From dad3399c32f066ebbbd18c9271b29a76dc45e3c0 Mon Sep 17 00:00:00 2001 From: Rocky Madden Date: Sat, 28 Dec 2013 13:55:50 -0700 Subject: Favored functional composition. --- .../stringmetric/phonetic/MetaphoneAlgorithm.scala | 19 +++++------ .../stringmetric/phonetic/MetaphoneMetric.scala | 4 +-- .../stringmetric/phonetic/NysiisAlgorithm.scala | 39 +++++++++++----------- .../stringmetric/phonetic/NysiisMetric.scala | 6 ++-- .../phonetic/RefinedNysiisAlgorithm.scala | 37 ++++++++++---------- .../phonetic/RefinedNysiisMetric.scala | 4 +-- .../phonetic/RefinedSoundexAlgorithm.scala | 2 +- .../phonetic/RefinedSoundexMetric.scala | 4 +-- .../stringmetric/phonetic/SoundexAlgorithm.scala | 2 +- .../stringmetric/phonetic/SoundexMetric.scala | 4 +-- .../similarity/DiceSorensenMetric.scala | 2 +- .../stringmetric/similarity/HammingMetric.scala | 2 +- .../stringmetric/similarity/JaroMetric.scala | 8 +++-- .../similarity/LevenshteinMetric.scala | 38 ++++++++++----------- .../stringmetric/similarity/NGramMetric.scala | 2 +- .../stringmetric/similarity/OverlapMetric.scala | 2 +- .../similarity/RatcliffObershelpMetric.scala | 6 ++-- .../similarity/WeightedLevenshteinMetric.scala | 33 +++++++++--------- .../stringmetric/tokenize/NGramTokenizer.scala | 2 +- 19 files changed, 110 insertions(+), 106 deletions(-) diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/MetaphoneAlgorithm.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/MetaphoneAlgorithm.scala index 126f170..64c16db 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/MetaphoneAlgorithm.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/MetaphoneAlgorithm.scala @@ -8,7 +8,7 @@ case object MetaphoneAlgorithm extends StringAlgorithmLike { override def compute(a: Array[Char]): Option[Array[Char]] = if (a.length == 0 || !(Alpha isSuperset a.head)) None else { - val th = deduplicate(transcodeHead(a.map(_.toLower))) + val th = (transcodeHead andThen deduplicate)(a.map(_.toLower)) val t = transcode(Array.empty[Char], th.head, th.tail, Array.empty[Char]) if (t.length == 0) None else Some(t) // Single Y or W would have 0 length. @@ -16,12 +16,12 @@ case object MetaphoneAlgorithm extends StringAlgorithmLike { override def compute(a: String): Option[String] = compute(a.toCharArray).map(_.mkString) - private def deduplicate(ca: Array[Char]) = + private val deduplicate: (Array[Char] => Array[Char]) = (ca) => if (ca.length <= 1) ca - else ca.sliding(2).withFilter(a => a(0) == 'c' || a(0) != a(1)).map(a => a(0)).toArray[Char] :+ ca.last + else ca.sliding(2).withFilter(a => a(0) == 'c' || a(0) != a(1)).map(_(0)).toArray[Char] :+ ca.last @annotation.tailrec - private def transcode(l: Array[Char], c: Char, r: Array[Char], o: Array[Char]): Array[Char] = { + private val transcode: ((Array[Char], Char, Array[Char], Array[Char]) => Array[Char]) = (l, c, r, o) => { if (c == '\0' && r.length == 0) o else { def shift(d: Int, ca: Array[Char]) = { @@ -89,19 +89,18 @@ case object MetaphoneAlgorithm extends StringAlgorithmLike { } } - private def transcodeHead(ca: Array[Char]) = { + private val transcodeHead: (Array[Char] => Array[Char]) = (ca) => (ca.length: @annotation.switch) match { case 0 => ca case 1 => if (ca.head == 'x') Array('s') else ca case _ => (ca.head: @annotation.switch) match { - case 'a' if (ca(1) == 'e') => ca.tail - case 'g' | 'k' | 'p' if (ca(1) == 'n') => ca.tail - case 'w' if (ca(1) == 'r') => ca.tail - case 'w' if (ca(1) == 'h') => 'w' +: ca.drop(2) + case 'a' if ca(1) == 'e' => ca.tail + case 'g' | 'k' | 'p' if ca(1) == 'n' => ca.tail + case 'w' if ca(1) == 'r' => ca.tail + case 'w' if ca(1) == 'h' => 'w' +: ca.drop(2) case 'x' => 's' +: ca.tail case _ => ca } } - } } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/MetaphoneMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/MetaphoneMetric.scala index 083016c..7d9cd1d 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/MetaphoneMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/MetaphoneMetric.scala @@ -7,9 +7,9 @@ case object MetaphoneMetric extends StringMetricLike[Boolean] { override def compare(a: Array[Char], b: Array[Char]): Option[Boolean] = if (a.length == 0 || !(Alpha isSuperset a.head) || b.length == 0 || !(Alpha isSuperset b.head)) None - else MetaphoneAlgorithm.compute(a).filter(_.length > 0).flatMap(mp1 => + else MetaphoneAlgorithm.compute(a).filter(_.length > 0).flatMap { mp1 => MetaphoneAlgorithm.compute(b).filter(_.length > 0).map(mp1.sameElements(_)) - ) + } override def compare(a: String, b: String): Option[Boolean] = compare(a.toCharArray, b.toCharArray) } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/NysiisAlgorithm.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/NysiisAlgorithm.scala index 43c2bc2..c9245e0 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/NysiisAlgorithm.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/NysiisAlgorithm.scala @@ -20,26 +20,27 @@ case object NysiisAlgorithm extends StringAlgorithmLike { ) ++ tr._2 if (t.length == 1) Some(t) - else Some(t.head +: deduplicate(cleanTerminal(cleanLast(t.tail)))) + else Some(t.head +: (cleanLast andThen cleanTerminal andThen deduplicate)(t.tail)) } override def compute(a: String): Option[String] = compute(a.toCharArray).map(_.mkString) - private def cleanLast(ca: Array[Char]) = + private val cleanLast: (Array[Char] => Array[Char]) = (ca) => if (ca.length == 0) ca - else if(ca.last == 'a' || ca.last == 's') ca.dropRight(ca.reverseIterator.takeWhile(c => c == 'a' || c == 's').length) + else if(ca.last == 'a' || ca.last == 's') + ca.dropRight(ca.reverseIterator.takeWhile(c => c == 'a' || c == 's').length) else ca - private def cleanTerminal(ca: Array[Char]) = + private val cleanTerminal: (Array[Char] => Array[Char]) = (ca) => if (ca.length >= 2 && ca.last == 'y' && ca(ca.length - 2) == 'a') ca.dropRight(2) :+ 'y' else ca - private def deduplicate(ca: Array[Char]) = + private val deduplicate: (Array[Char] => Array[Char]) = (ca) => if (ca.length <= 1) ca - else ca.sliding(2).withFilter(a => a(0) != a(1)).map(a => a(0)).toArray[Char] :+ ca.last + else ca.sliding(2).withFilter(a => a(0) != a(1)).map(_(0)).toArray[Char] :+ ca.last @annotation.tailrec - private def transcodeCenter(l: Array[Char], c: Char, r: Array[Char], o: Array[Char]): Array[Char] = { + private val transcodeCenter: ((Array[Char], Char, Array[Char], Array[Char]) => Array[Char]) = (l, c, r, o) => if (c == '\0' && r.length == 0) o else { def shift(d: Int, ca: Array[Char]) = { @@ -61,7 +62,8 @@ case object NysiisAlgorithm extends StringAlgorithmLike { if (r.length >= 1 && r.head == 'v') shift(2, o ++ Array('a', 'f')) else shift(1, o :+ 'a') case 'h' => - if (l.length >= 1 && (!(LowercaseVowel isSuperset l.last) || (r.length >= 1 && !(LowercaseVowel isSuperset r.head)))) shift(1, o) + if (l.length >= 1 && (!(LowercaseVowel isSuperset l.last) || (r.length >= 1 && !