nlptoolkit-morphologicalanalysis 1.0.14 → 1.0.15

package/README.md

@@ -32,7 +32,7 @@ For Developers
 ============
 
 You can also see [Python](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis-Py), 
-[Java](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis), [C++](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis-CPP), 
+[Java](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis), [C++](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis-CPP), [C](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis-C),
 [Swift](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis-Swift), [Cython](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis-Cy),
 or [C#](https://github.com/starlangsoftware/TurkishMorphologicalAnalysis-CS) repository.
 

package/dist/Corpus/DisambiguationCorpus.d.ts

@@ -1,4 +1,11 @@
 import { Corpus } from "nlptoolkit-corpus/dist/Corpus";
 export declare class DisambiguationCorpus extends Corpus {
+    /**
+     * Constructor which takes a file name {@link String} as an input and reads the file line by line. It takes each word of the line,
+     * and creates a new {@link DisambiguatedWord} with current word and its {@link MorphologicalParse}. It also creates a new {@link Sentence}
+     * when a new sentence starts, and adds each word to this sentence till the end of that sentence.
+     *
+     * @param fileName File which will be read and parsed.
+     */
     constructor(fileName?: string);
 }

package/dist/Corpus/DisambiguationCorpus.js

@@ -16,6 +16,13 @@
     const MorphologicalParse_1 = require("../MorphologicalAnalysis/MorphologicalParse");
     const Sentence_1 = require("nlptoolkit-corpus/dist/Sentence");
     class DisambiguationCorpus extends Corpus_1.Corpus {
+        /**
+         * Constructor which takes a file name {@link String} as an input and reads the file line by line. It takes each word of the line,
+         * and creates a new {@link DisambiguatedWord} with current word and its {@link MorphologicalParse}. It also creates a new {@link Sentence}
+         * when a new sentence starts, and adds each word to this sentence till the end of that sentence.
+         *
+         * @param fileName File which will be read and parsed.
+         */
         constructor(fileName) {
             super();
             if (fileName != undefined) {

package/dist/Corpus/DisambiguationCorpus.js.map

@@ -1 +1 @@
-{"version":3,"file":"DisambiguationCorpus.js","sourceRoot":"","sources":["../../source/Corpus/DisambiguationCorpus.ts"],"names":[],"mappings":";;;;;;;;;;;;IAAA,0DAAqD;IACrD,yBAAyB;IACzB,2DAAsD;IACtD,oFAA+E;IAC/E,8DAAyD;IAEzD,MAAa,oBAAqB,SAAQ,eAAM;QAE5C,YAAY,QAAiB;YACzB,KAAK,EAAE,CAAC;YACR,IAAI,QAAQ,IAAI,SAAS,EAAC;gBACtB,IAAI,WAAW,GAAG,SAAS,CAAC;gBAC5B,IAAI,IAAI,GAAG,EAAE,CAAC,YAAY,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAA;gBAC5C,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;gBAC5B,KAAK,IAAI,IAAI,IAAI,KAAK,EAAE;oBACpB,IAAI,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;oBACjD,IAAI,KAAK,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;oBACnD,IAAI,IAAI,IAAI,EAAE,IAAI,KAAK,IAAI,EAAE,EAAE;wBAC3B,IAAI,OAAO,GAAG,IAAI,qCAAiB,CAAC,IAAI,EAAE,IAAI,uCAAkB,CAAC,KAAK,CAAC,CAAC,CAAC;wBACzE,IAAI,IAAI,IAAI,KAAK,EAAE;4BACf,WAAW,GAAG,IAAI,mBAAQ,EAAE,CAAC;yBAChC;6BAAM;4BACH,IAAI,IAAI,IAAI,MAAM,EAAE;gCAChB,IAAI,CAAC,WAAW,CAAC,WAAW,CAAC,CAAC;6BACjC;iCAAM;gCACH,IAAI,IAAI,IAAI,OAAO,IAAI,IAAI,IAAI,QAAQ,IAAI,IAAI,IAAI,SAAS,IAAI,IAAI,IAAI,UAAU,EAAE;iCACnF;qCAAM;oCACH,IAAI,WAAW,IAAI,IAAI,EAAE;wCACrB,WAAW,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;qCAChC;iCACJ;6BACJ;yBACJ;qBACJ;iBACJ;aACJ;QACL,CAAC;KAEJ;IAhCD,oDAgCC"}
