axyseo 2.1.12 → 2.1.13

package/build/cjs/languageProcessing/helpers/sentence/SentenceTokenizer.js

@@ -0,0 +1,565 @@
+"use strict";
+
+Object.defineProperty(exports, "__esModule", {
+  value: true
+});
+exports.default = void 0;
+var _lodash = require("lodash");
+var _core = _interopRequireDefault(require("tokenizer2/core"));
+var _quotes = require("../sanitize/quotes.js");
+var _abbreviations = _interopRequireDefault(require("../../languages/en/config/abbreviations"));
+var _createRegexFromArray = _interopRequireDefault(require("../regex/createRegexFromArray"));
+var _wordBoundaries = _interopRequireDefault(require("../../../config/wordBoundaries"));
+function _interopRequireDefault(e) { return e && e.__esModule ? e : { default: e }; }
+// All characters that indicate a sentence delimiter.
+const fullStop = ".";
+const fullStopRegex = new RegExp("^[" + fullStop + "]$");
+const smallerThanContentRegex = /^<[^><]*$/;
+const htmlStartRegex = /^<([^>\s/]+)[^>]*>$/mi;
+const htmlEndRegex = /^<\/([^>\s]+)[^>]*>$/mi;
+const blockStartRegex = /^\s*[[({]\s*$/;
+const blockEndRegex = /^\s*[\])}]\s*$/;
+const abbreviationsPreparedForRegex = _abbreviations.default.map(abbreviation => abbreviation.replace(".", "\\."));
+const abbreviationsRegex = (0, _createRegexFromArray.default)(abbreviationsPreparedForRegex);
+const wordBoundariesForRegex = "(^|$|[" + (0, _wordBoundaries.default)().map(boundary => "\\" + boundary).join("") + "])";
+const lastCharacterPartOfInitialsRegex = new RegExp(wordBoundariesForRegex + "[A-Za-z]$");
+
+// Constants to be used in isValidTagPair.
+// A regex to get the tag type.
+const tagTypeRegex = /<\/?([^\s]+?)(\s|>)/;
+// Semantic tags (as opposed to style tags) are tags that are used to structure the text.
+const semanticTags = ["p", "div", "h1", "h2", "h3", "h4", "h5", "h6", "span", "li", "main"];
+
+/**
+ * Class for tokenizing a (html) text into sentences.
+ */
+class SentenceTokenizer {
+  /**
+   * Constructor
+   * @constructor
+   */
+  constructor() {
+    /*
+           * \u2026 - ellipsis.
+           * \u06D4 - Urdu full stop.
+           * \u061f - Arabic question mark.
+          */
+    this.sentenceDelimiters = "”〞〟„』›»’‛`\"?!\u2026\u06d4\u061f";
+  }
+
+  /**
+   * Gets the sentence delimiters.
+   *
+   * @returns {string} The sentence delimiters.
+   */
+  getSentenceDelimiters() {
+    return this.sentenceDelimiters;
+  }
+
+  /**
+   * Returns whether or not a certain character is a number.
+   *
+   * @param {string} character The character to check.
+   * @returns {boolean} Whether or not the character is a capital letter.
+   */
+  isNumber(character) {
+    return !(0, _lodash.isNaN)(parseInt(character, 10));
+  }
+
+  /**
+   * Returns whether or not a given HTML tag is a break tag.
+   *
+   * @param {string} htmlTag The HTML tag to check.
+   * @returns {boolean} Whether or not the given HTML tag is a break tag.
+   */
+  isBreakTag(htmlTag) {
+    return /<\/?br/.test(htmlTag);
+  }
+
+  /**
+   * Returns whether or not a given character is quotation mark.
+   *
+   * @param {string} character The character to check.
+   * @returns {boolean} Whether or not the given character is a quotation mark.
+   */
+  isQuotation(character) {
+    character = (0, _quotes.normalize)(character);
+    return "'" === character || "\"" === character;
+  }
+
+  /**
+   * A mock definition of this function. This function is only used in extensions for languages that use an ordinal dot.
+   *
+   * @returns {boolean} Always returns false as it is a language specific implementation if a language has an ordinal dot.
+   */
+  endsWithOrdinalDot() {
+    return false;
+  }
+
+  /**
+   * Returns whether or not a given character is a punctuation mark that can be at the beginning
+   * of a sentence, like ¿ and ¡ used in Spanish.
+   *
+   * @param {string} character The character to check.
+   * @returns {boolean} Whether or not the given character is a punctuation mark.
+   */
+  isPunctuation(character) {
+    return "¿" === character || "¡" === character;
+  }
+
+  /**
+   * Removes duplicate whitespace from a given text.
+   *
+   * @param {string} text The text with duplicate whitespace.
+   * @returns {string} The text without duplicate whitespace.
+   */
+  removeDuplicateWhitespace(text) {
+    return text.replace(/\s+/, " ");
+  }
+
+  /**
+   * Returns whether or not a certain character is a capital letter.
+   *
+   * @param {string} character The character to check.
+   * @returns {boolean} Whether or not the character is a capital letter.
+   */
+  isCapitalLetter(character) {
+    return character !== character.toLocaleLowerCase();
+  }
+
+  /**
+   * Checks whether the given character is a smaller than sign.
+   *
+   * This function is used to make sure that tokenizing the content after
+   * the smaller than sign works as expected.
+   * E.g. 'A sentence. < Hello world!' = ['A sentence.', '< Hello world!'].
+   *
+   * @param {string} character The character to check.
+   * @returns {boolean} Whether the character is a smaller than sign ('<') or not.
+   */
+  isSmallerThanSign(character) {
+    return character === "<";
+  }
+
+  /**
+   * Retrieves the next two characters from an array with the two next tokens.
+   *
+   * @param {Array} nextTokens The two next tokens. Might be undefined.
+   * @returns {string} The next two characters.
+   */
+  getNextTwoCharacters(nextTokens) {
+    let next = "";
+    if (!(0, _lodash.isUndefined)(nextTokens[0])) {
+      next += nextTokens[0].src;
+    }
+    if (!(0, _lodash.isUndefined)(nextTokens[1])) {
+      next += nextTokens[1].src;
+    }
+    next = this.removeDuplicateWhitespace(next);
+    return next;
+  }
+
+  /**
+   * Checks whether a character is from a language that's written from right to left.
+   * These languages don't have capital letter forms. Therefore any letter from these languages is a
+   * potential sentence beginning.
+   *
+   * @param {string} letter The letter to check.
+   *
+   * @returns {boolean} Whether the letter is from an LTR language.
+   */
+  isLetterFromSpecificLanguage(letter) {
+    const ltrLetterRanges = [
+    // Hebrew characters.
+    /^[\u0590-\u05fe]+$/i,
+    // Arabic characters (used for Arabic, Farsi, Urdu).
+    /^[\u0600-\u06FF]+$/i,
+    // Additional Farsi characters.
+    /^[\uFB8A\u067E\u0686\u06AF]+$/i];
+    return ltrLetterRanges.some(ltrLetterRange => ltrLetterRange.test(letter));
+  }
+
+  /**
+   * Checks if the sentenceBeginning beginning is a valid beginning.
+   *
+   * @param {string} sentenceBeginning The beginning of the sentence to validate.
+   * @returns {boolean} Returns true if it is a valid beginning, false if it is not.
+   */
+  isValidSentenceBeginning(sentenceBeginning) {
+    return this.isCapitalLetter(sentenceBeginning) || this.isLetterFromSpecificLanguage(sentenceBeginning) || this.isNumber(sentenceBeginning) || this.isQuotation(sentenceBeginning) || this.isPunctuation(sentenceBeginning) || this.isSmallerThanSign(sentenceBeginning);
+  }
+
+  /**
+   * Checks if the token is a valid sentence start.
+   *
+   * @param {Object} token The token to validate.
+   * @returns {boolean} Returns true if the token is valid sentence start, false if it is not.
+   */
+  isSentenceStart(token) {
+    return !(0, _lodash.isUndefined)(token) && ("html-start" === token.type || "html-end" === token.type || "block-start" === token.type);
+  }
+
+  /**
+   * Checks if the token is a valid sentence ending. A valid sentence ending is either a full stop or another
+   * delimiter such as "?", "!", etc.
+   *
+   * @param {Object} token The token to validate.
+   * @returns {boolean} Returns true if the token is valid sentence ending, false if it is not.
+   */
+  isSentenceEnding(token) {
+    return !(0, _lodash.isUndefined)(token) && (token.type === "full-stop" || token.type === "sentence-delimiter");
+  }
+
+  /**
+   * Checks if a full stop is part of a person's initials.
+   *
+   * Tests if tokens exist. Then tests if the tokens are of the right type.
+   * For previous token, it checks if the sentence ends with a single letter.
+   * For nextToken it checks if it is a single letter.
+   * Checks if next token is followed by a full stop.
+   *
+   * @param {object} token The current token (must be a full stop).
+   * @param {object} previousToken The token before the full stop.
+   * @param {object} nextToken The token following the full stop.
+   * @param {object} secondToNextToken The second token after the full stop.
+   * @returns {boolean} True if a full stop is part of a person's initials, False if the full stop is not part of a person's initials.
