substring-structures 0.0.2

Structures for doing substring searches.

substring_structures

https://pypi.org/project/substring-structures/

Python library for substring search structures. Includes implementations of Ukkonen's algorithm, the Aho-Corasick algorithm, and the Knuth-Morris-Pratt algorithm.

Ukkonen's Suffix Tree

>>> from substring_structures import SuffixTree

>>> hello = SuffixTree("hello")
>>> "lo" in hello
True
>>> "he" in hello
True
>>> "hi" in hello
False

Suffix tree implementation based on descriptions from https://www.charliemistrata.com/posts/ukkonens-algorithm.

This suffix tree can be precomputed for a string S in O(N) time using Ukkonen's algorithm. It can then be used to check if a string W of length M is a substring of S in O(M) time.

Aho-Corasick String Finite State Machine

>>> from substring_structures import ACStringFSM

>>> greetings = ACStringFSM({"hello", "hi", "hey"})
>>> greetings.find_substrings_in_superstring(superstring="hi, hello")
{"hello", "hi"}

Structure used to preprocess a set of substrings to efficiently do substring checks.

Based on the Aho-Corasick algorithm, this structure is an FSM (finite state machine) that can be computed in O(N) time for a set of strings strings of total character count N. It can then be used to efficiently check which strings in strings are in an arbitrary larger string superstring using the find_substrings_in_superstring(superstring) method.

Knuth-Morris-Pratt Prefix Fallback

>>> from substring_structures import KMPPrefixFallback

>>> KMPPrefixFallback("bababooie")
KMPPrefixFallback(string='bababooie', fallback_length_by_prefix_length=[-1, 0, 0, 1, 2, 3, 0, 0, 0])
>>> baba = KMPPrefixFallback("baba")
>>> baba.contained_by("ababab")
True
>>> baba.contained_by("keke")
False

Structure used to preprocess a string W of length M in O(M) time so that you can check if its a substring of another string S of length N in O(N) time using KMPPrefixFallback(w).contained_by(s).

It does this by generating a structure of string W that allows you to "fallback" from a longer prefix of W to a shorter prefix of W on a character mismatch when searching S. The code here is based on descriptions at https://www.charliemistrata.com/posts/knuth-morris-pratt.

This class is mainly to illustrate the Knuth-Morris-Pratt algorithm as python has a fast built-in algorithm for substring search.