+{"version":3,"file":"DisambiguationCorpus.js","sourceRoot":"","sources":["../../source/Corpus/DisambiguationCorpus.ts"],"names":[],"mappings":";;;;;;;;;;;;IAAA,0DAAqD;IACrD,yBAAyB;IACzB,2DAAsD;IACtD,oFAA+E;IAC/E,8DAAyD;IAEzD,MAAa,oBAAqB,SAAQ,eAAM;QAE5C;;;;;;WAMG;QACH,YAAY,QAAiB;YACzB,KAAK,EAAE,CAAC;YACR,IAAI,QAAQ,IAAI,SAAS,EAAC;gBACtB,IAAI,WAAW,GAAG,SAAS,CAAC;gBAC5B,IAAI,IAAI,GAAG,EAAE,CAAC,YAAY,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAA;gBAC5C,IAAI,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;gBAC5B,KAAK,IAAI,IAAI,IAAI,KAAK,EAAE;oBACpB,IAAI,IAAI,GAAG,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;oBACjD,IAAI,KAAK,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;oBACnD,IAAI,IAAI,IAAI,EAAE,IAAI,KAAK,IAAI,EAAE,EAAE;wBAC3B,IAAI,OAAO,GAAG,IAAI,qCAAiB,CAAC,IAAI,EAAE,IAAI,uCAAkB,CAAC,KAAK,CAAC,CAAC,CAAC;wBACzE,IAAI,IAAI,IAAI,KAAK,EAAE;4BACf,WAAW,GAAG,IAAI,mBAAQ,EAAE,CAAC;yBAChC;6BAAM;4BACH,IAAI,IAAI,IAAI,MAAM,EAAE;gCAChB,IAAI,CAAC,WAAW,CAAC,WAAW,CAAC,CAAC;6BACjC;iCAAM;gCACH,IAAI,IAAI,IAAI,OAAO,IAAI,IAAI,IAAI,QAAQ,IAAI,IAAI,IAAI,SAAS,IAAI,IAAI,IAAI,UAAU,EAAE;iCACnF;qCAAM;oCACH,IAAI,WAAW,IAAI,IAAI,EAAE;wCACrB,WAAW,CAAC,OAAO,CAAC,OAAO,CAAC,CAAC;qCAChC;iCACJ;6BACJ;yBACJ;qBACJ;iBACJ;aACJ;QACL,CAAC;KAEJ;IAvCD,oDAuCC"}

package/dist/MorphologicalAnalysis/FsmMorphologicalAnalyzer.d.ts

@@ -10,10 +10,11 @@ export declare class FsmMorphologicalAnalyzer {
     private dictionaryTrie;
     private suffixTrie;
     private parsedSurfaceForms;
-    private finiteStateMachine;
+    private pronunciations;
+    private readonly finiteStateMachine;
     private static MAX_DISTANCE;
-    private dictionary;
-    private cache;
+    private readonly dictionary;
+    private readonly cache;
     private mostUsedPatterns;
     /**
      * Another constructor of FsmMorphologicalAnalyzer class. It generates a new TxtDictionary type dictionary from
@@ -24,9 +25,29 @@ export declare class FsmMorphologicalAnalyzer {
      * @param cacheSize  the size of the LRUCache.
      */
     constructor(fileName?: string, dictionaryFileNameOrDictionary?: any, cacheSize?: number);
+    /**
+     * Constructs and returns the reverse string of a given string.
+     * @param s String to be reversed.
+     * @return Reverse of a given string.
+     */
     private reverseString;
+    /**
+     * Constructs the suffix trie from the input file suffixes.txt. suffixes.txt contains the most frequent 6000
+     * suffixes that a verb or a noun can take. The suffix trie is a trie that stores these suffixes in reverse form,
+     * which can be then used to match a given word for its possible suffix content.
+     */
     private prepareSuffixTrie;
+    /**
+     * Reads the file for correct surface forms and their most frequent root forms, in other words, the surface forms
+     * which have at least one morphological analysis in  Turkish.
+     * @param fileName Input file containing analyzable surface forms and their root forms.
+     */
     addParsedSurfaceForms(fileName: string): void;
+    /**
+     * Reads the file for foreign words and their pronunciations.
+     * @param fileName Input file containing foreign words and their pronunciations.