+   */
+  isPartOfPersonInitial(token, previousToken, nextToken, secondToNextToken) {
+    return !(0, _lodash.isUndefined)(token) && !(0, _lodash.isUndefined)(nextToken) && !(0, _lodash.isUndefined)(secondToNextToken) && !(0, _lodash.isUndefined)(previousToken) && token.type === "full-stop" && previousToken.type === "sentence" && lastCharacterPartOfInitialsRegex.test(previousToken.src) && nextToken.type === "sentence" && nextToken.src.trim().length === 1 && secondToNextToken.type === "full-stop";
+  }
+
+  /**
+   * Tokens that represent a '<', followed by content until it enters another '<' or '>'
+   * gets another pass by the tokenizer.
+   *
+   * @param {Object} token A token of type 'smaller-than-sign-content'.
+   * @param {string[]} tokenSentences The current array of found sentences. Sentences may get added by this method.
+   * @param {string} currentSentence The current sentence. Sentence parts may get appended by this method.
+   * @returns {{tokenSentences, currentSentence}} The found sentences and the current sentence, appended when necessary.
+   */
+  tokenizeSmallerThanContent(token, tokenSentences, currentSentence) {
+    /*
+    	Remove the '<' from the text, to avoid matching this rule
+    	recursively again and again.
+    	We add it again later on.
+    */
+    const localText = token.src.substring(1);
+
+    // Tokenize the current smaller-than-content token without the first '<'.
+    const tokenizerResult = this.createTokenizer();
+    this.tokenize(tokenizerResult.tokenizer, localText);
+    const localSentences = this.getSentencesFromTokens(tokenizerResult.tokens, false);
+    localSentences[0] = (0, _lodash.isUndefined)(localSentences[0]) ? "<" : "<" + localSentences[0];
+
+    /*
+     * When the first sentence has a valid sentence beginning.
+     * Add the currently build sentence to the sentences.
+     * Start building the next sentence.
+     */
+    if (this.isValidSentenceBeginning(localSentences[0])) {
+      tokenSentences.push(currentSentence);
+      currentSentence = "";
+    }
+    currentSentence += localSentences[0];
+    if (localSentences.length > 1) {
+      /*
+      	There is a new sentence after the first,
+      	add and reset the current sentence.
+       */
+      tokenSentences.push(currentSentence);
+      currentSentence = "";
+
+      // Remove the first sentence (we do not need to add it again).
+      localSentences.shift();
+      // Last sentence gets special treatment.
+      const lastSentence = localSentences.pop();
+
+      // Add the remaining found sentences.
+      localSentences.forEach(sentence => {
+        tokenSentences.push(sentence);
+      });
+      const sentenceEndRegex = new RegExp("[" + fullStop + this.getSentenceDelimiters() + "]$");
+
+      // Check if the last sentence has a valid sentence ending.
+      if (lastSentence.match(sentenceEndRegex)) {
+        // If so, add it as a sentence.
+        tokenSentences.push(lastSentence);
+      } else {
+        // If not, start making a new one.
+        currentSentence = lastSentence;
+      }
+    }
+    return {
+      tokenSentences,
+      currentSentence
+    };
+  }
+
+  /**
+   * Creates a tokenizer.
+   *
+   * @returns {Object} The tokenizer and the tokens.
+   */
+  createTokenizer() {
+    const sentenceDelimiterRegex = new RegExp("^[" + this.getSentenceDelimiters() + "]$");
+    const sentenceRegex = new RegExp("^[^" + fullStop + this.getSentenceDelimiters() + "<\\(\\)\\[\\]]+$");
+    const tokens = [];
+    const tokenizer = (0, _core.default)(function (token) {
+      tokens.push(token);
+    });
+    tokenizer.addRule(fullStopRegex, "full-stop");
+    tokenizer.addRule(smallerThanContentRegex, "smaller-than-sign-content");
+    tokenizer.addRule(htmlStartRegex, "html-start");
+    tokenizer.addRule(htmlEndRegex, "html-end");
+    tokenizer.addRule(blockStartRegex, "block-start");
+    tokenizer.addRule(blockEndRegex, "block-end");
+    tokenizer.addRule(sentenceDelimiterRegex, "sentence-delimiter");
+    tokenizer.addRule(sentenceRegex, "sentence");
+    return {
+      tokenizer,
+      tokens
+    };
+  }
+
+  /**
+   * Tokenizes the given text using the given tokenizer.
+   *
+   * @param {Object} tokenizer The tokenizer to use.
+   * @param {string} text The text to tokenize.
+   * @returns {void}
+   */
+  tokenize(tokenizer, text) {
+    tokenizer.onText(text);
+    try {
+      tokenizer.end();
+    } catch (e) {
+      console.error("Tokenizer end error:", e, e.tokenizer2);
+    }
+  }
+
+  /**
+   * Checks if a string ends with an abbreviation.
+   * @param {string} currentSentence A (part of) a sentence.
+   * @returns {boolean} True if the string ends with an abbreviation that is in abbreviations.js. Otherwise, False.
+   */
+  endsWithAbbreviation(currentSentence) {
+    const matchedAbbreviations = currentSentence.match(abbreviationsRegex);
+    if (!matchedAbbreviations) {
+      return false;
+    }
+    const lastAbbreviation = matchedAbbreviations.pop();
+    return currentSentence.endsWith(lastAbbreviation);
+  }
+
+  /**
+   * Checks whether the given tokens are a valid html tag pair.
+   * Note that this method is not a full html tag validator. It should be replaced with a better solution once the html parser is implemented.
+   *
+   * @param {object} firstToken   The first token to check. It is asserted that this token contains/is an opening html tag.
+   * @param {object} lastToken    The last token to check. It is asserted that this token contains/is a closing html tag.
+   *
+   * @returns {boolean} True if the tokens are a valid html tag pair. Otherwise, False.
+   */
+  isValidTagPair(firstToken, lastToken) {
+    const firstTokenText = firstToken.src;
+    const lastTokenText = lastToken.src;
+
+    // Get the tag types.
+    const firstTagType = firstTokenText.match(tagTypeRegex)[1];
+    const lastTagType = lastTokenText.match(tagTypeRegex)[1];
+
+    // Check if the tags are the same and if they are a semantic tag (p, div, h1, h2, h3, h4, h5, h6, span).
+    return firstTagType === lastTagType && semanticTags.includes(firstTagType);
+  }
+
+  /**
+   * Returns an array of sentences for a given array of tokens, assumes that the text has already been split into blocks.
+   *
+   * @param {Object[]} tokenArray The tokens from the sentence tokenizer.
+   * @param {boolean} [trimSentences=true] Whether to trim the sentences at the end or not.
+   *
+   * @returns {string[]} A list of sentences.
+   */
+  getSentencesFromTokens(tokenArray, trimSentences = true) {
+    let tokenSentences = [],
+      currentSentence = "",
+      nextSentenceStart,
+      sliced;
+
+    // Drop the first and last HTML tag if both are present.
+    do {
+      sliced = false;
+      const firstToken = tokenArray[0];
+      const lastToken = tokenArray[tokenArray.length - 1];
+      if (firstToken && lastToken && firstToken.type === "html-start" && lastToken.type === "html-end" && this.isValidTagPair(firstToken, lastToken)) {
+        tokenArray = tokenArray.slice(1, tokenArray.length - 1);
+        sliced = true;
+      }
+    } while (sliced && tokenArray.length > 1);
+    tokenArray.forEach((token, i) => {
+      let hasNextSentence, nextCharacters, tokenizeResults;
+      const nextToken = tokenArray[i + 1];
+      const previousToken = tokenArray[i - 1];
+      const secondToNextToken = tokenArray[i + 2];
+      nextCharacters = this.getNextTwoCharacters([nextToken, secondToNextToken]);
+
+      // For a new sentence we need to check the next two characters.
+      hasNextSentence = nextCharacters.length >= 2;
+      nextSentenceStart = hasNextSentence ? nextCharacters[1] : "";
+      switch (token.type) {
+        case "html-start":
+        case "html-end":
+          if (this.isBreakTag(token.src)) {
+            tokenSentences.push(currentSentence);
+            currentSentence = "";
+          } else {
+            currentSentence += token.src;
+          }
+          break;
+        case "smaller-than-sign-content":
+          tokenizeResults = this.tokenizeSmallerThanContent(token, tokenSentences, currentSentence);
+          tokenSentences = tokenizeResults.tokenSentences;
+          currentSentence = tokenizeResults.currentSentence;
+          break;
+        case "sentence":
+          currentSentence += token.src;
+          break;
+        case "sentence-delimiter":
+          currentSentence += token.src;
+
+          /*
+              * Only split text into sentences if:
+              * the next token is defined, AND
+              * the next token type is neither "block-end" nor "sentence-delimiter", AND
+              * the next token first character is a white space
+             */
+          if (!(0, _lodash.isUndefined)(nextToken) && "block-end" !== nextToken.type && "sentence-delimiter" !== nextToken.type && this.isCharacterASpace(nextToken.src[0])) {
+            // Don't split on quotation marks unless they're preceded by a full stop.