(LowercaseVowel isSuperset r.head)))) + shift(1, o) else shift(1, o :+ c) case 'k' => if (r.length >= 1 && r.head == 'n') shift(2, o :+ 'n') else shift(1, o :+ 'c') case 'm' => shift(1, o :+ 'n') @@ -80,37 +82,34 @@ case object NysiisAlgorithm extends StringAlgorithmLike { transcodeCenter(t._1, t._2, t._3, t._4) } - } - private def transcodeLeft(ca: Array[Char]) = { + private val transcodeLeft: (Array[Char] => (Array[Char], Array[Char])) = (ca) => if (ca.length == 0) (Array.empty[Char], ca) else { lazy val tr2 = ca.takeRight(ca.length - 2) lazy val tr3 = ca.takeRight(ca.length - 3) (ca.head: @annotation.switch) match { - case 'k' if (ca.length >= 2 && ca(1) == 'n') => (Array('n', 'n'), tr2) + case 'k' if ca.length >= 2 && ca(1) == 'n' => (Array('n', 'n'), tr2) case 'k' => (Array('c'), ca.tail) - case 'm' if (ca.length >= 3 && (ca(1) == 'a' && ca(2) == 'c')) => (Array('m', 'c'), tr3) - case 'p' if (ca.length >= 2 && (ca(1) == 'h' || ca(1) == 'f')) => (Array('f', 'f'), tr2) - case 's' if (ca.length >= 3 && (ca(1) == 'c' && ca(2) == 'h')) => (Array('s', 's'), tr3) + case 'm' if ca.length >= 3 && (ca(1) == 'a' && ca(2) == 'c') => (Array('m', 'c'), tr3) + case 'p' if ca.length >= 2 && (ca(1) == 'h' || ca(1) == 'f') => (Array('f', 'f'), tr2) + case 's' if ca.length >= 3 && (ca(1) == 'c' && ca(2) == 'h') => (Array('s', 's'), tr3) case _ => (Array(ca.head), ca.tail) } } - } - private def transcodeRight(ca: Array[Char]) = { + private val transcodeRight: (Array[Char] => (Array[Char], Array[Char])) = (ca) => if (ca.length >= 2) { val lc = ca(ca.length - 1) val lcm1 = ca(ca.length - 2) lazy val t2 = ca.take(ca.length - 2) (lc: @annotation.switch) match { - case 'd' if (lcm1 == 'n' || lcm1 == 'r') => (t2, Array('d')) - case 'e' if (lcm1 == 'e' || lcm1 == 'i') => (t2, Array('y')) - case 't' if (lcm1 == 'd' || lcm1 == 'n' || lcm1 == 'r') => (t2, Array('d')) + case 'd' if lcm1 == 'n' || lcm1 == 'r' => (t2, Array('d')) + case 'e' if lcm1 == 'e' || lcm1 == 'i' => (t2, Array('y')) + case 't' if lcm1 == 'd' || lcm1 == 'n' || lcm1 == 'r' => (t2, Array('d')) case _ => (ca, Array.empty[Char]) } } else (ca, Array.empty[Char]) - } } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/NysiisMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/NysiisMetric.scala index 6316981..84c2073 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/NysiisMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/NysiisMetric.scala @@ -6,7 +6,7 @@ case object NysiisMetric extends StringMetricLike[Boolean] { import com.rockymadden.stringmetric.Alphabet.Alpha override def compare(a: Array[Char], b: Array[Char]): Option[Boolean] = { - val unequal = (c1: Char, c2: Char) => { + val unequal: ((Char, Char) => Boolean) = (c1, c2) => { val lc1 = c1.toLower val lc2 = c2.toLower @@ -15,9 +15,9 @@ case object NysiisMetric extends StringMetricLike[Boolean] { if (a.length == 0 || !(Alpha isSuperset a.head) || b.length == 0 || !(Alpha isSuperset b.head)) None else if (unequal(a.head, b.head)) Some(false) - else NysiisAlgorithm.compute(a).filter(_.length > 0).flatMap(ny1 => + else NysiisAlgorithm.compute(a).filter(_.length > 0).flatMap { ny1 => NysiisAlgorithm.compute(b).filter(_.length > 0).map(ny1.sameElements(_)) - ) + } } override def compare(a: String, b: String): Option[Boolean] = compare(a.toCharArray, b.toCharArray) diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedNysiisAlgorithm.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedNysiisAlgorithm.scala index 72bd84e..c35175d 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedNysiisAlgorithm.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedNysiisAlgorithm.scala @@ -9,30 +9,32 @@ case object RefinedNysiisAlgorithm extends StringAlgorithmLike { if (a.length == 0 || !