+     */
+    addPronunciations(fileName: string): void;
     /**
      * The getPossibleWords method takes {@link MorphologicalParse} and {@link MetamorphicParse} as input.
      * First it determines whether the given morphologicalParse is the root verb and whether it contains a verb tag.
@@ -319,6 +340,15 @@ export declare class FsmMorphologicalAnalyzer {
      * @return ArrayList type initialFsmParse which holds the analyses.
      */
     analysis(surfaceForm: string, isProper: boolean): Array<FsmParse>;
+    /**
+     * This method uses cache idea to speed up pattern matching in Fsm. mostUsedPatterns stores the compiled forms of
+     * the previously used patterns. When Fsm tries to match a string to a pattern, first we check if it exists in
+     * mostUsedPatterns. If it exists, we directly use the compiled pattern to match the string. Otherwise, new pattern
+     * is compiled and put in the mostUsedPatterns.
+     * @param expr Pattern to check
+     * @param value String to match the pattern
+     * @return True if the string matches the pattern, false otherwise.
+     */
     private patternMatches;
     /**
      * The isProperNoun method takes surfaceForm String as input and checks its each char whether they are in the range
@@ -335,6 +365,19 @@ export declare class FsmMorphologicalAnalyzer {
      * @return true if it is a code-like word, return false otherwise.
      */
     isCode(surfaceForm: string): boolean;
+    /**
+     * Identifies a possible new root word for a given surface form. It also adds the new root form to the dictionary
+     * for further usage. The method first searches the suffix trie for the reverse string of the surface form. This
+     * way, it can identify if the word has a suffix that is in the most frequently used suffix list. Since a word can
+     * have multiple possible suffixes, the method identifies the longest suffix and returns the substring of the
+     * surface form tht does not contain the suffix. Let say the word is 'googlelaştırdık', it will identify 'tık' as
+     * a suffix and will return 'googlelaştır' as a possible root form. Another example will be 'homelesslerimizle', it
+     * will identify 'lerimizle' as suffix and will return 'homeless' as a possible root form. If the root word ends
+     * with 'ğ', it is replacesd with 'k'. 'morfolojikliğini' will return 'morfolojikliğ' then which will be replaced
+     * with 'morfolojiklik'.
+     * @param surfaceForm Surface form for which we will identify a possible new root form.
+     * @return Possible new root form.
+     */
     private rootOfPossiblyNewWord;
     /**
      * The robustMorphologicalAnalysis is used to analyse surfaceForm String. First it gets the currentParse of the surfaceForm
@@ -383,9 +426,30 @@ export declare class FsmMorphologicalAnalyzer {
      * @return true if surfaceForm matches with the regex.
      */
     private isNumber;
+    /**
+     * Checks if a given surface form matches to a percent value. It should be something like %4, %45, %4.3 or %56.786
+     * @param surfaceForm Surface form to be checked.
+     * @return True if the surface form is in percent form
+     */
     private isPercent;
+    /**
+     * Checks if a given surface form matches to a time form. It should be something like 3:34, 12:56 etc.
+     * @param surfaceForm Surface form to be checked.
+     * @return True if the surface form is in time form
+     */
     private isTime;
+    /**
+     * Checks if a given surface form matches to a range form. It should be something like 123-1400 or 12:34-15:78 or
+     * 3.45-4.67.
+     * @param surfaceForm Surface form to be checked.
+     * @return True if the surface form is in range form
+     */
     private isRange;
+    /**
+     * Checks if a given surface form matches to a date form. It should be something like 3/10/2023 or 2.3.2012
+     * @param surfaceForm Surface form to be checked.
+     * @return True if the surface form is in date form
+     */
     private isDate;
     /**
      * The morphologicalAnalysis method is used to analyse a FsmParseList by comparing with the regex.