+            if (this.isQuotation(token.src) && previousToken && previousToken.src !== ".") {
+              break;
+            }
+            /*
+                   * Only split on ellipsis or quotation marks when:
+                * a) There is a next sentence, and the next character is a valid sentence beginning preceded by a white space, OR
+                * b) The next token is a sentence start
+               */
+            if (this.isQuotation(token.src) || token.src === "…") {
+              currentSentence = this.getValidSentence(hasNextSentence, nextSentenceStart, nextCharacters, nextToken, tokenSentences, currentSentence);
+            } else {
+              tokenSentences.push(currentSentence);
+              currentSentence = "";
+            }
+          }
+          break;
+        case "full-stop":
+          currentSentence += token.src;
+          nextCharacters = this.getNextTwoCharacters([nextToken, secondToNextToken]);
+
+          // For a new sentence we need to check the next two characters.
+          hasNextSentence = nextCharacters.length >= 2;
+          nextSentenceStart = hasNextSentence ? nextCharacters[1] : "";
+
+          // If the current sentence ends with an abbreviation, the full stop does not split the sentence.
+          if (this.endsWithAbbreviation(currentSentence)) {
+            break;
+          }
+
+          // It should not split the text if the first character of the potential next sentence is a number.
+          if (hasNextSentence && this.isNumber(nextCharacters[0])) {
+            break;
+          }
+
+          // If the full stop is part of a person's initials, don't split sentence.
+          if (this.isPartOfPersonInitial(token, previousToken, nextToken, secondToNextToken)) {
+            break;
+          }
+
+          // If the full stop is an ordinal dot (in German), then don't break the sentence.
+          // This check should be done after  hasNextSentence && this.isNumber( nextCharacters[ 0 ] ) (above).
+          // Because otherwise it could break before that test.
+          if (this.endsWithOrdinalDot(currentSentence)) {
+            break;
+          }
+
+          /*
+           * Only split on full stop when:
+           * a) There is a next sentence, and the next character is a valid sentence beginning preceded by a white space, OR
+           * b) The next token is a sentence start
+           */
+          currentSentence = this.getValidSentence(hasNextSentence, nextSentenceStart, nextCharacters, nextToken, tokenSentences, currentSentence);
+          break;
+        case "block-start":
+          currentSentence += token.src;
+          break;
+        case "block-end":
+          currentSentence += token.src;
+          nextCharacters = this.getNextTwoCharacters([nextToken, secondToNextToken]);
+
+          // For a new sentence we need to check the next two characters.
+          hasNextSentence = nextCharacters.length >= 2;
+          nextSentenceStart = hasNextSentence ? nextCharacters[0] : "";
+
+          /* Don't split if:
+           * - The next character is a number. For example: IPv4-numbers.
+           * - The block end is preceded by a valid sentence ending, but not followed by a valid sentence beginning.
+           */
+          if (hasNextSentence && this.isNumber(nextCharacters[0]) || this.isSentenceEnding(previousToken) && !(this.isValidSentenceBeginning(nextSentenceStart) || this.isSentenceStart(nextToken))) {
+            break;
+          }
+
+          /*
+           * Split if:
+           * - The block end is preceded by a sentence ending and followed by a valid sentence beginning.
+           */
+          if (this.isSentenceEnding(previousToken) && (this.isSentenceStart(nextToken) || this.isValidSentenceBeginning(nextSentenceStart))) {
+            tokenSentences.push(currentSentence);
+            currentSentence = "";
+          }
+          break;
+      }
+    });
+    if ("" !== currentSentence) {
+      tokenSentences.push(currentSentence);
+    }
+    if (trimSentences) {
+      tokenSentences = (0, _lodash.map)(tokenSentences, function (sentence) {
+        return sentence.trim();
+      });
+    }
+    return tokenSentences;
+  }
+
+  /**
+   * Gets the current sentence when:
+   * a) There is a next sentence, and the next character is a valid sentence beginning preceded by a white space, OR
+   * b) The next token is a sentence start
+   *
+   * @param {boolean} hasNextSentence     Whether the next characters are more than two.
+   * @param {string} nextSentenceStart    The second character of the next characters.
+   * @param {string} nextCharacters       The string values of the next two tokens.
+   * @param {object} nextToken            The next token object.
+   * @param {array} tokenSentences        The array of pushed valid sentences.
+   * @param {string} currentSentence      The current sentence.
+   *
+   * @returns {string} The current sentence.
+   */
+  getValidSentence(hasNextSentence, nextSentenceStart, nextCharacters, nextToken, tokenSentences, currentSentence) {
+    if (hasNextSentence && this.isValidSentenceBeginning(nextSentenceStart) && this.isCharacterASpace(nextCharacters[0]) || this.isSentenceStart(nextToken)) {
+      tokenSentences.push(currentSentence);
+      currentSentence = "";
+    }
+    return currentSentence;
+  }
+
+  /**
+   * Checks if the character is a whitespace.
+   *
+   * @param {string} character    The character to check.
+   * @returns {boolean}   Whether the character is a whitespace.
+   */
+  isCharacterASpace(character) {
+    return /\s/.test(character);
+  }
+}
+exports.default = SentenceTokenizer;
+//# sourceMappingURL=SentenceTokenizer.js.map

