack, frame):
...     yield ("enter", frame.node, frame.node.node_sequence)
...
>>> def _exit(stack, frame):
...     yield ("exit", frame.node, frame.node.node_sequence)
...
>>> for action, node, seq in t.node_visit(enter=_enter, exit=_exit):
...     print(action, node, seq)
enter SetNode(raw_data=None) ()
enter SetNode(raw_data=None) ('c', 'a')
enter SetNode(raw_data={5}) ('c', 'a', 'r')
enter SetNode(raw_data={4}) ('c', 'a', 'r', 't')
exit SetNode(raw_data={4}) ('c', 'a', 'r', 't')
exit SetNode(raw_data={5}) ('c', 'a', 'r')
enter SetNode(raw_data={3}) ('c', 'a', 't', 'e', 'g', 'o', 'r', 'i', 'z', 'a', 't', 'i', 'o', 'n')
exit SetNode(raw_data={3}) ('c', 'a', 't', 'e', 'g', 'o', 'r', 'i', 'z', 'a', 't', 'i', 'o', 'n')
exit SetNode(raw_data=None) ('c', 'a')
exit SetNode(raw_data=None) ()