package/dist/MorphologicalAnalysis/FsmMorphologicalAnalyzer.js

@@ -4,7 +4,7 @@
         if (v !== undefined) module.exports = v;
     }
     else if (typeof define === "function" && define.amd) {
-        define(["require", "exports", "nlptoolkit-dictionary/dist/Dictionary/Trie/Trie", "./FiniteStateMachine", "nlptoolkit-dictionary/dist/Dictionary/TxtDictionary", "nlptoolkit-datastructure/dist/LRUCache", "./FsmParseList", "nlptoolkit-dictionary/dist/Dictionary/WordComparator", "fs", "./Transition", "./MorphologicalTag", "nlptoolkit-dictionary/dist/Dictionary/TxtWord", "./FsmParse", "nlptoolkit-corpus/dist/Sentence", "nlptoolkit-dictionary/dist/Dictionary/Word", "./State", "nlptoolkit-datastructure/dist/Queue"], factory);
+        define(["require", "exports", "nlptoolkit-dictionary/dist/Dictionary/Trie/Trie", "./FiniteStateMachine", "nlptoolkit-dictionary/dist/Dictionary/TxtDictionary", "nlptoolkit-datastructure/dist/LRUCache", "./FsmParseList", "nlptoolkit-dictionary/dist/Dictionary/WordComparator", "fs", "./Transition", "./MorphologicalTag", "nlptoolkit-dictionary/dist/Dictionary/TxtWord", "./FsmParse", "nlptoolkit-corpus/dist/Sentence", "nlptoolkit-dictionary/dist/Dictionary/Word", "./State", "nlptoolkit-datastructure/dist/Queue", "nlptoolkit-util/dist/FileUtils"], factory);
     }
 })(function (require, exports) {
     "use strict";
@@ -25,6 +25,7 @@
     const Word_1 = require("nlptoolkit-dictionary/dist/Dictionary/Word");
     const State_1 = require("./State");
     const Queue_1 = require("nlptoolkit-datastructure/dist/Queue");
+    const FileUtils_1 = require("nlptoolkit-util/dist/FileUtils");
     class FsmMorphologicalAnalyzer {
         /**
          * Another constructor of FsmMorphologicalAnalyzer class. It generates a new TxtDictionary type dictionary from
@@ -36,6 +37,7 @@
          */
         constructor(fileName, dictionaryFileNameOrDictionary, cacheSize) {
             this.parsedSurfaceForms = undefined;
+            this.pronunciations = undefined;
             this.cache = undefined;
             this.mostUsedPatterns = new Map();
             if (dictionaryFileNameOrDictionary == undefined) {
@@ -60,7 +62,13 @@
             if (cacheSize > 0) {
                 this.cache = new LRUCache_1.LRUCache(cacheSize);
             }
+            this.addPronunciations("pronunciations.txt");
         }
+        /**
+         * Constructs and returns the reverse string of a given string.
+         * @param s String to be reversed.
+         * @return Reverse of a given string.
+         */
         reverseString(s) {
             let result = "";
             for (let i = s.length - 1; i >= 0; i--) {
@@ -68,6 +76,11 @@
             }
             return result;
         }
+        /**
+         * Constructs the suffix trie from the input file suffixes.txt. suffixes.txt contains the most frequent 6000
+         * suffixes that a verb or a noun can take. The suffix trie is a trie that stores these suffixes in reverse form,
+         * which can be then used to match a given word for its possible suffix content.
+         */
         prepareSuffixTrie() {
             this.suffixTrie = new Trie_1.Trie();
             let data = fs.readFileSync("suffixes.txt", 'utf8');
@@ -77,14 +90,20 @@
                 this.suffixTrie.addWord(reverseSuffix, new Word_1.Word(reverseSuffix));
             }
         }
+        /**
+         * Reads the file for correct surface forms and their most frequent root forms, in other words, the surface forms
+         * which have at least one morphological analysis in  Turkish.
+         * @param fileName Input file containing analyzable surface forms and their root forms.
+         */
         addParsedSurfaceForms(fileName) {
-            this.parsedSurfaceForms = new Map();
-            let data = fs.readFileSync(fileName, 'utf8');
-            let lines = data.split("\n");
-            for (let line of lines) {
-                let items = line.split(" ");
-                this.parsedSurfaceForms.set(items[0], items[1]);
-            }
+            this.parsedSurfaceForms = FileUtils_1.FileUtils.readHashMap(fileName);
+        }
+        /**
+         * Reads the file for foreign words and their pronunciations.
+         * @param fileName Input file containing foreign words and their pronunciations.
+         */
+        addPronunciations(fileName) {
+            this.pronunciations = FileUtils_1.FileUtils.readHashMap(fileName);
         }
         /**
          * The getPossibleWords method takes {@link MorphologicalParse} and {@link MetamorphicParse} as input.