package/build/cjs/languageProcessing/helpers/sentence/SentenceTokenizer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"SentenceTokenizer.js","names":["_lodash","require","_core","_interopRequireDefault","_quotes","_abbreviations","_createRegexFromArray","_wordBoundaries","e","__esModule","default","fullStop","fullStopRegex","RegExp","smallerThanContentRegex","htmlStartRegex","htmlEndRegex","blockStartRegex","blockEndRegex","abbreviationsPreparedForRegex","abbreviations","map","abbreviation","replace","abbreviationsRegex","createRegexFromArray","wordBoundariesForRegex","wordBoundaries","boundary","join","lastCharacterPartOfInitialsRegex","tagTypeRegex","semanticTags","SentenceTokenizer","constructor","sentenceDelimiters","getSentenceDelimiters","isNumber","character","isNaN","parseInt","isBreakTag","htmlTag","test","isQuotation","normalizeQuotes","endsWithOrdinalDot","isPunctuation","removeDuplicateWhitespace","text","isCapitalLetter","toLocaleLowerCase","isSmallerThanSign","getNextTwoCharacters","nextTokens","next","isUndefined","src","isLetterFromSpecificLanguage","letter","ltrLetterRanges","some","ltrLetterRange","isValidSentenceBeginning","sentenceBeginning","isSentenceStart","token","type","isSentenceEnding","isPartOfPersonInitial","previousToken","nextToken","secondToNextToken","trim","length","tokenizeSmallerThanContent","tokenSentences","currentSentence","localText","substring","tokenizerResult","createTokenizer","tokenize","tokenizer","localSentences","getSentencesFromTokens","tokens","push","shift","lastSentence","pop","forEach","sentence","sentenceEndRegex","match","sentenceDelimiterRegex","sentenceRegex","core","addRule","onText","end","console","error","tokenizer2","endsWithAbbreviation","matchedAbbreviations","lastAbbreviation","endsWith","isValidTagPair","firstToken","lastToken","firstTokenText","lastTokenText","firstTagType","lastTagType","includes","tokenArray","trimSentences","nextSentenceStart","sliced","slice","i","hasNextSentence","nextCharacters","tokenizeResults","isCharacterASpace","getValidSentence","exports"],"sources":["../../../../../src/languageProcessing/helpers/sentence/SentenceTokenizer.js"],"sourcesContent":["import { isNaN, isUndefined, map } from \"lodash\";\n\nimport core from \"tokenizer2/core\";\n\nimport { normalize as normalizeQuotes } from \"../sanitize/quotes.js\";\n\nimport abbreviations from \"../../languages/en/config/abbreviations\";\n\nimport createRegexFromArray from \"../regex/createRegexFromArray\";\nimport wordBoundaries from \"../../../config/wordBoundaries\";\n\n// All characters that indicate a sentence delimiter.\nconst fullStop = \".\";\n\nconst fullStopRegex = new RegExp( \"^[\" + fullStop + \"]$\" );\nconst smallerThanContentRegex = /^<[^><]*$/;\nconst htmlStartRegex = /^<([^>\\s/]+)[^>]*>$/mi;\nconst htmlEndRegex = /^<\\/([^>\\s]+)[^>]*>$/mi;\n\nconst blockStartRegex = /^\\s*[[({]\\s*$/;\nconst blockEndRegex = /^\\s*[\\])}]\\s*$/;\n\nconst abbreviationsPreparedForRegex = abbreviations.map( ( abbreviation ) => abbreviation.replace( \".\", \"\\\\.\" ) );\nconst abbreviationsRegex = createRegexFromArray( abbreviationsPreparedForRegex );\n\nconst wordBoundariesForRegex = \"(^|$|[\" + wordBoundaries().map( ( boundary ) => \"\\\\\" + boundary ).join( \"\" ) + \"])\";\nconst lastCharacterPartOfInitialsRegex = new RegExp( wordBoundariesForRegex + \"[A-Za-z]$\" );\n\n// Constants to be used in isValidTagPair.\n// A regex to get the tag type.\nconst tagTypeRegex = /<\\/?([^\\s]+?)(\\s|>)/;\n// Semantic tags (as opposed to style tags) are tags that are used to structure the text.\nconst semanticTags = [ \"p\", \"div\", \"h1\", \"h2\", \"h3\", \"h4\", \"h5\", \"h6\", \"span\", \"li\", \"main\" ];\n\n/**\n * Class for tokenizing a (html) text into sentences.\n */\nexport default class SentenceTokenizer {\n\t/**\n\t * Constructor\n\t * @constructor\n\t */\n\tconstructor() {\n\t\t/*\n         * \\u2026 - ellipsis.\n         * \\u06D4 - Urdu full stop.\n         * \\u061f - Arabic question mark.\n        */\n\t\tthis.sentenceDelimiters = \"”〞〟„』›»’‛`\\\"?!\\u2026\\u06d4\\u061f\";\n\t}\n\n\t/**\n\t * Gets the sentence delimiters.\n\t *\n\t * @returns {string} The sentence delimiters.\n\t */\n\tgetSentenceDelimiters() {\n\t\treturn this.sentenceDelimiters;\n\t}\n\n\t/**\n\t * Returns whether or not a certain character is a number.\n\t *\n\t * @param {string} character The character to check.\n\t * @returns {boolean} Whether or not the character is a capital letter.\n\t */\n\tisNumber( character ) {\n\t\treturn ! isNaN( parseInt( character, 10 ) );\n\t}\n\n\t/**\n\t * Returns whether or not a given HTML tag is a break tag.\n\t *\n\t * @param {string} htmlTag The HTML tag to check.\n\t * @returns {boolean} Whether or not the given HTML tag is a break tag.\n\t */\n\tisBreakTag( htmlTag ) {\n\t\treturn /<\\/?br/.test( htmlTag );\n\t}\n\n\t/**\n\t * Returns whether or not a given character is quotation mark.\n\t *\n\t * @param {string} character The character to check.\n\t * @returns {boolean} Whether or not the given character is a quotation mark.\n\t */\n\tisQuotation( character ) {\n\t\tcharacter = normalizeQuotes( character );\n\n\t\treturn \"'\" === character ||\n\t\t\t\"\\\"\" === character;\n\t}\n\n\t/**\n\t * A mock definition of this function. This function is only used in extensions for languages that use an ordinal dot.\n\t *\n\t * @returns {boolean} Always returns false as it is a language specific implementation if a language has an ordinal dot.\n\t */\n\tendsWithOrdinalDot() {\n\t\treturn false;\n\t}\n\n\t/**\n\t * Returns whether or not a given character is a punctuation mark that can be at the beginning\n\t * of a sentence, like ¿ and ¡ used in Spanish.\n\t *\n\t * @param {string} character The character to check.\n\t * @returns {boolean} Whether or not the given character is a punctuation mark.\n\t */\n\tisPunctuation( character ) {\n\t\treturn \"¿\" === character ||\n\t\t\t\"¡\" === character;\n\t}\n\n\t/**\n\t * Removes duplicate whitespace from a given text.\n\t *\n\t * @param {string} text The text with duplicate whitespace.\n\t * @returns {string} The text without duplicate whitespace.\n\t */\n\tremoveDuplicateWhitespace( text ) {\n\t\treturn text.replace( /\\s+/, \" \" );\n\t}\n\n\t/**\n\t * Returns whether or not a certain character is a capital letter.\n\t *\n\t * @param {string} character The character to check.\n\t * @returns {boolean} Whether or not the character is a capital letter.\n\t */\n\tisCapitalLetter( character ) {\n\t\treturn character !== character.toLocaleLowerCase();\n\t}\n\n\t/**\n\t * Checks whether the given character is a smaller than sign.\n\t *\n\t * This function is used to make sure that tokenizing the content after\n\t * the smaller than sign works as expected.\n\t * E.g. 'A sentence. < Hello world!' = ['A sentence.', '< Hello world!'].\n\t *\n\t * @param {string} character The character to check.\n\t * @returns {boolean} Whether the character is a smaller than sign ('<') or not.\n\t */\n\tisSmallerThanSign( character ) {\n\t\treturn character === \"<\";\n\t}\n\n\t/**\n\t * Retrieves the next two characters from an array with the two next tokens.\n\t *\n\t * @param {Array} nextTokens The two next tokens. Might be undefined.\n\t * @returns {string} The next two characters.\n\t */\n\tgetNextTwoCharacters( nextTokens ) {\n\t\tlet next = \"\";\n\n\t\tif ( ! isUndefined( nextTokens[ 0 ] ) ) {\n\t\t\tnext += nextTokens[ 0 ].src;\n\t\t}\n\n\t\tif ( ! isUndefined( nextTokens[ 1 ] ) ) {\n\t\t\tnext += nextTokens[ 1 ].src;\n\t\t}\n\n\t\tnext = this.removeDuplicateWhitespace( next );\n\n\t\treturn next;\n\t}\n\n\t/**\n\t * Checks whether a character is from a language that's written from right to left.