Installation

The package is registered in the General registry and so can be installed at the REPL with ] add StringDistances.

Evaluate

The function evaluate returns the distance between two strings. Its syntax is:

evaluate(dist, s1, s2)

where s1 and s2 can be any iterator with a length method (e.g. AbstractString, GraphemeIterator, AbstractVector...), and dist is one of the following distances:

• Edit Distances

  • Jaro Distance Jaro()
  • Levenshtein Distance Levenshtein()
  • Damerau-Levenshtein Distance DamerauLevenshtein()
  • RatcliffObershelp Distance RatcliffObershelp()

• Q-gram distances compare the set of all substrings of length q in each string.

  • QGram Distance Qgram(q::Int)
  • Cosine Distance Cosine(q::Int)
  • Jaccard Distance Jaccard(q::Int)
  • Overlap Distance Overlap(q::Int)
  • Sorensen-Dice Distance SorensenDice(q::Int)

• The package includes distance "modifiers", that can be applied to any distance.

  • Winkler diminishes the distance of strings with common prefixes. The Winkler adjustment was originally defined for the Jaro similarity score but this package defines it for any string distance.
  • Partial returns the minimum distance between the shorter string and substrings of the longer string.
  • TokenSort adjusts for differences in word orders by reording words alphabetically.
  • TokenSet adjusts for differences in word orders and word numbers by comparing the intersection of two strings with each string.
  • TokenMax combines scores using the base distance, the Partial, TokenSort and TokenSet modifiers, with penalty terms depending on string lengths.

Some examples:

evaluate(Jaro(), "martha", "marhta")
evaluate(Winkler(Jaro()), "martha", "marhta")
evaluate(QGram(2), "martha", "marhta")
evaluate(Winkler(QGram(2)), "martha", "marhta")
evaluate(Levenshtein(), "martha", "marhta")
evaluate(Partial(Levenshtein()), "martha", "marhta")
evaluate(Jaro(), "martha", "marhta")
evaluate(TokenSet(Jaro()), "martha", "marhta")
evaluate(TokenMax(RatcliffObershelp()), "martha", "marhta")

Alternatively, each distance can be used as a callable to call the evaluate function of each metric or modified metric, for example:

Jaro()("martha", "marhta")
Winkler(Jaro())("martha", "marhta")
QGram(2)("martha", "marhta")

A good distance to match strings composed of multiple words (like addresses) is TokenMax(Levenshtein()) (see fuzzywuzzy).

Compare

The function compare is defined as 1 minus the normalized distance between two strings. It always returns a number between 0 and 1: a value of 0 means completely different and a value of 1 means completely similar.

evaluate(Levenshtein(), "New York", "New York")
#> 0
compare("New York", "New York", Levenshtein())
#> 1.0

Find

• findmax returns the value and index of the element in itr with the highest similarity score with s. Its syntax is:

  findmax(s, itr, dist::StringDistance; min_score = 0.0)
  

• findall returns the indices of all elements in itr with a similarity score with s higher than a minimum value (default to 0.8). Its syntax is:

  findall(s, itr, dist::StringDistance; min_score = 0.8)
  

The functions findmax and findall are particularly optimized for Levenshtein and DamerauLevenshtein distances (as well as their modifications via Partial, TokenSort, TokenSet, or TokenMax).

References

• The stringdist Package for Approximate String Matching Mark P.J. van der Loo
• fuzzywuzzy