add Hamming + restrict pairwise to vectors + handle missings

src/StringDistances.jl

@@ -7,10 +7,11 @@ include("distances/edit.jl")
 include("distances/qgram.jl")
 include("normalize.jl")
 
-const StringDistance = Union{Jaro, Levenshtein, DamerauLevenshtein, RatcliffObershelp, QGramDistance, Winkler, Partial, TokenSort, TokenSet, TokenMax, Normalize}
+const StringDistance = Union{Hamming, Jaro, Levenshtein, DamerauLevenshtein, RatcliffObershelp, QGramDistance, Winkler, Partial, TokenSort, TokenSet, TokenMax, Normalize}
 # Distances API
 Distances.result_type(dist::StringDistance, s1::Type, s2::Type) = typeof(dist("", ""))
-Distances.result_type(dist::StringDistance, s1, s2) = result_type(dist, eltype(s1), eltype(s2))
+Distances.result_type(dist::StringDistance, s1, s2) = result_type(dist, typeof(s1), typeof(s2))
+
 
 include("find.jl")
 include("pairwise.jl")
@@ -23,10 +24,12 @@ include("pairwise.jl")
 
 export
 StringDistance,
+Hamming,
 Levenshtein,
 DamerauLevenshtein,
 Jaro,
 RatcliffObershelp,
+QGramDistance,
 QGram,
 Cosine,
 Jaccard,

src/distances/edit.jl

@@ -1,3 +1,27 @@
+"""
+    Hamming()
+
+Creates the Hamming distance
+
+The Hamming distance is defined as the number of characters that do not match
+"""
+struct Hamming{V <: Union{Integer, Nothing}} <: SemiMetric
+   max_dist::V
+end
+Hamming() = Hamming(nothing)
+
+function (dist::Hamming)(s1, s2)
+    ((s1 === missing) | (s2 === missing)) && return missing
+    current = abs(length(s2) - length(s1))
+    dist.max_dist !== nothing && current > dist.max_dist && return dist.max_dist + 1
+    for (ch1, ch2) in zip(s1, s2)
+        current += ch1 != ch2
+        dist.max_dist !== nothing && current > dist.max_dist && return dist.max_dist + 1
+    end
+    return current
+end
+
+
 """
     Jaro()
 

src/normalize.jl

@@ -11,6 +11,14 @@ end
 normalize(dist::SemiMetric, max_dist = 1.0) = Normalize(dist)
 normalize(dist::Normalize, max_dist = 1.0) = Normalize(dist.dist)
 
+function (dist::Normalize{<:Hamming})(s1, s2, max_dist = 1.0)
+    ((s1 === missing) | (s2 === missing)) && return missing
+    s1, s2 = reorder(s1, s2)
+    len1, len2 = length(s1), length(s2)
+    len2 == 0 && return 1.0
+    out = dist.dist(s1, s2) / len2
+    out > max_dist ? 1.0 : out
+end
 
 # A normalized distance is between 0 and 1, and accept a third argument, max_dist.
 function (dist::Normalize{<: Union{Levenshtein, DamerauLevenshtein}})(s1, s2, max_dist = 1.0)

src/pairwise.jl

@@ -1,12 +1,12 @@
 @doc """
-    pairwise(dist::StringDistance, itr; preprocess = nothing)
+    pairwise(dist::StringDistance, xs::AbstractVector; preprocess = nothing)
-    pairwise(dist::StringDistance, itr1, itr2; preprocess = nothing)
+    pairwise(dist::StringDistance, xs::AbstractVector, ys::AbstractVector; preprocess = nothing)
 
-Compute distances between all pairs of elements in `itr` according to the
+Compute distances between all pairs of elements in `xs`  and `ys` according to the
 `StringDistance` `dist`.
 
 For QGramDistances preprocessing will be used either if `preprocess` is set 
-to true or if there are more than 5 elements in `itr`. Set `preprocess` to 
+to true or if there are more than 5 elements in `xs`. Set `preprocess` to 
 false if no preprocessing should be used, regardless of length.
 