@@ -677,7 +696,7 @@
          */
         parseWordLength(fsmParse, maxLength) {
             let result = new Array();
-            let resultSuffixList = new Array();
+            let resultTransitionList = new Array();
             let parseQueue = new Queue_1.Queue(1000);
             parseQueue.enqueueAll(fsmParse);
             while (!parseQueue.isEmpty()) {
@@ -687,11 +706,11 @@
                 let currentState = currentFsmParse.getFinalSuffix();
                 let currentSurfaceForm = currentFsmParse.getSurfaceForm();
                 if (currentState.isEndState() && currentSurfaceForm.length <= maxLength) {
-                    let currentSuffixList = currentFsmParse.getSuffixList();
-                    if (!resultSuffixList.includes(currentSuffixList)) {
+                    let currentTransitionList = currentSurfaceForm + " " + currentFsmParse.getFsmParseTransitionList();
+                    if (!resultTransitionList.includes(currentTransitionList)) {
                         result.push(currentFsmParse);
                         currentFsmParse.constructInflectionalGroups();
-                        resultSuffixList.push(currentSuffixList);
+                        resultTransitionList.push(currentTransitionList);
                     }
                 }
                 this.addNewParsesFromCurrentParseLength(currentFsmParse, parseQueue, maxLength, root);
@@ -708,7 +727,7 @@
          */
         parseWordSurfaceForm(fsmParse, surfaceForm) {
             let result = new Array();
-            let resultSuffixList = new Array();
+            let resultTransitionList = new Array();
             let parseQueue = new Queue_1.Queue(1000);
             parseQueue.enqueueAll(fsmParse);
             while (!parseQueue.isEmpty()) {
@@ -718,11 +737,11 @@
                 let currentState = currentFsmParse.getFinalSuffix();
                 let currentSurfaceForm = currentFsmParse.getSurfaceForm();
                 if (currentState.isEndState() && currentSurfaceForm == surfaceForm) {
-                    let currentSuffixList = currentFsmParse.getSuffixList();
-                    if (!resultSuffixList.includes(currentSuffixList)) {
+                    let currentTransitionList = currentFsmParse.getFsmParseTransitionList();
+                    if (!resultTransitionList.includes(currentTransitionList)) {
                         result.push(currentFsmParse);
                         currentFsmParse.constructInflectionalGroups();
-                        resultSuffixList.push(currentSuffixList);
+                        resultTransitionList.push(currentTransitionList);
                     }
                 }
                 this.addNewParsesFromCurrentParseSurfaceForm(currentFsmParse, parseQueue, surfaceForm, root);
@@ -986,6 +1005,15 @@
             initialFsmParse = this.initializeParseListFromSurfaceForm(surfaceForm, isProper);
             return this.parseWordSurfaceForm(initialFsmParse, surfaceForm);
         }
+        /**
+         * This method uses cache idea to speed up pattern matching in Fsm. mostUsedPatterns stores the compiled forms of
+         * the previously used patterns. When Fsm tries to match a string to a pattern, first we check if it exists in
+         * mostUsedPatterns. If it exists, we directly use the compiled pattern to match the string. Otherwise, new pattern
+         * is compiled and put in the mostUsedPatterns.
+         * @param expr Pattern to check
+         * @param value String to match the pattern
+         * @return True if the string matches the pattern, false otherwise.
+         */
         patternMatches(expr, value) {
             let p = this.mostUsedPatterns.get(expr);
             if (p == undefined) {
@@ -1021,31 +1049,39 @@
             }
             return this.patternMatches("^.*[0-9].*$", surfaceForm) && this.patternMatches("^.*[a-zA-ZçöğüşıÇÖĞÜŞİ].*$", surfaceForm);
         }
+        /**
+         * Identifies a possible new root word for a given surface form. It also adds the new root form to the dictionary
+         * for further usage. The method first searches the suffix trie for the reverse string of the surface form. This
+         * way, it can identify if the word has a suffix that is in the most frequently used suffix list. Since a word can
+         * have multiple possible suffixes, the method identifies the longest suffix and returns the substring of the
+         * surface form tht does not contain the suffix. Let say the word is 'googlelaştırdık', it will identify 'tık' as
+         * a suffix and will return 'googlelaştır' as a possible root form. Another example will be 'homelesslerimizle', it
+         * will identify 'lerimizle' as suffix and will return 'homeless' as a possible root form. If the root word ends
+         * with 'ğ', it is replacesd with 'k'. 'morfolojikliğini' will return 'morfolojikliğ' then which will be replaced
+         * with 'morfolojiklik'.