\n\t * These languages don't have capital letter forms. Therefore any letter from these languages is a\n\t * potential sentence beginning.\n\t *\n\t * @param {string} letter The letter to check.\n\t *\n\t * @returns {boolean} Whether the letter is from an LTR language.\n\t */\n\tisLetterFromSpecificLanguage( letter ) {\n\t\tconst ltrLetterRanges = [\n\t\t\t// Hebrew characters.\n\t\t\t/^[\\u0590-\\u05fe]+$/i,\n\t\t\t// Arabic characters (used for Arabic, Farsi, Urdu).\n\t\t\t/^[\\u0600-\\u06FF]+$/i,\n\t\t\t// Additional Farsi characters.\n\t\t\t/^[\\uFB8A\\u067E\\u0686\\u06AF]+$/i,\n\t\t];\n\n\t\treturn (\n\t\t\tltrLetterRanges.some( ltrLetterRange => ltrLetterRange.test( letter ) )\n\t\t);\n\t}\n\n\t/**\n\t * Checks if the sentenceBeginning beginning is a valid beginning.\n\t *\n\t * @param {string} sentenceBeginning The beginning of the sentence to validate.\n\t * @returns {boolean} Returns true if it is a valid beginning, false if it is not.\n\t */\n\tisValidSentenceBeginning( sentenceBeginning ) {\n\t\treturn ( this.isCapitalLetter( sentenceBeginning ) ||\n\t\t\t\tthis.isLetterFromSpecificLanguage( sentenceBeginning ) ||\n\t\t\t\tthis.isNumber( sentenceBeginning ) ||\n\t\t\t\tthis.isQuotation( sentenceBeginning ) ||\n\t\t\t\tthis.isPunctuation( sentenceBeginning ) ||\n\t\t\t\tthis.isSmallerThanSign( sentenceBeginning ) );\n\t}\n\n\t/**\n\t * Checks if the token is a valid sentence start.\n\t *\n\t * @param {Object} token The token to validate.\n\t * @returns {boolean} Returns true if the token is valid sentence start, false if it is not.\n\t */\n\tisSentenceStart( token ) {\n\t\treturn ( ! isUndefined( token ) && (\n\t\t\t\"html-start\" === token.type ||\n\t\t\t\"html-end\" === token.type ||\n\t\t\t\"block-start\" === token.type\n\t\t) );\n\t}\n\n\t/**\n\t * Checks if the token is a valid sentence ending. A valid sentence ending is either a full stop or another\n\t * delimiter such as \"?\", \"!\", etc.\n\t *\n\t * @param {Object} token The token to validate.\n\t * @returns {boolean} Returns true if the token is valid sentence ending, false if it is not.\n\t */\n\tisSentenceEnding( token ) {\n\t\treturn (\n\t\t\t! isUndefined( token ) &&\n\t\t\t( token.type === \"full-stop\" || token.type === \"sentence-delimiter\" )\n\t\t);\n\t}\n\n\t/**\n\t * Checks if a full stop is part of a person's initials.\n\t *\n\t * Tests if tokens exist. Then tests if the tokens are of the right type.\n\t * For previous token, it checks if the sentence ends with a single letter.\n\t * For nextToken it checks if it is a single letter.\n\t * Checks if next token is followed by a full stop.\n\t *\n\t * @param {object} token The current token (must be a full stop).\n\t * @param {object} previousToken The token before the full stop.\n\t * @param {object} nextToken The token following the full stop.\n\t * @param {object} secondToNextToken The second token after the full stop.\n\t * @returns {boolean} True if a full stop is part of a person's initials, False if the full stop is not part of a person's initials.\n\t */\n\tisPartOfPersonInitial( token, previousToken, nextToken, secondToNextToken ) {\n\t\treturn ( ! isUndefined( token ) &&\n\t\t\t! isUndefined( nextToken ) &&\n\t\t\t! isUndefined( secondToNextToken ) &&\n\t\t\t! isUndefined( previousToken ) &&\n\t\t\ttoken.type === \"full-stop\" &&\n\t\t\tpreviousToken.type === \"sentence\" &&\n\t\t\tlastCharacterPartOfInitialsRegex.test( previousToken.src ) &&\n\t\t\tnextToken.type === \"sentence\" &&\n\t\t\tnextToken.src.trim().length === 1 &&\n\t\t\tsecondToNextToken.type === \"full-stop\"\n\t\t);\n\t}\n\n\t/**\n\t * Tokens that represent a '<', followed by content until it enters another '<' or '>'\n\t * gets another pass by the tokenizer.\n\t *\n\t * @param {Object} token A token of type 'smaller-than-sign-content'.\n\t * @param {string[]} tokenSentences The current array of found sentences. Sentences may get added by this method.\n\t * @param {string} currentSentence The current sentence. Sentence parts may get appended by this method.\n\t * @returns {{tokenSentences, currentSentence}} The found sentences and the current sentence, appended when necessary.\n\t */\n\ttokenizeSmallerThanContent( token, tokenSentences, currentSentence ) {\n\t\t/*\n\t\t\tRemove the '<' from the text, to avoid matching this rule\n\t\t\trecursively again and again.\n\t\t\tWe add it again later on.\n\t\t*/\n\t\tconst localText = token.src.substring( 1 );\n\n\t\t// Tokenize the current smaller-than-content token without the first '<'.\n\t\tconst tokenizerResult = this.createTokenizer();\n\t\tthis.tokenize( tokenizerResult.tokenizer, localText );\n\t\tconst localSentences = this.getSentencesFromTokens( tokenizerResult.tokens, false );\n\n\t\tlocalSentences[ 0 ] = isUndefined( localSentences[ 0 ] ) ? \"<\" : \"<\" + localSentences[ 0 ];\n\n\t\t/*\n\t\t * When the first sentence has a valid sentence beginning.\n\t\t * Add the currently build sentence to the sentences.\n\t\t * Start building the next sentence.\n\t\t */\n\t\tif ( this.isValidSentenceBeginning( localSentences[ 0 ] ) ) {\n\t\t\ttokenSentences.push( currentSentence );\n\t\t\tcurrentSentence = \"\";\n\t\t}\n\t\tcurrentSentence += localSentences[ 0 ];\n\n\t\tif ( localSentences.length > 1 ) {\n\t\t\t/*\n\t\t\t\tThere is a new sentence after the first,\n\t\t\t\tadd and reset the current sentence.\n\t\t\t */\n\t\t\ttokenSentences.push( currentSentence );\n\t\t\tcurrentSentence = \"\";\n\n\t\t\t// Remove the first sentence (we do not need to add it again).\n\t\t\tlocalSentences.shift();\n\t\t\t// Last sentence gets special treatment.\n\t\t\tconst lastSentence = localSentences.pop();\n\n\t\t\t// Add the remaining found sentences.\n\t\t\tlocalSentences.forEach( sentence => {\n\t\t\t\ttokenSentences.push( sentence );\n\t\t\t} );\n\n\t\t\tconst sentenceEndRegex = new RegExp( \"[\" + fullStop + this.getSentenceDelimiters() + \"]$\" );\n\n\t\t\t// Check if the last sentence has a valid sentence ending.\n\t\t\tif ( lastSentence.match( sentenceEndRegex ) ) {\n\t\t\t\t// If so, add it as a sentence.\n\t\t\t\ttokenSentences.push( lastSentence );\n\t\t\t} else {\n\t\t\t\t// If not, start making a new one.\n\t\t\t\tcurrentSentence = lastSentence;\n\t\t\t}\n\t\t}\n\t\treturn {\n\t\t\ttokenSentences,\n\t\t\tcurrentSentence,\n\t\t};\n\t}\n\n\t/**\n\t * Creates a tokenizer.\n\t *\n\t * @returns {Object} The tokenizer and the tokens.