 Both symmetric and asymmetric versions are available.
@@ -28,55 +28,66 @@ julia> pairwise(Levenshtein(), iter, iter2) # asymmetric
 """
 Distances.pairwise
 
-function Distances.pairwise(dist::StringDistance, X, Y; preprocess = length(X) >= 5)
+function Distances.pairwise(dist::StringDistance, xs::AbstractVector, ys::AbstractVector; preprocess = nothing)
-    T = result_type(dist, eltype(X), eltype(Y))
+    T = result_type(dist, eltype(xs), eltype(ys))
-    R = Matrix{T}(undef, length(X), length(Y))
+    if Missing <: Union{eltype(xs), eltype(ys)}
-    pairwise!(R, dist, X, Y; preprocess = preprocess)
+        T = Union{T, Missing}
+    end
+    R = Matrix{T}(undef, length(xs), length(ys))
+    pairwise!(R, dist, xs, ys; preprocess = preprocess)
 end
 
-function Distances.pairwise(dist::StringDistance, X; preprocess = nothing)
+function Distances.pairwise(dist::StringDistance, xs::AbstractVector; preprocess = nothing)
-    T = result_type(dist, eltype(X), eltype(X))
+    T = result_type(dist, eltype(xs), eltype(xs))
-    R = Matrix{T}(undef, length(X), length(X))
+    if Missing <: eltype(xs)
-    pairwise!(R, dist, X; preprocess = preprocess)
+        T = Union{T, Missing}
+    end
+    R = Matrix{T}(undef, length(xs), length(xs))
+    pairwise!(R, dist, xs; preprocess = preprocess)
 end
 
 @doc """
-    pairwise!(r::AbstractMatrix, dist::StringDistance, itr; preprocess = nothing)
+    pairwise!(R::AbstractMatrix, dist::StringDistance, xs::AbstractVector; preprocess = nothing)
-    pairwise!(r::AbstractMatrix, dist::StringDistance, itr1, itr2; preprocess = nothing)
+    pairwise!(R::AbstractMatrix, dist::StringDistance, xs::AbstractVector, ys::AbstractVector; preprocess = nothing)
 
-Compute distances between all pairs of elements in `itr` according to the
+Compute distances between all pairs of elements in `xs` and `ys` according to the
-`StringDistance` `dist` and write the result in `r`.
+`StringDistance` `dist` and write the result in `R`.
 
 For QGramDistances preprocessing will be used either if `preprocess` is set 
-to true or if there are more than 5 elements in `itr`. Set `preprocess` to 
+to true or if there are more than 5 elements in `xs`. Set `preprocess` to 
 false if no preprocessing should be used, regardless of length.
 """
 Distances.pairwise!
 
-function Distances.pairwise!(R::AbstractMatrix, dist::StringDistance, X, Y; preprocess = nothing)
+function Distances.pairwise!(R::AbstractMatrix, dist::StringDistance, xs::AbstractVector, ys::AbstractVector; preprocess = nothing)
-    _asymmetric_pairwise!(R, dist, X, Y; preprocess = preprocess)
+    length(xs) == size(R, 1) || throw(DimensionMismatch("inconsistent length"))
+    length(ys) == size(R, 2) || throw(DimensionMismatch("inconsistent length"))
+    _asymmetric_pairwise!(R, dist, xs, ys; preprocess = preprocess)
 end
 
-function Distances.pairwise!(R::AbstractMatrix, dist::StringDistance, X; preprocess = nothing)
+function Distances.pairwise!(R::AbstractMatrix, dist::StringDistance, xs::AbstractVector; preprocess = nothing)
+    length(xs) == size(R, 1) || throw(DimensionMismatch("inconsistent length"))
+    length(xs) == size(R, 2) || throw(DimensionMismatch("inconsistent length"))
     (dist isa SemiMetric) ?
-        _symmetric_pairwise!(R, dist, X; preprocess = preprocess) :
+        _symmetric_pairwise!(R, dist, xs; preprocess = preprocess) :
-        _asymmetric_pairwise!(R, dist, X, X; preprocess = preprocess)
+        _asymmetric_pairwise!(R, dist, xs, xs; preprocess = preprocess)
 end
 