(Alpha isSuperset a.head)) None else { val lca = a.map(_.toLower) - val tlh = transcodeLast(transcodeHead(lca.head +: cleanLast(lca.tail, Set('s', 'z')))) + val tlh = (transcodeHead andThen transcodeLast)(lca.head +: cleanLast(lca.tail, Set('s', 'z'))) val t = transcode(Array.empty[Char], tlh.head, tlh.tail, Array.empty[Char]) if (t.length == 1) Some(t) - else Some(deduplicate(t.head +: cleanTerminal(cleanLast(t.tail, Set('a'))))) + else Some(deduplicate( + t.head +: (cleanLast.tupled andThen cleanTerminal)(t.tail, Set('a')) + )) } override def compute(string: String): Option[String] = compute(string.toCharArray).map(_.mkString) - private def cleanLast(ca: Array[Char], s: Set[Char]) = + private val cleanLast: ((Array[Char], Set[Char]) => Array[Char]) = (ca, s) => if (ca.length == 0) ca else if(s.contains(ca.last)) ca.dropRight(ca.reverseIterator.takeWhile(c => s.contains(c)).length) else ca - private def cleanTerminal(ca: Array[Char]) = + private val cleanTerminal: (Array[Char] => Array[Char]) = (ca) => if (ca.length >= 2 && ca.last == 'y' && ca(ca.length - 2) == 'a') ca.dropRight(2) :+ 'y' else ca - private def deduplicate(ca: Array[Char]) = + private val deduplicate: (Array[Char] => Array[Char]) = (ca) => if (ca.length <= 1) ca else ca.sliding(2).withFilter(a => a(0) != a(1)).map(a => a(0)).toArray[Char] :+ ca.last @annotation.tailrec - private def transcode(l: Array[Char], c: Char, r: Array[Char], o: Array[Char]): Array[Char] = { + private val transcode: ((Array[Char], Char, Array[Char], Array[Char]) => Array[Char]) = (l, c, r, o) => if (c == '\0' && r.length == 0) o else { def shift(d: Int, ca: Array[Char]) = { @@ -63,7 +65,8 @@ case object RefinedNysiisAlgorithm extends StringAlgorithmLike { else shift(1, o :+ c) case 'h' => if (l.length == 0) shift(1, o :+ c) - else if (!(LowercaseVowel isSuperset l.last) || (r.length >= 1 && !(LowercaseVowel isSuperset r.head))) shift(1, o) + else if (!(LowercaseVowel isSuperset l.last) || (r.length >= 1 && !(LowercaseVowel isSuperset r.head))) + shift(1, o) else shift(1, o :+ c) case 'k' => if (r.length >= 1 && r.head == 'n') shift(2, o :+ 'n') else shift(1, o :+ 'c') case 'm' => if (l.length == 0) shift(1, o :+ c) else shift(1, o :+ 'n') @@ -89,29 +92,29 @@ case object RefinedNysiisAlgorithm extends StringAlgorithmLike { transcode(t._1, t._2, t._3, t._4) } - } - private def transcodeHead(ca: Array[Char]) = + private val transcodeHead: (Array[Char] => Array[Char]) = (ca) => if (ca.length == 0) ca else (ca.head: @annotation.switch) match { - case 'm' if (ca.length >= 3 && ca(1) == 'a' && ca(2) == 'c') => Array('m', 'c') ++ ca.takeRight(ca.length - 3) - case 'p' if (ca.length >= 2 && ca(1) == 'f') => 'f' +: ca.takeRight(ca.length - 2) + case 'm' if ca.length >= 3 && ca(1) == 'a' && ca(2) == 'c' => + Array('m', 'c') ++ ca.takeRight(ca.length - 3) + case 'p' if ca.length >= 2 && ca(1) == 'f' =>'f' +: ca.takeRight(ca.length - 2) case _ => ca } - private def transcodeLast(ca: Array[Char]) = + private val transcodeLast: (Array[Char] => Array[Char]) = (ca) => if (ca.length >= 2) { val lc = ca(ca.length - 1) val lcm1 = ca(ca.length - 2) lazy val t2 = ca.take(ca.length - 2) (lc: @annotation.switch) match { - case 'd' if (lcm1 == 'n' || lcm1 == 'r') => t2 :+ 'd' - case 'e' if (lcm1 == 'e' || lcm1 == 'i' || lcm1 =='y') => t2 :+ 'y' - case 't' if (lcm1 == 'd' || lcm1 == 'n' || lcm1 == 'r') => t2 :+ 'd' - case 'x' if (lcm1 == 'e') => t2 ++ Array('e', 'c') - case 'x' if (lcm1 == 'i') => t2 ++ Array('i', 'c') + case 'd' if lcm1 == 'n' || lcm1 == 'r' => t2 :+ 'd' + case 'e' if lcm1 == 'e' || lcm1 == 'i' || lcm1 =='y' => t2 :+ 'y' + case 't' if lcm1 == 'd' || lcm1 == 'n' || lcm1 == 'r' => t2 :+ 'd' + case 'x' if lcm1 == 'e' => t2 ++ Array('e', 'c') + case 'x' if lcm1 == 'i' => t2 ++ Array('i', 'c') case _ => ca } } else ca diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedNysiisMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedNysiisMetric.scala index 73795a0..86111e4 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedNysiisMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedNysiisMetric.scala @@ -15,9 +15,9 @@ case object RefinedNysiisMetric extends StringMetricLike[Boolean] { if (a.length == 0 || !