+         * @param surfaceForm Surface form for which we will identify a possible new root form.
+         * @return Possible new root form.
+         */
         rootOfPossiblyNewWord(surfaceForm) {
             let words = this.suffixTrie.getWordsWithPrefix(this.reverseString(surfaceForm));
-            let maxLength = 0;
-            let longestWord = null;
+            let candidateWord = null;
+            let candidateList = new Array();
             for (let word of words) {
-                if (word.getName().length > maxLength) {
-                    longestWord = surfaceForm.substring(0, surfaceForm.length - word.getName().length);
-                    maxLength = word.getName().length;
-                }
-            }
-            if (maxLength != 0) {
+                candidateWord = surfaceForm.substring(0, surfaceForm.length - word.getName().length);
                 let newWord;
-                if (longestWord.endsWith("ğ")) {
-                    longestWord = longestWord.substring(0, longestWord.length - 1) + "k";
-                    newWord = new TxtWord_1.TxtWord(longestWord, "CL_ISIM");
+                if (candidateWord.endsWith("ğ")) {
+                    candidateWord = candidateWord.substring(0, candidateWord.length - 1) + "k";
+                    newWord = new TxtWord_1.TxtWord(candidateWord, "CL_ISIM");
                     newWord.addFlag("IS_SD");
                 }
                 else {
-                    newWord = new TxtWord_1.TxtWord(longestWord, "CL_ISIM");
+                    newWord = new TxtWord_1.TxtWord(candidateWord, "CL_ISIM");
                     newWord.addFlag("CL_FIIL");
                 }
-                this.dictionaryTrie.addWord(longestWord, newWord);
-                return newWord;
+                candidateList.push(newWord);
+                this.dictionaryTrie.addWord(candidateWord, newWord);
             }
-            return null;
+            return candidateList;
         }
         /**
          * The robustMorphologicalAnalysis is used to analyse surfaceForm String. First it gets the currentParse of the surfaceForm
@@ -1066,21 +1102,17 @@
                 if (this.isProperNoun(surfaceForm)) {
                     fsmParse.push(new FsmParse_1.FsmParse(surfaceForm, this.finiteStateMachine.getState("ProperRoot")));
                 }
-                else {
-                    if (this.isCode(surfaceForm)) {
-                        fsmParse.push(new FsmParse_1.FsmParse(surfaceForm, this.finiteStateMachine.getState("CodeRoot")));
-                    }
-                    else {
-                        let newRoot = this.rootOfPossiblyNewWord(surfaceForm);
-                        if (newRoot != null) {
-                            fsmParse.push(new FsmParse_1.FsmParse(newRoot, this.finiteStateMachine.getState("VerbalRoot")));
-                            fsmParse.push(new FsmParse_1.FsmParse(newRoot, this.finiteStateMachine.getState("NominalRoot")));
-                        }
-                        else {
-                            fsmParse.push(new FsmParse_1.FsmParse(surfaceForm, this.finiteStateMachine.getState("NominalRoot")));
-                        }
+                if (this.isCode(surfaceForm)) {
+                    fsmParse.push(new FsmParse_1.FsmParse(surfaceForm, this.finiteStateMachine.getState("CodeRoot")));
+                }
+                let newCandidateList = this.rootOfPossiblyNewWord(surfaceForm);
+                if (newCandidateList.length != 0) {
+                    for (let word of newCandidateList) {
+                        fsmParse.push(new FsmParse_1.FsmParse(word, this.finiteStateMachine.getState("VerbalRoot")));
+                        fsmParse.push(new FsmParse_1.FsmParse(word, this.finiteStateMachine.getState("NominalRoot")));
                     }
                 }
+                fsmParse.push(new FsmParse_1.FsmParse(surfaceForm, this.finiteStateMachine.getState("NominalRoot")));
                 return new FsmParseList_1.FsmParseList(this.parseWordSurfaceForm(fsmParse, surfaceForm));
             }
             else {
@@ -1186,19 +1218,40 @@
             }
             return word == "" && count > 1;
         }
+        /**
+         * Checks if a given surface form matches to a percent value. It should be something like %4, %45, %4.3 or %56.786
+         * @param surfaceForm Surface form to be checked.