\n\t */\n\tcreateTokenizer() {\n\t\tconst sentenceDelimiterRegex = new RegExp( \"^[\" + this.getSentenceDelimiters() + \"]$\" );\n\t\tconst sentenceRegex = new RegExp( \"^[^\" + fullStop + this.getSentenceDelimiters() + \"<\\\\(\\\\)\\\\[\\\\]]+$\" );\n\n\t\tconst tokens = [];\n\t\tconst tokenizer = core( function( token ) {\n\t\t\ttokens.push( token );\n\t\t} );\n\n\t\ttokenizer.addRule( fullStopRegex, \"full-stop\" );\n\t\ttokenizer.addRule( smallerThanContentRegex, \"smaller-than-sign-content\" );\n\t\ttokenizer.addRule( htmlStartRegex, \"html-start\" );\n\t\ttokenizer.addRule( htmlEndRegex, \"html-end\" );\n\t\ttokenizer.addRule( blockStartRegex, \"block-start\" );\n\t\ttokenizer.addRule( blockEndRegex, \"block-end\" );\n\t\ttokenizer.addRule( sentenceDelimiterRegex, \"sentence-delimiter\" );\n\t\ttokenizer.addRule( sentenceRegex, \"sentence\" );\n\n\t\treturn {\n\t\t\ttokenizer,\n\t\t\ttokens,\n\t\t};\n\t}\n\n\t/**\n\t * Tokenizes the given text using the given tokenizer.\n\t *\n\t * @param {Object} tokenizer The tokenizer to use.\n\t * @param {string} text The text to tokenize.\n\t * @returns {void}\n\t */\n\ttokenize( tokenizer, text ) {\n\t\ttokenizer.onText( text );\n\n\t\ttry {\n\t\t\ttokenizer.end();\n\t\t} catch ( e ) {\n\t\t\tconsole.error( \"Tokenizer end error:\", e, e.tokenizer2 );\n\t\t}\n\t}\n\n\t/**\n\t * Checks if a string ends with an abbreviation.\n\t * @param {string} currentSentence A (part of) a sentence.\n\t * @returns {boolean} True if the string ends with an abbreviation that is in abbreviations.js. Otherwise, False.\n\t */\n\tendsWithAbbreviation( currentSentence ) {\n\t\tconst matchedAbbreviations = currentSentence.match( abbreviationsRegex );\n\n\t\tif ( ! matchedAbbreviations ) {\n\t\t\treturn false;\n\t\t}\n\n\t\tconst lastAbbreviation = matchedAbbreviations.pop();\n\t\treturn currentSentence.endsWith( lastAbbreviation );\n\t}\n\n\t/**\n\t * Checks whether the given tokens are a valid html tag pair.\n\t * Note that this method is not a full html tag validator. It should be replaced with a better solution once the html parser is implemented.\n\t *\n\t * @param {object} firstToken   The first token to check. It is asserted that this token contains/is an opening html tag.\n\t * @param {object} lastToken    The last token to check. It is asserted that this token contains/is a closing html tag.\n\t *\n\t * @returns {boolean} True if the tokens are a valid html tag pair. Otherwise, False.\n\t */\n\tisValidTagPair( firstToken, lastToken ) {\n\t\tconst firstTokenText = firstToken.src;\n\t\tconst lastTokenText = lastToken.src;\n\n\t\t// Get the tag types.\n\t\tconst firstTagType = firstTokenText.match( tagTypeRegex )[ 1 ];\n\t\tconst lastTagType  = lastTokenText.match( tagTypeRegex )[ 1 ];\n\n\n\t\t// Check if the tags are the same and if they are a semantic tag (p, div, h1, h2, h3, h4, h5, h6, span).\n\t\treturn firstTagType === lastTagType && semanticTags.includes( firstTagType );\n\t}\n\n\t/**\n\t * Returns an array of sentences for a given array of tokens, assumes that the text has already been split into blocks.\n\t *\n\t * @param {Object[]} tokenArray The tokens from the sentence tokenizer.\n\t * @param {boolean} [trimSentences=true] Whether to trim the sentences at the end or not.\n\t *\n\t * @returns {string[]} A list of sentences.\n\t */\n\tgetSentencesFromTokens( tokenArray, trimSentences = true ) {\n\t\tlet tokenSentences = [], currentSentence = \"\", nextSentenceStart, sliced;\n\n\t\t// Drop the first and last HTML tag if both are present.\n\t\tdo {\n\t\t\tsliced = false;\n\t\t\tconst firstToken = tokenArray[ 0 ];\n\t\t\tconst lastToken = tokenArray[ tokenArray.length - 1 ];\n\n\t\t\tif ( firstToken && lastToken && firstToken.type === \"html-start\" &&\n\t\t\t\tlastToken.type === \"html-end\" && this.isValidTagPair( firstToken, lastToken ) ) {\n\t\t\t\ttokenArray = tokenArray.slice( 1, tokenArray.length - 1 );\n\n\t\t\t\tsliced = true;\n\t\t\t}\n\t\t} while ( sliced && tokenArray.length > 1 );\n\n\t\ttokenArray.forEach( ( token, i ) => {\n\t\t\tlet hasNextSentence, nextCharacters, tokenizeResults;\n\t\t\tconst nextToken = tokenArray[ i + 1 ];\n\t\t\tconst previousToken = tokenArray[ i - 1 ];\n\t\t\tconst secondToNextToken = tokenArray[ i + 2 ];\n\t\t\tnextCharacters = this.getNextTwoCharacters( [ nextToken, secondToNextToken ] );\n\n\t\t\t// For a new sentence we need to check the next two characters.\n\t\t\thasNextSentence = nextCharacters.length >= 2;\n\t\t\tnextSentenceStart = hasNextSentence ? nextCharacters[ 1 ] : \"\";\n\n\t\t\tswitch ( token.type ) {\n\t\t\t\tcase \"html-start\":\n\t\t\t\tcase \"html-end\":\n\t\t\t\t\tif ( this.isBreakTag( token.src ) ) {\n\t\t\t\t\t\ttokenSentences.push( currentSentence );\n\t\t\t\t\t\tcurrentSentence = \"\";\n\t\t\t\t\t} else {\n\t\t\t\t\t\tcurrentSentence += token.src;\n\t\t\t\t\t}\n\t\t\t\t\tbreak;\n\n\t\t\t\tcase \"smaller-than-sign-content\":\n\t\t\t\t\ttokenizeResults = this.tokenizeSmallerThanContent( token, tokenSentences, currentSentence );\n\t\t\t\t\ttokenSentences = tokenizeResults.tokenSentences;\n\t\t\t\t\tcurrentSentence = tokenizeResults.currentSentence;\n\t\t\t\t\tbreak;\n\t\t\t\tcase \"sentence\":\n\t\t\t\t\tcurrentSentence += token.src;\n\t\t\t\t\tbreak;\n\t\t\t\tcase \"sentence-delimiter\":\n\t\t\t\t\tcurrentSentence += token.src;\n\n\t\t\t\t\t/*\n\t\t\t\t     * Only split text into sentences if:\n\t\t\t\t     * the next token is defined, AND\n\t\t\t\t     * the next token type is neither \"block-end\" nor \"sentence-delimiter\", AND\n\t\t\t\t     * the next token first character is a white space\n\t\t\t\t    */\n\t\t\t\t\tif ( ! isUndefined( nextToken ) &&\n\t\t\t\t\t\t\"block-end\" !== nextToken.type &&\n\t\t\t\t\t\t\"sentence-delimiter\" !== nextToken.type &&\n\t\t\t\t\t\tthis.isCharacterASpace( nextToken.src[ 0 ] ) ) {\n\t\t\t\t\t\t// Don't split on quotation marks unless they're preceded by a full stop.\n\t\t\t\t\t\tif ( this.isQuotation( token.src ) && previousToken && previousToken.src !== \".\" ) {\n\t\t\t\t\t\t\tbreak;\n\t\t\t\t\t\t}\n\t\t\t\t\t\t/*\n\t\t\t\t         * Only split on ellipsis or quotation marks when:\n\t\t\t\t\t     * a) There is a next sentence, and the next character is a valid sentence beginning preceded by a white space, OR\n\t\t\t\t\t     * b) The next token is a sentence start\n\t\t\t\t\t    */\n\t\t\t\t\t\tif ( this.isQuotation( token.src ) || token.src === \"…\" ) {\n\t\t\t\t\t\t\tcurrentSentence = this.getValidSentence( hasNextSentence,\n\t\t\t\t\t\t\t\tnextSentenceStart,\n\t\t\t\t\t\t\t\tnextCharacters,\n\t\t\t\t\t\t\t\tnextToken,\n\t\t\t\t\t\t\t\ttokenSentences,\n\t\t\t\t\t\t\t\tcurrentSentence );\n\t\t\t\t\t\t} else {\n\t\t\t\t\t\t\ttokenSentences.push( currentSentence );\n\t\t\t\t\t\t\tcurrentSentence = \"\";\n\t\t\t\t\t\t}\n\t\t\t\t\t}\n\t\t\t\t\tbreak;\n\n\t\t\t\tcase \"full-stop\":\n\t\t\t\t\tcurrentSentence += token.src;\n\t\t\t\t\tnextCharacters = this.getNextTwoCharacters( [ nextToken, secondToNextToken ] );\n\n\t\t\t\t\t// For a new sentence we need to check the next two characters.