Python radix tree implementation for IPv4 and IPv6 prefix matching.





GitHub Stars



Last Commit

4yrs ago











.. image:: :target:

.. image:: :target:

py-radix implements the radix tree data structure for the storage and retrieval of IPv4 and IPv6 network prefixes.

The radix tree is commonly used for routing table lookups. It efficiently stores network prefixes of varying lengths and allows fast lookups of containing networks.


Installation is a breeze via pip: ::

pip install py-radix

Or with the standard Python distutils incantation: ::

python build
python install

The C extension will be built for supported python versions. If you do not want the C extension, set the environment variable RADIX_NO_EXT=1.

Tests are in the tests/ directory and can be run with python nosetests.


A simple example that demonstrates most of the features: ::

import radix

# Create a new tree
rtree = radix.Radix()

# Adding a node returns a RadixNode object. You can create
# arbitrary members in its 'data' dict to store your data
rnode = rtree.add("")["blah"] = "whatever you want"

# You can specify nodes as CIDR addresses, or networks with
# separate mask lengths. The following three invocations are
# identical:
rnode = rtree.add("")
rnode = rtree.add("", 16)
rnode = rtree.add(network = "", masklen = 16)

# It is also possible to specify nodes using binary packed
# addresses, such as those returned by the socket module
# functions. In this case, the radix module will assume that
# a four-byte address is an IPv4 address and a sixteen-byte
# address is an IPv6 address. For example:
binary_addr = inet_ntoa("")
rnode = rtree.add(packed = binary_addr, masklen = 23)

# Exact search will only return prefixes you have entered
# You can use all of the above ways to specify the address
rnode = rtree.search_exact("")
# Get your data back out
# Use a packed address
addr = socket.inet_ntoa("")
rnode = rtree.search_exact(packed = addr, masklen = 8)

# Best-match search will return the longest matching prefix
# that contains the search term (routing-style lookup)
rnode = rtree.search_best("")

# Worst-search will return the shortest matching prefix
# that contains the search term (inverse routing-style lookup)
rnode = rtree.search_worst("")

# Covered search will return all prefixes inside the given
# search term, as a list (including the search term itself,
# if present in the tree)
rnodes = rtree.search_covered("")

# There are a couple of implicit members of a RadixNode:
print # -> ""
print rnode.prefix  # -> ""
print rnode.prefixlen   # -> 8
print  # -> socket.AF_INET
print rnode.packed  # -> '\n\x00\x00\x00'

# IPv6 prefixes are fully supported in the same tree
rnode = rtree.add("2001:DB8::/3")
rnode = rtree.add("::/0")

# Use the nodes() method to return all RadixNodes created
nodes = rtree.nodes()
for rnode in nodes:
    print rnode.prefix

# The prefixes() method will return all the prefixes (as a
# list of strings) that have been entered
prefixes = rtree.prefixes()

# You can also directly iterate over the tree itself
# this would save some memory if the tree is big
# NB. Don't modify the tree (add or delete nodes) while
# iterating otherwise you will abort the iteration and
# receive a RuntimeWarning. Changing a node's data dict
# is permitted.
for rnode in rtree:

print rnode.prefix


py-radix is licensed under a ISC/BSD licence. The underlying radix tree implementation is taken (and modified) from MRTd and is subject to a 4-term BSD license. See the LICENSE file for details.


Please report bugs via GitHub at Code changes can be contributed through a pull request on GitHub or emailed directly to me

The main portions of the directory tree are as follows: ::

├── radix/*.py      # Pure Python code
├── radix/_radix.c  # C extension code (compatible with pure python code)
├── radix/_radix/*  # C extension code (compatible with pure python code)
├── tests/          # Tests (regression and unit)
└──        # Standard for installation/testing/etc.

Rate & Review

Great Documentation0
Easy to Use0
Highly Customizable0
Bleeding Edge0
Responsive Maintainers0
Poor Documentation0
Hard to Use0
Unwelcoming Community0
No reviews found
Be the first to rate


No alternatives found


No tutorials found
Add a tutorial