+         * @return True if the surface form is in percent form
+         */
         isPercent(surfaceForm) {
             return this.patternMatches("^%(\\d\\d|\\d)$", surfaceForm) ||
                 this.patternMatches("^%(\\d\\d|\\d)\\.\\d+$", surfaceForm);
         }
+        /**
+         * Checks if a given surface form matches to a time form. It should be something like 3:34, 12:56 etc.
+         * @param surfaceForm Surface form to be checked.
+         * @return True if the surface form is in time form
+         */
         isTime(surfaceForm) {
             return this.patternMatches("^(\\d\\d|\\d):(\\d\\d|\\d):(\\d\\d|\\d)$", surfaceForm) ||
                 this.patternMatches("^(\\d\\d|\\d):(\\d\\d|\\d)$", surfaceForm);
         }
+        /**
+         * Checks if a given surface form matches to a range form. It should be something like 123-1400 or 12:34-15:78 or
+         * 3.45-4.67.
+         * @param surfaceForm Surface form to be checked.
+         * @return True if the surface form is in range form
+         */
         isRange(surfaceForm) {
             return this.patternMatches("^\\d+-\\d+$", surfaceForm) ||
                 this.patternMatches("^(\\d\\d|\\d):(\\d\\d|\\d)-(\\d\\d|\\d):(\\d\\d|\\d)$", surfaceForm) ||
                 this.patternMatches("^(\\d\\d|\\d)\\.(\\d\\d|\\d)-(\\d\\d|\\d)\\.(\\d\\d|\\d)$", surfaceForm);
         }
+        /**
+         * Checks if a given surface form matches to a date form. It should be something like 3/10/2023 or 2.3.2012
+         * @param surfaceForm Surface form to be checked.
+         * @return True if the surface form is in date form
+         */
         isDate(surfaceForm) {
             return this.patternMatches("^(\\d\\d|\\d)/(\\d\\d|\\d)/\\d+$", surfaceForm) ||
                 this.patternMatches("^(\\d\\d|\\d)\\.(\\d\\d|\\d)\\.\\d+$", surfaceForm);
@@ -1222,6 +1275,8 @@
          */
         morphologicalAnalysis(surfaceForm) {
             let lowerCased = surfaceForm.toLocaleLowerCase("tr");
+            let possibleRootLowerCased = "", pronunciation = "";
+            let isRootReplaced = false;
             if (this.parsedSurfaceForms != undefined && this.parsedSurfaceForms.has(lowerCased) &&
                 !this.isInteger(surfaceForm) && !this.isDouble(surfaceForm) && !this.isPercent(surfaceForm) &&
                 !this.isTime(surfaceForm) && !this.isRange(surfaceForm) && !this.isDate(surfaceForm)) {
@@ -1233,9 +1288,9 @@
                 return this.cache.get(surfaceForm);
             }
             if (this.patternMatches("^(\\w|Ç|Ş|İ|Ü|Ö)\\.$", surfaceForm)) {
-                this.dictionaryTrie.addWord(surfaceForm.toLocaleLowerCase("tr"), new TxtWord_1.TxtWord(surfaceForm.toLocaleLowerCase("tr"), "IS_OA"));
+                this.dictionaryTrie.addWord(lowerCased, new TxtWord_1.TxtWord(lowerCased, "IS_OA"));
             }
-            let defaultFsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+            let defaultFsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
             if (defaultFsmParse.length > 0) {
                 let fsmParseList = new FsmParseList_1.FsmParseList(defaultFsmParse);
                 if (this.cache != undefined) {
@@ -1249,57 +1304,69 @@
                 if (possibleRoot != "") {
                     if (possibleRoot.includes("/") || possibleRoot.includes("\\/")) {
                         this.dictionaryTrie.addWord(possibleRoot, new TxtWord_1.TxtWord(possibleRoot, "IS_KESIR"));
-                        fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                        fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                     }
                     else {
                         if (this.isDate(possibleRoot)) {
                             this.dictionaryTrie.addWord(possibleRoot, new TxtWord_1.TxtWord(possibleRoot, "IS_DATE"));
-                            fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                            fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                         }
                         else {
                             if (this.patternMatches("^\\d+/\\d+$", possibleRoot)) {
                                 this.dictionaryTrie.addWord(possibleRoot, new TxtWord_1.TxtWord(possibleRoot, "IS_KESIR"));
-                                fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                                fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                             }
                             else {
                                 if (this.isPercent(possibleRoot)) {
                                     this.dictionaryTrie.addWord(possibleRoot, new TxtWord_1.TxtWord(possibleRoot, "IS_PERCENT"));
-                                    fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                                    fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                                 }
                                 else {
                                     if (this.isTime(surfaceForm)) {
                                         this.dictionaryTrie.addWord(possibleRoot, new TxtWord_1.TxtWord(possibleRoot, "IS_ZAMAN"));
-                                        fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                                        fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                                     }
                                     else {
                                         if (this.isRange(surfaceForm)) {
                                             this.dictionaryTrie.addWord(possibleRoot, new TxtWord_1.TxtWord(possibleRoot, "IS_RANGE"));