(Alpha isSuperset a.head) || b.length == 0 || !(Alpha isSuperset b.head)) None else if (unequal(a.head, b.head)) Some(false) - else RefinedNysiisAlgorithm.compute(a).filter(_.length > 0).flatMap(rny1 => + else RefinedNysiisAlgorithm.compute(a).filter(_.length > 0).flatMap { rny1 => RefinedNysiisAlgorithm.compute(b).filter(_.length > 0).map(rny1.sameElements(_)) - ) + } } override def compare(a: String, b: String): Option[Boolean] = compare(a.toCharArray, b.toCharArray) diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedSoundexAlgorithm.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedSoundexAlgorithm.scala index 9f7fce9..7fb6dfc 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedSoundexAlgorithm.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedSoundexAlgorithm.scala @@ -12,7 +12,7 @@ case object RefinedSoundexAlgorithm extends StringAlgorithmLike { override def compute(a: String): Option[String] = compute(a.toCharArray).map(_.mkString) @annotation.tailrec - private def transcode(i: Array[Char], o: Array[Char]): Array[Char] = + private val transcode: ((Array[Char], Array[Char]) => Array[Char]) = (i, o) => if (i.length == 0) o else { val c = i.head.toLower diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedSoundexMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedSoundexMetric.scala index 5ad0e30..a7d3984 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedSoundexMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/RefinedSoundexMetric.scala @@ -8,9 +8,9 @@ case object RefinedSoundexMetric extends StringMetricLike[Boolean] { override def compare(a: Array[Char], b: Array[Char]): Option[Boolean] = if (a.length == 0 || !(Alpha isSuperset a.head) || b.length == 0 || !(Alpha isSuperset b.head)) None else if (a.head.toLower != b.head.toLower) Some(false) - else RefinedSoundexAlgorithm.compute(a).filter(_.length > 0).flatMap(rse1 => + else RefinedSoundexAlgorithm.compute(a).filter(_.length > 0).flatMap { rse1 => RefinedSoundexAlgorithm.compute(b).filter(_.length > 0).map(rse1.sameElements(_)) - ) + } override def compare(a: String, b: String): Option[Boolean] = compare(a.toCharArray, b.toCharArray) } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/SoundexAlgorithm.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/SoundexAlgorithm.scala index d615144..98898e9 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/SoundexAlgorithm.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/SoundexAlgorithm.scala @@ -16,7 +16,7 @@ case object SoundexAlgorithm extends StringAlgorithmLike { override def compute(string: String): Option[String] = compute(string.toCharArray).map(_.mkString) @annotation.tailrec - private def transcode(i: Array[Char], pc: Char, o: Array[Char]): Array[Char] = + private val transcode: ((Array[Char], Char, Array[Char]) => Array[Char]) = (i, pc, o) => if (i.length == 0) o else { val c = i.head.toLower diff --git a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/SoundexMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/SoundexMetric.scala index 7e0bf5c..29e2606 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/phonetic/SoundexMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/phonetic/SoundexMetric.scala @@ -8,9 +8,9 @@ case object SoundexMetric extends StringMetricLike[Boolean] { override def compare(a: Array[Char], b: Array[Char]): Option[Boolean] = if (a.length == 0 || !