version 2.1.0

- much faster processing of strings containing multiple words - move utility functions to a separate file - cleaned up testing
cityssm · May 14, 2021 · aa7b97d · aa7b97d
1 parent d78551f
commit aa7b97d
Show file tree

Hide file tree

Showing 14 changed files with 966 additions and 528 deletions.
diff --git a/README.md b/README.md
@@ -24,13 +24,13 @@ npm install @cityssm/unleet
 
 ```javascript
 unleet("b@d w0rd");
-= [ "bad word" ]
+// => [ "bad word" ]
 
 unleet("1337 $P33K");
-= [ 'ieet zpeek', 'ieet speek', 'leet zpeek', 'leet speek' ]
+// => [ 'ieet zpeek', 'ieet speek', 'leet zpeek', 'leet speek' ]
 
 unleet("0rg@ni℠");
-= [ 'organizm', 'organism' ]
+// => [ 'organizm', 'organism' ]
 ```
 
 ## Contributing

diff --git a/index.d.ts b/index.d.ts
@@ -1 +1,2 @@
-export declare const unleet: (leetString: string) => string[];
+export declare const unleet: (leetString: string | number) => string[];
+export default unleet;
diff --git a/index.js b/index.js
@@ -1,35 +1,10 @@
 import diacritic from "diacritic";
-import { leetSymbolTranslationKeys, simpleTranslations, complexTranslations } from "./translations/translations.js";
-const indiciesOf = (sourceString, searchString) => {
-    const indicies = [];
-    for (let index = 0; index < sourceString.length - searchString.length; index += 1) {
-        if (sourceString.substring(index, searchString.length + index) === searchString) {
-            indicies.push(index);
-        }
-    }
-    return indicies;
-};
-const isLetter = (potentialLetter) => {
-    if ("abcdefghijklmnopqrstuvwxyz".includes(potentialLetter)) {
-        return true;
-    }
-    return false;
-};
-const isPotentialLeet = (potentialLeetString) => {
-    for (const leetSymbol of leetSymbolTranslationKeys) {
-        if (isLetter(leetSymbol)) {
-            continue;
-        }
-        if (potentialLeetString.includes(leetSymbol)) {
-            return true;
-        }
-    }
-    return false;
-};
+import * as utils from "./utils.js";
+import { simpleTranslations, complexTranslations } from "./translations/translations.js";
 const unleetRecurse = (lowerCaseLeetString, unleetStrings, previousStrings, complexTranslationKeys) => {
     for (const leetSymbol of complexTranslationKeys) {
         if (lowerCaseLeetString.includes(leetSymbol)) {
-            let matchingIndicies = indiciesOf(lowerCaseLeetString, leetSymbol);
+            let matchingIndicies = utils.indiciesOf(lowerCaseLeetString, leetSymbol);
             if (matchingIndicies.length === 0) {
                 matchingIndicies = [lowerCaseLeetString.indexOf(leetSymbol)];
             }
@@ -46,7 +21,7 @@ const unleetRecurse = (lowerCaseLeetString, unleetStrings, previousStrings, comp
             }
         }
     }
-    if (!isPotentialLeet(lowerCaseLeetString)) {
+    if (!utils.isPotentialLeet(lowerCaseLeetString)) {
         unleetStrings.add(lowerCaseLeetString);
         return unleetStrings;
     }
@@ -56,7 +31,7 @@ export const unleet = (leetString) => {
     if (leetString === null || leetString === undefined || leetString === "") {
         return [""];
     }
-    let cleanLeetString = (leetString + "").toLowerCase();
+    let cleanLeetString = leetString.toString().toLowerCase();
     cleanLeetString = cleanLeetString.replace(/\./g, " ");
     cleanLeetString = cleanLeetString.replace(/ +/g, " ");
     cleanLeetString = diacritic.clean(cleanLeetString);
@@ -68,5 +43,14 @@ export const unleet = (leetString) => {
     const complexTranslationKeys = Object.keys(complexTranslations).filter(function (leetSymbol) {
         return cleanLeetString.includes(leetSymbol);
     });
-    return Array.from(unleetRecurse(cleanLeetString.trim(), new Set(), new Set(), complexTranslationKeys));
+    const cleanLeetStringSplit = cleanLeetString.split(" ");
+    const unleetResults = [];
+    for (const cleanLeetStringPiece of cleanLeetStringSplit) {
+        unleetResults.push(Array.from(unleetRecurse(cleanLeetStringPiece.trim(), new Set(), new Set(), complexTranslationKeys)));
+    }
+    if (unleetResults.length === 1) {
+        return unleetResults[0];
+    }
+    return utils.combineStringArrays(unleetResults);
 };
+export default unleet;
diff --git a/index.ts b/index.ts
@@ -1,46 +1,6 @@
 import diacritic from "diacritic";
-
-import { leetSymbolTranslationKeys, simpleTranslations, complexTranslations } from "./translations/translations.js";
-
-
-const indiciesOf = (sourceString: string, searchString: string) => {
-
-  const indicies: number[] = [];
-
-  for (let index = 0; index < sourceString.length - searchString.length; index += 1) {
-
-    if (sourceString.substring(index, searchString.length + index) === searchString) {
-      indicies.push(index);
-    }
-  }
-
-  return indicies;
-};
-
-
-const isLetter = (potentialLetter: string) => {
-  if ("abcdefghijklmnopqrstuvwxyz".includes(potentialLetter)) {
-    return true;
-  }
-  return false;
-};
-
-
-const isPotentialLeet = (potentialLeetString: string) => {
-
-  for (const leetSymbol of leetSymbolTranslationKeys) {
-
-    if (isLetter(leetSymbol)) {
-      continue;
-    }
-
-    if (potentialLeetString.includes(leetSymbol)) {
-      return true;
-    }
-  }
-
-  return false;
-};
+import * as utils from "./utils.js";
+import { simpleTranslations, complexTranslations } from "./translations/translations.js";
 