-                                            fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                                            fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                                         }
                                         else {
                                             if (this.isInteger(possibleRoot)) {
                                                 this.dictionaryTrie.addWord(possibleRoot, new TxtWord_1.TxtWord(possibleRoot, "IS_SAYI"));
-                                                fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                                                fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                                             }
                                             else {
                                                 if (this.isDouble(possibleRoot)) {
                                                     this.dictionaryTrie.addWord(possibleRoot, new TxtWord_1.TxtWord(possibleRoot, "IS_REELSAYI"));
-                                                    fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                                                    fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                                                 }
                                                 else {
-                                                    if (Word_1.Word.isCapital(possibleRoot)) {
+                                                    if (Word_1.Word.isCapital(possibleRoot) || "QXW".includes(possibleRoot.substring(0, 1))) {
                                                         let newWord = undefined;
-                                                        if (this.dictionary.getWord(possibleRoot.toLocaleLowerCase("tr")) != null) {
-                                                            this.dictionary.getWord(possibleRoot.toLocaleLowerCase("tr")).addFlag("IS_OA");
+                                                        possibleRootLowerCased = possibleRoot.toLocaleLowerCase("tr");
+                                                        if (this.pronunciations.has(possibleRootLowerCased)) {
+                                                            isRootReplaced = true;
+                                                            pronunciation = this.pronunciations.get(possibleRootLowerCased);
+                                                            if (this.dictionary.getWord(pronunciation) != null) {
+                                                                this.dictionary.getWord(pronunciation).addFlag("IS_OA");
+                                                            }
+                                                            else {
+                                                                newWord = new TxtWord_1.TxtWord(pronunciation, "IS_OA");
+                                                                this.dictionaryTrie.addWord(pronunciation, newWord);
+                                                            }
+                                                            let replacedWord = pronunciation + lowerCased.substring(possibleRootLowerCased.length);
+                                                            fsmParse = this.analysis(replacedWord, this.isProperNoun(surfaceForm));
                                                         }
                                                         else {
-                                                            newWord = new TxtWord_1.TxtWord(possibleRoot.toLocaleLowerCase("tr"), "IS_OA");
-                                                            this.dictionaryTrie.addWord(possibleRoot.toLocaleLowerCase("tr"), newWord);
-                                                        }
-                                                        fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
-                                                        if (fsmParse.length == 0 && newWord != undefined) {
-                                                            newWord.addFlag("IS_KIS");
-                                                            fsmParse = this.analysis(surfaceForm.toLocaleLowerCase("tr"), this.isProperNoun(surfaceForm));
+                                                            if (this.dictionary.getWord(possibleRootLowerCased) != null) {
+                                                                this.dictionary.getWord(possibleRootLowerCased).addFlag("IS_OA");
+                                                            }
+                                                            else {
+                                                                newWord = new TxtWord_1.TxtWord(possibleRootLowerCased, "IS_OA");
+                                                                this.dictionaryTrie.addWord(possibleRootLowerCased, newWord);
+                                                            }
+                                                            fsmParse = this.analysis(lowerCased, this.isProperNoun(surfaceForm));
                                                         }
                                                     }
                                                 }
@@ -1312,6 +1379,11 @@
                     }
                 }
             }
+            if (!isRootReplaced) {
+                for (let parse of fsmParse) {
+                    parse.restoreOriginalForm(possibleRootLowerCased, pronunciation);
+                }
+            }
             let fsmParseList = new FsmParseList_1.FsmParseList(fsmParse);
             if (this.cache != undefined && fsmParseList.size() > 0) {
                 this.cache.add(surfaceForm, fsmParseList);