FEAT: improved the speed of the soundex code and fixed it to produc…

…e the same results as PHP.

src/boot/sysobj.reb

@@ -256,6 +256,7 @@ modules: object [
 	httpd:            https://src.rebol.tech/modules/httpd.reb
 	prebol:           https://src.rebol.tech/modules/prebol.reb
 	scheduler:        https://src.rebol.tech/modules/scheduler.reb
+	soundex:          https://src.rebol.tech/modules/soundex.reb
 	spotify:          https://src.rebol.tech/modules/spotify.reb
 	thru-cache:       https://src.rebol.tech/modules/thru-cache.reb
 	to-ascii:         https://src.rebol.tech/modules/to-ascii.reb

src/modules/soundex.reb

@@ -1,78 +1,81 @@
 REBOL [
-    Title: "Soundex"
-    Date: 16-Jul-2024
-    File: %soundex.r
-    Author: "Allen Kamp, Oldes"
-    Purpose: {Soundex Encoding returns similar codes for similar sounding words or names. eg Stephens, Stevens are both S315, Smith and Smythe are both S53. Useful for adding Sounds-like searching to databases}
-    Comment: {
-        This is the basic Soundex algorithm (There are a number of different
-        one floating around)
+	Title: "Soundex"
+	Date: 16-Jul-2024
+	File: %soundex.reb
+	Author: "Allen Kamp, Oldes"
+	Purpose: {Soundex Encoding returns similar codes for similar sounding words or names. eg Stephens, Stevens are both S315, Smith and Smythe are both S53. Useful for adding Sounds-like searching to databases}
+	Comment: {
+		This is the basic Soundex algorithm: https://en.wikipedia.org/wiki/Soundex
 
-        1. Remove vowels, H, W and Y
-        2. Encode each char with its code value
-        3. Remove adjacent duplicate numbers
+		1. Remove vowels, H, W and Y
+		2. Encode each char with its code value
+		3. Remove adjacent duplicate numbers
 
-        4. Return First letter, followed by the next 3 letter's code
-           numbers, if they exist.
+		4. Return First letter, followed by the next 3 letter's code
+		   numbers, if they exist.
 
-        TODO: Other algorithms: Extended Soundex, Metaphone and the LC Cutter table
-    }
-    Language: "English"
-    Email: %allenk--powerup--com--au
-    library: [
-        level: 'intermediate 
-        platform: 'all 
-        type: 'tool 
-        domain: [DB text text-processing] 
-        tested-under: none 
-        support: none 
-        license: none 
-        see-also: none
-    ]
-    Version: 1.1.0
-    Type: module
-    Exports: [soundex]
-    Needs: 3.0.0
-    History: [
-    	17-Jul-1999 @Allen "Initial version"
-    	16-Jul-2024 @Oldes "Ported to Rebol3"
+		TODO: Other algorithms: Extended Soundex, Metaphone and the LC Cutter table
+	}
+	Version: 2.0.0
+	Type: module
+	Name: soundex
+	Exports: [soundex]
+	Needs: 3.0.0
+	History: [
+		17-Jul-1999 @Allen "Initial version"
+		16-Jul-2024 @Oldes "Ported to Rebol3"
 
-    ]
+	]
 ]
 
 soundex: function/with [
-    {Returns the Census Soundex Code for the given string}
-    string [any-string!] "String to Encode"
+	{Returns the Census Soundex Code for the given string}
+	string [any-string!] "String to Encode"
 ][
+	code: make string! 4
+	prev: none
 
-    code: make string! ""
+	if empty? string [return "0000"]
 
-    either all [string? string string <> ""] [
-        string: uppercase trim copy string
-
-        foreach letter string [
-            parse to string! letter [soundex-match | soundex-no-match]
-            if 4 = length? code [break] ;maximum length for code is 4
-        ]
-    ] [
-        return string ; return unchanged
-    ]
-    change code first string ; replace first number with first letter
-    return code
+	foreach letter string [
+		either val: mapping/:letter [ 
+			if val != prev [append code val prev: val]
+		][
+			if find "aeiouhwy" letter [prev: #" "]
+			if empty? code [append code #"0"]
+		]
+		if 4 = length? code [break] ;maximum length for code is 4
+	]
+	change code uppercase first string
+	pad/with code 4 #"0"
+	code
 ][
-	code: val: none
-    ; Create Rules
-    set1: [[#"B" | #"F" | #"P" | #"V"](val: #"1")]
-    set2: [[#"C" | #"G" | #"J" | #"K" | #"Q" | #"S" | #"X" | #"Z"](val: #"2")]
-    set3: [[#"D" | #"T"](val: #"3")]
-    set4: [[#"L"](val: "4")]
-    set5: [[#"M" | #"N"] (val: #"5")]
-    set6: [[#"R"](val: #"6")]
-    ; Append val to code if not a duplicate of previous code val
-    soundex-match: [[set1 | set2 | set3 | set4 | set5 | set6 ] 
-        (if val <> back tail code [append code val]) ]
+	code: val: prev: none 
+	mapping: make map! [
+		;Set1
+		#"B" #"1"
+		#"F" #"1"
+		#"P" #"1"
+		#"V" #"1"
+		;Set2
+		#"C" #"2"
+		#"G" #"2"
+		#"J" #"2"
+		#"K" #"2"
+		#"Q" #"2"
+		#"S" #"2"
+		#"X" #"2"
+		#"Z" #"2"
+		;Set3
+		#"D" #"3"
+		#"T" #"3"
+		;Set4
+		#"L" #"4"
+		;Set5
+		#"M" #"5"
+		#"N" #"5"
+		;Set6
+		#"R" #"6"
+	]
+]
 
-    ; If letter not a matched letter its val is 0, but we only care
-    ; about it if it is the first letter.
-    soundex-no-match: [(if (length? code) = 0 [append code "0"])]
-]

src/tests/test-soundex.r3

@@ -0,0 +1,22 @@
+Rebol [
+	Title:    "Test Soundex function"
+	Date:     16-Jul-2024
+	Author:   "Oldes"
+	File:     %test-soundex.r3
+	Version:  1.0.0
+]
+use [tmp][
+	tmp: none
+	foreach [code name] [
+		"R163" "Robert"
+		"R163" "Rupert"
+		"R150" "Rubin"
+		"A226" "Ashcraft"
+		"A226" "Ashcroft"
+		"T522" "Tymczak" ;; the chars 'z' and 'k' in the name are coded as 2 twice since a vowel lies in between them
+		"P236" "Pfister"
+		"H555" "Honeyman"
+	][
+		printf [5 9 5] reduce [code name tmp: soundex name code == tmp]
+	]
+]