 
 const unleetRecurse = (lowerCaseLeetString: string,
@@ -54,7 +14,7 @@ const unleetRecurse = (lowerCaseLeetString: string,
     // If the current leet symbol is found in the string
     if (lowerCaseLeetString.includes(leetSymbol)) {
 
-      let matchingIndicies = indiciesOf(lowerCaseLeetString, leetSymbol);
+      let matchingIndicies = utils.indiciesOf(lowerCaseLeetString, leetSymbol);
 
       if (matchingIndicies.length === 0) {
         matchingIndicies = [lowerCaseLeetString.indexOf(leetSymbol)];
@@ -81,7 +41,7 @@ const unleetRecurse = (lowerCaseLeetString: string,
   }
 
   // If no leet symbols exist in the string that are able to be translated
-  if (!isPotentialLeet(lowerCaseLeetString)) {
+  if (!utils.isPotentialLeet(lowerCaseLeetString)) {
     unleetStrings.add(lowerCaseLeetString);
     return unleetStrings;
   }
@@ -90,14 +50,14 @@ const unleetRecurse = (lowerCaseLeetString: string,
 };
 
 
-export const unleet = (leetString: string): string[] => {
+export const unleet = (leetString: string | number): string[] => {
 
   if (leetString === null || leetString === undefined || leetString === "") {
     return [""];
   }
 
   // Convert to lower case
-  let cleanLeetString = (leetString + "").toLowerCase();
+  let cleanLeetString = leetString.toString().toLowerCase();
 
   // Remove periods
   cleanLeetString = cleanLeetString.replace(/\./g, " ");
@@ -120,5 +80,20 @@ export const unleet = (leetString: string): string[] => {
     return cleanLeetString.includes(leetSymbol);
   });
 
-  return Array.from(unleetRecurse(cleanLeetString.trim(), new Set(), new Set(), complexTranslationKeys));
+  const cleanLeetStringSplit = cleanLeetString.split(" ");
+
+  const unleetResults: string[][] = [];
+
+  for (const cleanLeetStringPiece of cleanLeetStringSplit) {
+    unleetResults.push(Array.from(unleetRecurse(cleanLeetStringPiece.trim(), new Set(), new Set(), complexTranslationKeys)));
+  }
+
+  if (unleetResults.length === 1) {
+    return unleetResults[0];
+  }
+
+  return utils.combineStringArrays(unleetResults);
 };
+
+
+export default unleet;