\n\t\t\t\t\thasNextSentence = nextCharacters.length >= 2;\n\t\t\t\t\tnextSentenceStart = hasNextSentence ? nextCharacters[ 1 ] : \"\";\n\n\t\t\t\t\t// If the current sentence ends with an abbreviation, the full stop does not split the sentence.\n\t\t\t\t\tif ( this.endsWithAbbreviation( currentSentence ) ) {\n\t\t\t\t\t\tbreak;\n\t\t\t\t\t}\n\n\t\t\t\t\t// It should not split the text if the first character of the potential next sentence is a number.\n\t\t\t\t\tif ( hasNextSentence && this.isNumber( nextCharacters[ 0 ] ) ) {\n\t\t\t\t\t\tbreak;\n\t\t\t\t\t}\n\n\t\t\t\t\t// If the full stop is part of a person's initials, don't split sentence.\n\t\t\t\t\tif ( this.isPartOfPersonInitial( token, previousToken, nextToken, secondToNextToken ) ) {\n\t\t\t\t\t\tbreak;\n\t\t\t\t\t}\n\n\t\t\t\t\t// If the full stop is an ordinal dot (in German), then don't break the sentence.\n\t\t\t\t\t// This check should be done after  hasNextSentence && this.isNumber( nextCharacters[ 0 ] ) (above).\n\t\t\t\t\t// Because otherwise it could break before that test.\n\t\t\t\t\tif ( this.endsWithOrdinalDot( currentSentence ) ) {\n\t\t\t\t\t\tbreak;\n\t\t\t\t\t}\n\n\t\t\t\t\t/*\n\t\t\t\t\t * Only split on full stop when:\n\t\t\t\t\t * a) There is a next sentence, and the next character is a valid sentence beginning preceded by a white space, OR\n\t\t\t\t\t * b) The next token is a sentence start\n\t\t\t\t\t */\n\t\t\t\t\tcurrentSentence = this.getValidSentence( hasNextSentence,\n\t\t\t\t\t\tnextSentenceStart,\n\t\t\t\t\t\tnextCharacters,\n\t\t\t\t\t\tnextToken,\n\t\t\t\t\t\ttokenSentences,\n\t\t\t\t\t\tcurrentSentence );\n\n\t\t\t\t\tbreak;\n\n\t\t\t\tcase \"block-start\":\n\t\t\t\t\tcurrentSentence += token.src;\n\t\t\t\t\tbreak;\n\n\t\t\t\tcase \"block-end\":\n\t\t\t\t\tcurrentSentence += token.src;\n\n\t\t\t\t\tnextCharacters = this.getNextTwoCharacters( [ nextToken, secondToNextToken ] );\n\n\t\t\t\t\t// For a new sentence we need to check the next two characters.\n\t\t\t\t\thasNextSentence = nextCharacters.length >= 2;\n\t\t\t\t\tnextSentenceStart = hasNextSentence ? nextCharacters[ 0 ] : \"\";\n\n\t\t\t\t\t/* Don't split if:\n\t\t\t\t\t * - The next character is a number. For example: IPv4-numbers.\n\t\t\t\t\t * - The block end is preceded by a valid sentence ending, but not followed by a valid sentence beginning.\n\t\t\t\t\t */\n\t\t\t\t\tif (\n\t\t\t\t\t\thasNextSentence && this.isNumber( nextCharacters[ 0 ] ) ||\n\t\t\t\t\t\t( this.isSentenceEnding( previousToken ) &&\n\t\t\t\t\t\t\t( ! ( this.isValidSentenceBeginning( nextSentenceStart ) || this.isSentenceStart( nextToken ) ) ) )\n\t\t\t\t\t) {\n\t\t\t\t\t\tbreak;\n\t\t\t\t\t}\n\n\t\t\t\t\t/*\n\t\t\t\t\t * Split if:\n\t\t\t\t\t * - The block end is preceded by a sentence ending and followed by a valid sentence beginning.\n\t\t\t\t\t */\n\t\t\t\t\tif (\n\t\t\t\t\t\tthis.isSentenceEnding( previousToken ) &&\n\t\t\t\t\t\t( this.isSentenceStart( nextToken ) || this.isValidSentenceBeginning( nextSentenceStart ) )\n\t\t\t\t\t) {\n\t\t\t\t\t\ttokenSentences.push( currentSentence );\n\t\t\t\t\t\tcurrentSentence = \"\";\n\t\t\t\t\t}\n\t\t\t\t\tbreak;\n\t\t\t}\n\t\t} );\n\n\t\tif ( \"\" !== currentSentence ) {\n\t\t\ttokenSentences.push( currentSentence );\n\t\t}\n\n\t\tif ( trimSentences ) {\n\t\t\ttokenSentences = map( tokenSentences, function( sentence ) {\n\t\t\t\treturn sentence.trim();\n\t\t\t} );\n\t\t}\n\n\t\treturn tokenSentences;\n\t}\n\n\t/**\n\t * Gets the current sentence when:\n\t * a) There is a next sentence, and the next character is a valid sentence beginning preceded by a white space, OR\n\t * b) The next token is a sentence start\n\t *\n\t * @param {boolean} hasNextSentence     Whether the next characters are more than two.\n\t * @param {string} nextSentenceStart    The second character of the next characters.\n\t * @param {string} nextCharacters       The string values of the next two tokens.\n\t * @param {object} nextToken            The next token object.\n\t * @param {array} tokenSentences        The array of pushed valid sentences.\n\t * @param {string} currentSentence      The current sentence.\n\t *\n\t * @returns {string} The current sentence.\n\t */\n\tgetValidSentence( hasNextSentence, nextSentenceStart, nextCharacters, nextToken, tokenSentences, currentSentence ) {\n\t\tif ( ( hasNextSentence && this.isValidSentenceBeginning( nextSentenceStart ) && this.isCharacterASpace( nextCharacters[ 0 ] ) ) ||\n\t\t\tthis.isSentenceStart( nextToken ) ) {\n\t\t\ttokenSentences.push( currentSentence );\n\t\t\tcurrentSentence = \"\";\n\t\t}\n\t\treturn currentSentence;\n\t}\n\n\t/**\n\t * Checks if the character is a whitespace.\n\t *\n\t * @param {string} character    The character to check.\n\t * @returns {boolean}   Whether the character is a whitespace.\n\t */\n\tisCharacterASpace( character ) {\n\t\treturn /\\s/.test( character );\n\t}\n}\n"],"mappings":";;;;;;AAAA,IAAAA,OAAA,GAAAC,OAAA;AAEA,IAAAC,KAAA,GAAAC,sBAAA,CAAAF,OAAA;AAEA,IAAAG,OAAA,GAAAH,OAAA;AAEA,IAAAI,cAAA,GAAAF,sBAAA,CAAAF,OAAA;AAEA,IAAAK,qBAAA,GAAAH,sBAAA,CAAAF,OAAA;AACA,IAAAM,eAAA,GAAAJ,sBAAA,CAAAF,OAAA;AAA4D,SAAAE,uBAAAK,CAAA,WAAAA,CAAA,IAAAA,CAAA,CAAAC,UAAA,GAAAD,CAAA,KAAAE,OAAA,EAAAF,CAAA;AAE5D;AACA,MAAMG,QAAQ,GAAG,GAAG;AAEpB,MAAMC,aAAa,GAAG,IAAIC,MAAM,CAAE,IAAI,GAAGF,QAAQ,GAAG,IAAK,CAAC;AAC1D,MAAMG,uBAAuB,GAAG,WAAW;AAC3C,MAAMC,cAAc,GAAG,uBAAuB;AAC9C,MAAMC,YAAY,GAAG,wBAAwB;AAE7C,MAAMC,eAAe,GAAG,eAAe;AACvC,MAAMC,aAAa,GAAG,gBAAgB;AAEtC,MAAMC,6BAA6B,GAAGC,sBAAa,CAACC,GAAG,CAAIC,YAAY,IAAMA,YAAY,CAACC,OAAO,CAAE,GAAG,EAAE,KAAM,CAAE,CAAC;AACjH,MAAMC,kBAAkB,GAAG,IAAAC,6BAAoB,EAAEN,6BAA8B,CAAC;AAEhF,MAAMO,sBAAsB,GAAG,QAAQ,GAAG,IAAAC,uBAAc,EAAC,CAAC,CAACN,GAAG,CAAIO,QAAQ,IAAM,IAAI,GAAGA,QAAS,CAAC,CAACC,IAAI,CAAE,EAAG,CAAC,GAAG,IAAI;AACnH,MAAMC,gCAAgC,GAAG,IAAIjB,MAAM,CAAEa,sBAAsB,GAAG,WAAY,CAAC;;AAE3F;AACA;AACA,MAAMK,YAAY,GAAG,qBAAqB;AAC1C;AACA,MAAMC,YAAY,GAAG,CAAE,GAAG,EAAE,KAAK,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,CAAE;;AAE7F;AACA;AACA;AACe,MAAMC,iBAAiB,CAAC;EACtC;AACD;AACA;AACA;EACCC,WAAWA,CAAA,EAAG;IACb;AACF;AACA;AACA;AACA;IACE,IAAI,CAACC,kBAAkB,GAAG,kCAAkC;EAC7D;;EAEA;AACD;AACA;AACA;AACA;EACCC,qBAAqBA,CAAA,EAAG;IACvB,OAAO,IAAI,CAACD,kBAAkB;EAC/B;;EAEA;AACD;AACA;AACA;AACA;AACA;EACCE,QAAQA,CAAEC,SAAS,EAAG;IACrB,OAAO,CAAE,IAAAC,aAAK,EAAEC,QAAQ,CAAEF,SAAS,EAAE,EAAG,CAAE,CAAC;EAC5C;;EAEA;AACD;AACA;AACA;AACA;AACA;EACCG,UAAUA,CAAEC,OAAO,EAAG;IACrB,OAAO,QAAQ,CAACC,IAAI,CAAED,OAAQ,CAAC;EAChC;;EAEA;AACD;AACA;AACA;AACA;AACA;EACCE,WAAWA,CAAEN,SAAS,EAAG;IACxBA,SAAS,GAAG,IAAAO,iBAAe,EAAEP,