(Alpha isSuperset a.head) || b.length == 0 || !(Alpha isSuperset b.head)) None else if (a.head.toLower != b.head.toLower) Some(false) - else SoundexAlgorithm.compute(a).filter(_.length > 0).flatMap(se1 => + else SoundexAlgorithm.compute(a).filter(_.length > 0).flatMap { se1 => SoundexAlgorithm.compute(b).filter(_.length > 0).map(se1.sameElements(_)) - ) + } final override def compare(a: String, b: String): Option[Boolean] = compare(a.toCharArray, b.toCharArray) } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/similarity/DiceSorensenMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/similarity/DiceSorensenMetric.scala index 21c9f16..2c16bba 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/similarity/DiceSorensenMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/similarity/DiceSorensenMetric.scala @@ -26,5 +26,5 @@ final case class DiceSorensenMetric(private val n: Int) extends StringMetricLike override def compare(a: String, b: String): Option[Double] = compare(a.toCharArray, b.toCharArray) - private[this] def scoreMatches(mt: MatchTuple[String]) = mt._1.intersect(mt._2).length + private val scoreMatches: (MatchTuple[String] => Int) = (mt) => mt._1.intersect(mt._2).length } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/similarity/HammingMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/similarity/HammingMetric.scala index 09b62bc..f1527d2 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/similarity/HammingMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/similarity/HammingMetric.scala @@ -12,7 +12,7 @@ case object HammingMetric extends StringMetricLike[Int] { override def compare(a: String, b: String): Option[Int] = compare(a.toCharArray, b.toCharArray) - private def hamming(ct: CompareTuple[Char]) = + private val hamming: (CompareTuple[Char] => Int) = (ct) => if (ct._1.length == 0) 0 else ct._1.zip(ct._2).count(t => t._1 != t._2) } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/similarity/JaroMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/similarity/JaroMetric.scala index b9d1434..c58b864 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/similarity/JaroMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/similarity/JaroMetric.scala @@ -1,6 +1,7 @@ package com.rockymadden.stringmetric.similarity import com.rockymadden.stringmetric.Metric.StringMetricLike +import scala.Some /** * An implementation of the Jaro metric. One differing detail in this implementation is that if a character is matched @@ -27,7 +28,7 @@ case object JaroMetric extends StringMetricLike[Double] { override def compare(a: String, b: String): Option[Double] = compare(a.toCharArray, b.toCharArray) - private def `match`(ct: CompareTuple[Char]): MatchTuple[Char] = { + private val `match`: (CompareTuple[Char] => MatchTuple[Char]) = (ct) => { lazy val window = math.abs((math.max(ct._1.length, ct._2.length) / 2d).floor.toInt - 1) val one = ArrayBuffer.empty[Int] val two = HashSet.empty[Int] @@ -58,7 +59,8 @@ case object JaroMetric extends StringMetricLike[Double] { (one.toArray.map(ct._1(_)), two.toArray.sortWith(_ < _).map(ct._2(_))) } - private def scoreMatches(mt: MatchTuple[Char]) = mt._1.length + private val scoreMatches: (MatchTuple[Char] => Int) = (mt) => mt._1.length - private def scoreTranspositions(mt: MatchTuple[Char]) = (mt._1.zip(mt._2).count(t => t._1 != t._2) / 2d).floor.toInt + private val scoreTranspositions: (MatchTuple[Char] => Int) = (mt) => + (mt._1.zip(mt._2).count(t => t._1 != t._2) / 2d).floor.toInt } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/similarity/LevenshteinMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/similarity/LevenshteinMetric.scala index 9f78aed..e145e1f 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/similarity/LevenshteinMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/similarity/LevenshteinMetric.scala @@ -12,28 +12,26 @@ case object LevenshteinMetric extends StringMetricLike[Int] { override def compare(a: String, b: String): Option[Int] = compare(a.toCharArray, b.toCharArray) - private def levenshtein(ct: CompareTuple[Char]) = { + private val levenshtein: (CompareTuple[Char] => Int) = (ct) => { val m = Array.fill[Int](ct._1.length + 1, ct._2.length + 1)(-1) - def distance(t: (Int, Int)): Int = { - t match { - case (r, 0) => r - case (0, c) => c - case (r, c) if m(r)(c) != -1 => m(r)(c) - case (r, c) => { - val min = - if (ct._1(r - 1) == ct._2(c - 1)) distance(r - 1, c - 1) - else math.min( - math.min( - distance(r - 1, c) + 1, // Delete (left). - distance(r, c - 1) + 1 // Insert (up). - ), - distance(r - 1, c - 1) + 1 // Substitute (left-up). - ) - - m(r)(c) = min - min - } + def distance(t: (Int, Int)): Int = t match { + case (r, 0) => r + case (0, c) => c + case (r, c) if m(r)(c) != -1 => m(r)(c) + case (r, c) => { + val min = + if (ct._1(r - 1) == ct._2(c - 1)) distance(r - 1, c - 1) + else math.min( + math.min( + distance(r - 1, c) + 1, // Delete (left). + distance(r, c - 1) + 1 // Insert (up). + ), + distance(r - 1, c - 1) + 1 // Substitute (left-up). + ) + + m(r)(c) = min + min } } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/similarity/NGramMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/similarity/NGramMetric.scala index f6397f1..2b2fc01 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/similarity/NGramMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/similarity/NGramMetric.scala @@ -23,5 +23,5 @@ final case class NGramMetric(private val n: Int) extends StringMetricLike[Double override def compare(a: String, b: String): Option[Double] = compare(a.toCharArray, b.toCharArray) - private[this] def scoreMatches(mt: MatchTuple[String]) = mt._1.intersect(mt._2).length + private val scoreMatches: (MatchTuple[String] => Int) = (mt) => mt._1.intersect(mt._2).length } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/similarity/OverlapMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/similarity/OverlapMetric.scala index f4c66b7..b6b3bb9 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/similarity/OverlapMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/similarity/OverlapMetric.scala @@ -23,5 +23,5 @@ final case class OverlapMetric(private val n: Int) extends StringMetricLike[Doub override def compare(a: String, b: String): Option[Double] = compare(a.toCharArray, b.toCharArray) - private[this] def scoreMatches(mt: MatchTuple[String]) = mt._1.intersect(mt._2).length + private val scoreMatches: (MatchTuple[String] => Int) = (mt) => mt._1.intersect(mt._2).length } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/similarity/RatcliffObershelpMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/similarity/RatcliffObershelpMetric.scala index 197e59a..1562f79 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/similarity/RatcliffObershelpMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/similarity/RatcliffObershelpMetric.scala @@ -27,7 +27,7 @@ case object RatcliffObershelpMetric extends StringMetricLike[Double] { lrc } - private def commonSequences(ct: CompareTuple[Char]): Array[Array[Char]] = { + private val commonSequences: (CompareTuple[Char] => Array[Array[Char]]) = (ct) => { val lcs = longestCommonSubsequence(ct) if (lcs._1 == 0) Array.empty @@ -35,7 +35,9 @@ case object RatcliffObershelpMetric extends StringMetricLike[Double] { val sct1 = (ct._1.take(lcs._2 - lcs._1), ct._1.takeRight(ct._1.length - lcs._2)) val sct2 = (ct._2.take(lcs._3 - lcs._1), ct._2.takeRight(ct._2.length - lcs._3)) - Array(ct._1.slice(lcs._2 - lcs._1, lcs._2)) ++ commonSequences(sct1._1, sct2._1) ++ commonSequences(sct1._2, sct2._2) + Array(ct._1.slice(lcs._2 - lcs._1, lcs._2)) ++ + commonSequences(sct1._1, sct2._1) ++ + commonSequences(sct1._2, sct2._2) } } } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/similarity/WeightedLevenshteinMetric.scala b/core/source/main/scala/com/rockymadden/stringmetric/similarity/WeightedLevenshteinMetric.scala index 2564eb7..74fb320 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/similarity/WeightedLevenshteinMetric.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/similarity/WeightedLevenshteinMetric.scala @@ -14,22 +14,23 @@ final case class WeightedLevenshteinMetric(delete: BigDecimal, insert: BigDecima override def compare(a: String, b: String): Option[Double] = compare(a.toCharArray, b.toCharArray) - private[this] def weightedLevenshtein(ct: CompareTuple[Char], w: (BigDecimal, BigDecimal, BigDecimal)) = { - val m = Array.ofDim[BigDecimal](ct._1.length + 1, ct._2.length + 1) - - for (r <- 0 to ct._1.length) m(r)(0) = w._1 * r - for (c <- 0 to ct._2.length) m(0)(c) = w._2 * c - - for (r <- 1 to ct._1.length; c <- 1 to ct._2.length) { - m(r)(c) = - if (ct._1(r - 1) == ct._2(c - 1)) m(r - 1)(c - 1) - else (m(r - 1)(c) + w._1).min( // Delete (left). - (m(r)(c - 1) + w._2).min( // Insert (up). - m(r - 1)(c - 1) + w._3 // Substitute (left-up). + private val weightedLevenshtein: ((CompareTuple[Char], (BigDecimal, BigDecimal, BigDecimal)) => BigDecimal) = + (ct, w) => { + val m = Array.ofDim[BigDecimal](ct._1.length + 1, ct._2.length + 1) + + for (r <- 0 to ct._1.length) m(r)(0) = w._1 * r + for (c <- 0 to ct._2.length) m(0)(c) = w._2 * c + + for (r <- 1 to ct._1.length; c <- 1 to ct._2.length) { + m(r)(c) = + if (ct._1(r - 1) == ct._2(c - 1)) m(r - 1)(c - 1) + else (m(r - 1)(c) + w._1).min( // Delete (left). + (m(r)(c - 1) + w._2).min( // Insert (up). + m(r - 1)(c - 1) + w._3 // Substitute (left-up). + ) ) - ) - } + } - m(ct._1.length)(ct._2.length) - } + m(ct._1.length)(ct._2.length) + } } diff --git a/core/source/main/scala/com/rockymadden/stringmetric/tokenize/NGramTokenizer.scala b/core/source/main/scala/com/rockymadden/stringmetric/tokenize/NGramTokenizer.scala index 09d2a42..e2a2731 100755 --- a/core/source/main/scala/com/rockymadden/stringmetric/tokenize/NGramTokenizer.scala +++ b/core/source/main/scala/com/rockymadden/stringmetric/tokenize/NGramTokenizer.scala @@ -13,7 +13,7 @@ final case class NGramTokenizer(private val n: Int) extends StringTokenizerLike override def tokenize(a: String): Option[Array[String]] = tokenize(a.toCharArray).map(_.map(_.mkString)) @annotation.tailrec - private[this] def sequence(i: Array[Char], o: Array[Array[Char]], n: Int): Array[Array[Char]] = + private val sequence: ((Array[Char], Array[Array[Char]], Int) => Array[Array[Char]]) = (i, o, n) => if (i.length <= n) o :+ i else sequence(i.tail, o :+ i.take(n), n) } -- cgit v1.2.3