mirrors
/
StringDistances.jl
mirror of https://github.com/matthieugomez/StringDistances.jl.git
1
0
Fork 0
Code Issues Releases Wiki Activity

add tests

pull/1/head
matthieugomez 2015-10-22 12:23:10 -04:00
parent 3ecd33e732
commit 6bb6d706aa
3 changed files with 25 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
README.md
View File

@ -1,2 +1,12 @@
# StringDistances
String Distances in Julia
- Hamming Distance
- Jaro distance
- Jaro-Winkler Distance. Options are `p` (scaling factor), `b` (boosting parameter), and `minlen` (minimum length for long string adjustment)
- Levenshtein distance
To be implemented
- Damerau-Levenshtein Distance
- qgram

16
src/StringDistances.jl
View File

@ -74,9 +74,9 @@ levenshtein(s1::AbstractString, s2::AbstractString) = evaluate(Levenshtein(), s1
##############################################################################
type JaroWinkler{T1 <: Number, T2 <: Number, T3 <: Integer}
p::T1 # scaling factor. Default to 0.1
b::T2 # boost threshold. Default to 0.7
long::T3 # long string adjustment. Default to 5
scaling_factor::T1 # scaling factor. Default to 0.1
boosting_threshold::T2 # boost threshold. Default to 0.7
long_threshold::T3 # long string adjustment. Default to 5
end
function evaluate(dist::JaroWinkler, s1::AbstractString, s2::AbstractString)
@ -108,25 +108,25 @@ function evaluate(dist::JaroWinkler, s1::AbstractString, s2::AbstractString)
score = (m / length(s1) + m / length(s2) + (m - t) / m) / 3.0
# common prefix adjustment
if (dist.p > 0 && score >= dist.b) || (length(s1) >= dist.long)
if (dist.scaling_factor > 0 && score >= dist.boosting_threshold) || (length(s1) >= dist.long_threshold)
l = 0
last = min(4, length(s1))
while l < last && s1[l+1] == s2[l+1]
l += 1
end
# common prefix adjustment
if (dist.p > 0 && score >= dist.b)
score += l * (1 - score) * dist.p
if (dist.scaling_factor > 0 && score >= dist.boosting_threshold)
score += l * (1 - score) * dist.scaling_factor
end
# longer string adjustment
if (length(s1) >= dist.long) && (m - l >= 2) && ((m - l) >= (length(s1) - l) / 2)
if (length(s1) >= dist.long_threshold) && (m - l >= 2) && ((m - l) >= (length(s1) - l) / 2)
score += (1 - score) * (m - (l + 1)) / (length(s1) + length(s2) - (2 * (l - 1)))
end
end
return score
end
jaro_winkler(s1::AbstractString, s2::AbstractString; p = 0.1, b = 0.7, long = 5) = evaluate(JaroWinkler(p, b, long), s1, s2)
jaro_winkler(s1::AbstractString, s2::AbstractString; scaling_factor = 0.1, boosting_threshold = 0.7, long_threshold = 5) = evaluate(JaroWinkler(scaling_factor, boosting_threshold, long_threshold), s1, s2)
jaro(s1::AbstractString, s2::AbstractString) = evaluate(JaroWinkler(0.0, 0.0, 0), s1, s2)
end # module FixedEffectModels

14
test/distances.jl
View File

@ -2,13 +2,13 @@
using Base.Test
@test_approx_eq_eps jaro_winkler("MARTHA", "MARHTA", b = 0.0, long = 100) 0.9611 1e-4
@test_approx_eq_eps jaro_winkler("DWAYNE", "DUANE", b = 0.0, long = 100) 0.84 1e-4
@test_approx_eq_eps jaro_winkler("DIXON", "DICKSONX", b = 0.0, long = 100) 0.81333 1e-4
@test_approx_eq_eps jaro_winkler("William", "Williams", b = 0.0, long = 100) 0.975 1e-4
@test_approx_eq_eps jaro_winkler("", "foo", b = 0.0, long = 100) 0.0 1e-4
@test_approx_eq_eps jaro_winkler("a", "a", b = 0.0, long = 100) 1.0 1e-4
@test_approx_eq_eps jaro_winkler("abc", "xyz", b = 0.0, long = 100) 0.0 1e-4
@test_approx_eq_eps jaro_winkler("MARTHA", "MARHTA", boosting_threshold = 0.0, long_threshold = 100) 0.9611 1e-4
@test_approx_eq_eps jaro_winkler("DWAYNE", "DUANE", boosting_threshold = 0.0, long_threshold = 100) 0.84 1e-4
@test_approx_eq_eps jaro_winkler("DIXON", "DICKSONX", boosting_threshold = 0.0, long_threshold = 100) 0.81333 1e-4
@test_approx_eq_eps jaro_winkler("William", "Williams", boosting_threshold = 0.0, long_threshold = 100) 0.975 1e-4
@test_approx_eq_eps jaro_winkler("", "foo", boosting_threshold = 0.0, long_threshold = 100) 0.0 1e-4
@test_approx_eq_eps jaro_winkler("a", "a", boosting_threshold = 0.0, long_threshold = 100) 1.0 1e-4
@test_approx_eq_eps jaro_winkler("abc", "xyz", boosting_threshold = 0.0, long_threshold = 100) 0.0 1e-4