-function _preprocess(X, dist::QGramDistance, preprocess)
+function _preprocess(xs, dist::QGramDistance, preprocess)
-    if (preprocess === true) || (isnothing(preprocess) && length(X) >= 5)
+    if (preprocess === true) || (isnothing(preprocess) && length(xs) >= 5)
-        return map(x -> QGramSortedVector(x, dist.q), X)
+        return map(x -> x === missing ? x : QGramSortedVector(x, dist.q), xs)
     else
-        return X
+        return xs
     end
 end
-_preprocess(X, dist::StringDistance, preprocess) = X
+_preprocess(xs, dist::StringDistance, preprocess) = xs
 
 
-function _symmetric_pairwise!(R::AbstractMatrix, dist::StringDistance, X; preprocess = nothing)
+function _symmetric_pairwise!(R::AbstractMatrix, dist::StringDistance, xs::AbstractVector; preprocess = nothing)
-    objs = _preprocess(X, dist, preprocess)
+    objs = _preprocess(xs, dist, preprocess)
     for i in 1:length(objs)
-        R[i, i] = 0
+        # handle missing
+        R[i, i] = objs[i] != objs[i]
         Threads.@threads for j in (i+1):length(objs)
             R[i, j] = R[j, i] = evaluate(dist, objs[i], objs[j])
         end
@@ -84,12 +95,12 @@ function _symmetric_pairwise!(R::AbstractMatrix, dist::StringDistance, X; prepro
     return R
 end
 
-function _asymmetric_pairwise!(R::AbstractMatrix, dist::StringDistance, X, Y; preprocess = nothing)
+function _asymmetric_pairwise!(R::AbstractMatrix, dist::StringDistance, xs::AbstractVector, ys::AbstractVector; preprocess = nothing)
-    objsX = _preprocess(X, dist, preprocess)
+    objsxs = _preprocess(xs, dist, preprocess)
-    objsY = _preprocess(Y, dist, preprocess)
+    objsys = _preprocess(ys, dist, preprocess)
-    for i in 1:length(objsX)
+    for i in 1:length(objsxs)
-        Threads.@threads for j in 1:length(objsY)
+        Threads.@threads for j in 1:length(objsys)
-            R[i, j] = evaluate(dist, objsX[i], objsY[j])
+            R[i, j] = evaluate(dist, objsxs[i], objsys[j])
         end
     end
     return R

test/distances.jl

@@ -1,6 +1,13 @@
 using StringDistances, Unicode, Test, Random
 
 @testset "Distances" begin
+	@testset "Hamming" begin
+		@test evaluate(Hamming(), "martha", "marhta") ≈  2
+		@test evaluate(Hamming(), "es an ", " vs an") ≈ 6
+		@test evaluate(Hamming(), [1, 2, 3], [1,2, 4]) ≈ 1
+		@inferred evaluate(Hamming(), "", "")
+		@test ismissing(evaluate(Hamming(), "", missing))
+	end
 
 	@testset "Jaro" begin
 		@test evaluate(Jaro(), "martha", "marhta") ≈  0.05555555555555547

test/pairwise.jl

@@ -1,11 +1,12 @@
 using StringDistances, Unicode, Test, Random
-using StringDistances: pairwise, pairwise!, QGramDistance
-
 @testset "pairwise" begin
 
 TestStrings1 = ["", "abc", "bc", "kitten"]
 TestStrings2 = ["mew", "ab"]
 
+TestStrings1missing = ["", "abc", "bc", missing]
+TestStrings2missing = ["mew", missing]
+
 @testset "pairwise" begin
 	for DT in [Jaro, Levenshtein, DamerauLevenshtein, RatcliffObershelp,
 				QGram, Cosine, Jaccard, SorensenDice, Overlap]
@@ -79,6 +80,9 @@ TestStrings2 = ["mew", "ab"]
 				end
 			end
 		end
+		# ensures missing
+		R5 = pairwise(d, TestStrings1missing; preprocess = true)
+		@test eltype(R5) == Union{result_type(d, String, String), Missing}
 	end
 end