frej.fuzzy
Class Fuzzy

Object
  frej.fuzzy.Fuzzy

public final class Fuzzy
extends Object

Class providing fuzzy string comparison. It is used in fuzzy regexp matching, but could also be used alone, for fuzzy string matching or fuzzy substring search. Based on Demerau-Levenshtein distance evaluation, i.e. there are four types of "mistakes" each counting as 1 point (char deletion, char adding, char replacement, swap of two adjacent chars). Methods are static, and resulting variables too, so necessary values should be read before new matching/searching attempt.

Author:

Rodion Gorkovenko

Field Summary

static String	matchedPattern keeps best matched string after matching against a list of strings
static double	result "distance" of last match (roughly mistakes count divided by length
static int	resultEnd keeps ending position of matched region after substring search
static int	resultIndex keeps index of best match after matching against a list of strings
static int	resultStart keeps starting position of matched region after substring search
static double	threshold if result of match is higher than threshold, boolean methods return "false"

Constructor Summary

Fuzzy()

Method Summary

static double	bestEqual(String string, Object patterns, boolean equality) Given list or array of strings, searches for one which is best matched with whole original string (equality=true) or with some its substring (equality=false).
static double	containability(CharSequence source, CharSequence pattern) Core method for searching substring.
static boolean	containsOneOf(CharSequence source, CharSequence... patterns) Tests whether any of "patterns" is presented in "source" as substring.
static boolean	equals(CharSequence source, CharSequence pattern) Tests whether "source" matches "pattern".
static double	similarity(CharSequence source, CharSequence pattern) Core method for measuring Demerau-Levenshtein distance between two strings.
static int	substrEnd(CharSequence source, CharSequence pattern) Tries to find substring "pattern" in the "source" and if successful, returns the position of the end of the match.
static int	substrStart(CharSequence source, CharSequence pattern) Tries to find substring "pattern" in the "source" and if successful, returns the position of the beginning of the match.

Methods inherited from class Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

resultStart

public static int resultStart

keeps starting position of matched region after substring search

resultEnd

public static int resultEnd

keeps ending position of matched region after substring search

resultIndex

public static int resultIndex

keeps index of best match after matching against a list of strings

matchedPattern

public static String matchedPattern

keeps best matched string after matching against a list of strings

result

public static double result

"distance" of last match (roughly mistakes count divided by length

threshold

public static double threshold

if result of match is higher than threshold, boolean methods return "false"

Constructor Detail

Fuzzy

public Fuzzy()

Method Detail

substrStart

public static int substrStart(CharSequence source,
                              CharSequence pattern)

Tries to find substring "pattern" in the "source" and if successful, returns the position of the beginning of the match.

Returns:

position of found substring (0 .. source.length() - 1) or (-1) if substring was not found (with given threshold).

substrEnd

public static int substrEnd(CharSequence source,
                            CharSequence pattern)

Tries to find substring "pattern" in the "source" and if successful, returns the position of the end of the match.

Returns:

position of found substring end (0 .. source.length() - 1) or (-1) if substring was not found (with given threshold).

equals

public static boolean equals(CharSequence source,
                             CharSequence pattern)

Tests whether "source" matches "pattern".

Returns:

true or false depending on match quality.

containsOneOf

public static boolean containsOneOf(CharSequence source,
                                    CharSequence... patterns)

Tests whether any of "patterns" is presented in "source" as substring. Stops on first good match.

Returns:

true or false depending on whether any of pattern search succeeds.

containability

public static double containability(CharSequence source,
                                    CharSequence pattern)

Core method for searching substring. Finds the region for which Demerau-Levenshtein distance is minimal.

Returns:

normalized best distance (i.e. distance / pattern.length())

bestEqual

public static double bestEqual(String string,
                               Object patterns,
                               boolean equality)

Given list or array of strings, searches for one which is best matched with whole original string (equality=true) or with some its substring (equality=false).

Returns:

best match result (normalized distance).

similarity

public static double similarity(CharSequence source,
                                CharSequence pattern)

Core method for measuring Demerau-Levenshtein distance between two strings.

Returns:

normalized distance (distance / average(source.length(), pattern.length()))