SAAU,CAAC;IAExC,OAAO,GAAG,KAAKA,SAAS,IACvB,IAAI,KAAKA,SAAS;EACpB;;EAEA;AACD;AACA;AACA;AACA;EACCQ,kBAAkBA,CAAA,EAAG;IACpB,OAAO,KAAK;EACb;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;EACCC,aAAaA,CAAET,SAAS,EAAG;IAC1B,OAAO,GAAG,KAAKA,SAAS,IACvB,GAAG,KAAKA,SAAS;EACnB;;EAEA;AACD;AACA;AACA;AACA;AACA;EACCU,yBAAyBA,CAAEC,IAAI,EAAG;IACjC,OAAOA,IAAI,CAAC1B,OAAO,CAAE,KAAK,EAAE,GAAI,CAAC;EAClC;;EAEA;AACD;AACA;AACA;AACA;AACA;EACC2B,eAAeA,CAAEZ,SAAS,EAAG;IAC5B,OAAOA,SAAS,KAAKA,SAAS,CAACa,iBAAiB,CAAC,CAAC;EACnD;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;EACCC,iBAAiBA,CAAEd,SAAS,EAAG;IAC9B,OAAOA,SAAS,KAAK,GAAG;EACzB;;EAEA;AACD;AACA;AACA;AACA;AACA;EACCe,oBAAoBA,CAAEC,UAAU,EAAG;IAClC,IAAIC,IAAI,GAAG,EAAE;IAEb,IAAK,CAAE,IAAAC,mBAAW,EAAEF,UAAU,CAAE,CAAC,CAAG,CAAC,EAAG;MACvCC,IAAI,IAAID,UAAU,CAAE,CAAC,CAAE,CAACG,GAAG;IAC5B;IAEA,IAAK,CAAE,IAAAD,mBAAW,EAAEF,UAAU,CAAE,CAAC,CAAG,CAAC,EAAG;MACvCC,IAAI,IAAID,UAAU,CAAE,CAAC,CAAE,CAACG,GAAG;IAC5B;IAEAF,IAAI,GAAG,IAAI,CAACP,yBAAyB,CAAEO,IAAK,CAAC;IAE7C,OAAOA,IAAI;EACZ;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;AACA;AACA;EACCG,4BAA4BA,CAAEC,MAAM,EAAG;IACtC,MAAMC,eAAe,GAAG;IACvB;IACA,qBAAqB;IACrB;IACA,qBAAqB;IACrB;IACA,gCAAgC,CAChC;IAED,OACCA,eAAe,CAACC,IAAI,CAAEC,cAAc,IAAIA,cAAc,CAACnB,IAAI,CAAEgB,MAAO,CAAE,CAAC;EAEzE;;EAEA;AACD;AACA;AACA;AACA;AACA;EACCI,wBAAwBA,CAAEC,iBAAiB,EAAG;IAC7C,OAAS,IAAI,CAACd,eAAe,CAAEc,iBAAkB,CAAC,IAChD,IAAI,CAACN,4BAA4B,CAAEM,iBAAkB,CAAC,IACtD,IAAI,CAAC3B,QAAQ,CAAE2B,iBAAkB,CAAC,IAClC,IAAI,CAACpB,WAAW,CAAEoB,iBAAkB,CAAC,IACrC,IAAI,CAACjB,aAAa,CAAEiB,iBAAkB,CAAC,IACvC,IAAI,CAACZ,iBAAiB,CAAEY,iBAAkB,CAAC;EAC9C;;EAEA;AACD;AACA;AACA;AACA;AACA;EACCC,eAAeA,CAAEC,KAAK,EAAG;IACxB,OAAS,CAAE,IAAAV,mBAAW,EAAEU,KAAM,CAAC,KAC9B,YAAY,KAAKA,KAAK,CAACC,IAAI,IAC3B,UAAU,KAAKD,KAAK,CAACC,IAAI,IACzB,aAAa,KAAKD,KAAK,CAACC,IAAI,CAC5B;EACF;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;EACCC,gBAAgBA,CAAEF,KAAK,EAAG;IACzB,OACC,CAAE,IAAAV,mBAAW,EAAEU,KAAM,CAAC,KACpBA,KAAK,CAACC,IAAI,KAAK,WAAW,IAAID,KAAK,CAACC,IAAI,KAAK,oBAAoB,CAAE;EAEvE;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;EACCE,qBAAqBA,CAAEH,KAAK,EAAEI,aAAa,EAAEC,SAAS,EAAEC,iBAAiB,EAAG;IAC3E,OAAS,CAAE,IAAAhB,mBAAW,EAAEU,KAAM,CAAC,IAC9B,CAAE,IAAAV,mBAAW,EAAEe,SAAU,CAAC,IAC1B,CAAE,IAAAf,mBAAW,EAAEgB,iBAAkB,CAAC,IAClC,CAAE,IAAAhB,mBAAW,EAAEc,aAAc,CAAC,IAC9BJ,KAAK,CAACC,IAAI,KAAK,WAAW,IAC1BG,aAAa,CAACH,IAAI,KAAK,UAAU,IACjCrC,gCAAgC,CAACa,IAAI,CAAE2B,aAAa,CAACb,GAAI,CAAC,IAC1Dc,SAAS,CAACJ,IAAI,KAAK,UAAU,IAC7BI,SAAS,CAACd,GAAG,CAACgB,IAAI,CAAC,CAAC,CAACC,MAAM,KAAK,CAAC,IACjCF,iBAAiB,CAACL,IAAI,KAAK,WAAW;EAExC;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;AACA;AACA;EACCQ,0BAA0BA,CAAET,KAAK,EAAEU,cAAc,EAAEC,eAAe,EAAG;IACpE;AACF;AACA;AACA;AACA;IACE,MAAMC,SAAS,GAAGZ,KAAK,CAACT,GAAG,CAACsB,SAAS,CAAE,CAAE,CAAC;;IAE1C;IACA,MAAMC,eAAe,GAAG,IAAI,CAACC,eAAe,CAAC,CAAC;IAC9C,IAAI,CAACC,QAAQ,CAAEF,eAAe,CAACG,SAAS,EAAEL,SAAU,CAAC;IACrD,MAAMM,cAAc,GAAG,IAAI,CAACC,sBAAsB,CAAEL,eAAe,CAACM,MAAM,EAAE,KAAM,CAAC;IAEnFF,cAAc,CAAE,CAAC,CAAE,GAAG,IAAA5B,mBAAW,EAAE4B,cAAc,CAAE,CAAC,CAAG,CAAC,GAAG,GAAG,GAAG,GAAG,GAAGA,cAAc,CAAE,CAAC,CAAE;;IAE1F;AACF;AACA;AACA;AACA;IACE,IAAK,IAAI,CAACrB,wBAAwB,CAAEqB,cAAc,CAAE,CAAC,CAAG,CAAC,EAAG;MAC3DR,cAAc,CAACW,IAAI,CAAEV,eAAgB,CAAC;MACtCA,eAAe,GAAG,EAAE;IACrB;IACAA,eAAe,IAAIO,cAAc,CAAE,CAAC,CAAE;IAEtC,IAAKA,cAAc,CAACV,MAAM,GAAG,CAAC,EAAG;MAChC;AACH;AACA;AACA;MACGE,cAAc,CAACW,IAAI,CAAEV,eAAgB,CAAC;MACtCA,eAAe,GAAG,EAAE;;MAEpB;MACAO,cAAc,CAACI,KAAK,CAAC,CAAC;MACtB;MACA,MAAMC,YAAY,GAAGL,cAAc,CAACM,GAAG,CAAC,CAAC;;MAEzC;MACAN,cAAc,CAACO,OAAO,CAAEC,QAAQ,IAAI;QACnChB,cAAc,CAACW,IAAI,CAAEK,QAAS,CAAC;MAChC,CAAE,CAAC;MAEH,MAAMC,gBAAgB,GAAG,IAAIhF,MAAM,CAAE,GAAG,GAAGF,QAAQ,GAAG,IAAI,CAACyB,qBAAqB,CAAC,CAAC,GAAG,IAAK,CAAC;;MAE3F;MACA,IAAKqD,YAAY,CAACK,KAAK,CAAED,gBAAiB,CAAC,EAAG;QAC7C;QACAjB,cAAc,CAACW,IAAI,CAAEE,YAAa,CAAC;MACpC,CAAC,MAAM;QACN;QACAZ,eAAe,GAAGY,YAAY;MAC/B;IACD;IACA,OAAO;MACNb,cAAc;MACdC;IACD,CAAC;EACF;;EAEA;AACD;AACA;AACA;AACA;EACCI,eAAeA,CAAA,EAAG;IACjB,MAAMc,sBAAsB,GAAG,IAAIlF,MAAM,CAAE,IAAI,GAAG,IAAI,CAACuB,qBAAqB,CAAC,CAAC,GAAG,IAAK,CAAC;IACvF,MAAM4D,aAAa,GAAG,IAAInF,MAAM,CAAE,KAAK,GAAGF,QAAQ,GAAG,IAAI,CAACyB,qBAAqB,CAAC,CAAC,GAAG,kBAAmB,CAAC;IAExG,MAAMkD,MAAM,GAAG,EAAE;IACjB,MAAMH,SAAS,GAAG,IAAAc,aAAI,EAAE,UAAU/B,KAAK,EAAG;MACzCoB,MAAM,CAACC,IAAI,CAAErB,KAAM,CAAC;IACrB,CAAE,CAAC;IAEHiB,SAAS,CAACe,OAAO,CAAEtF,aAAa,EAAE,WAAY,CAAC;IAC/CuE,SAAS,CAACe,OAAO,CAAEpF,uBAAuB,EAAE,2BAA4B,CAAC;IACzEqE,SAAS,CAACe,OAAO,CAAEnF,cAAc,EAAE,YAAa,CAAC;IACjDoE,SAAS,CAACe,OAAO,CAAElF,YAAY,EAAE,UAAW,CAAC;IAC7CmE,SAAS,CAACe,OAAO,CAAEjF,eAAe,EAAE,aAAc,CAAC;IACnDkE,SAAS,CAACe,OAAO,CAAEhF,aAAa,EAAE,WAAY,CAAC;IAC/CiE,SAAS,CAACe,OAAO,CAAEH,sBAAsB,EAAE,oBAAqB,CAAC;IACjEZ,SAAS,CAACe,OAAO,CAAEF,aAAa,EAAE,UAAW,CAAC;IAE9C,OAAO;MACNb,SAAS;MACTG;IACD,CAAC;EACF;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;EACCJ,QAAQA,CAAEC,SAAS,EAAElC,IAAI,EAAG;IAC3BkC,SAAS,CAACgB,MAAM,CAAElD,IAAK,CAAC;IAExB,IAAI;MACHkC,SAAS,CAACiB,GAAG,CAAC,CAAC;IAChB,CAAC,CAAC,OAAQ5F,CAAC,EAAG;MACb6F,OAAO,CAACC,KAAK,CAAE,sBAAsB,EAAE9F,CAAC,EAAEA,CAAC,CAAC+F,UAAW,CAAC;IACzD;EACD;;EAEA;AACD;AACA;AACA;AACA;EACCC,oBAAoBA,CAAE3B,eAAe,EAAG;IACvC,MAAM4B,oBAAoB,GAAG5B,eAAe,CAACiB,KAAK,CAAEtE,kBAAmB,CAAC;IAExE,IAAK,CAAEiF,oBAAoB,EAAG;MAC7B,OAAO,KAAK;IACb;IAEA,MAAMC,gBAAgB,GAAGD,oBAAoB,CAACf,GAAG,CAAC,CAAC;IACnD,OAAOb,eAAe,CAAC8B,QAAQ,CAAED,gBAAiB,CAAC;EACpD;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;AACA;AACA;EACCE,cAAcA,CAAEC,UAAU,EAAEC,SAAS,EAAG;IACvC,MAAMC,cAAc,GAAGF,UAAU,CAACpD,GAAG;IACrC,MAAMuD,aAAa,GAAGF,SAAS,CAACrD,GAAG;;IAEnC;IACA,MAAMwD,YAAY,GAAGF,cAAc,CAACjB,KAAK,CAAE/D,YAAa,CAAC,CAAE,CAAC,CAAE;IAC9D,MAAMmF,WAAW,GAAIF,aAAa,CAAClB,KAAK,CAAE/D,YAAa,CAAC,CAAE,CAAC,CAAE;;IAG7D;IACA,OAAOkF,YAAY,KAAKC,WAAW,IAAIlF,YAAY,CAACmF,QAAQ,CAAEF,YAAa,CAAC;EAC7E;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;AACA;EACC5B,sBAAsBA,CAAE+B,UAAU,EAAEC,aAAa,GAAG,IAAI,EAAG;IAC1D,IAAIzC,cAAc,GAAG,EAAE;MAAEC,eAAe,GAAG,EAAE;MAAEyC,iBAAiB;MAAEC,MAAM;;IAExE;IACA,GAAG;MACFA,MAAM,GAAG,KAAK;MACd,MAAMV,UAAU,GAAGO,UAAU,CAAE,CAAC,CAAE;MAClC,MAAMN,SAAS,GAAGM,UAAU,CAAEA,UAAU,CAAC1C,MAAM,GAAG,CAAC,CAAE;MAErD,IAAKmC,UAAU,IAAIC,SAAS,IAAID,UAAU,CAAC1C,IAAI,KAAK,YAAY,IAC/D2C,SAAS,CAAC3C,IAAI,KAAK,UAAU,IAAI,IAAI,CAACyC,cAAc,CAAEC,UAAU,EAAEC,SAAU,CAAC,EAAG;QAChFM,UAAU,GAAGA,UAAU,CAACI,KAAK,CAAE,CAAC,EAAEJ,UAAU,CAAC1C,MAAM,GAAG,CAAE,CAAC;QAEzD6C,MAAM,GAAG,IAAI;MACd;IACD,CAAC,QAASA,MAAM,IAAIH,UAAU,CAAC1C,MAAM,GAAG,CAAC;IAEzC0C,UAAU,CAACzB,OAAO,CAAE,CAAEzB,KAAK,EAAEuD,CAAC,KAAM;MACnC,IAAIC,eAAe,EAAEC,cAAc,EAAEC,eAAe;MACpD,MAAMrD,SAAS,GAAG6C,UAAU,CAAEK,CAAC,GAAG,CAAC,CAAE;MACrC,MAAMnD,aAAa,GAAG8C,UAAU,CAAEK,CAAC,GAAG,CAAC,CAAE;MACzC,MAAMjD,iBAAiB,GAAG4C,UAAU,CAAEK,CAAC,GAAG,CAAC,CAAE;MAC7CE,cAAc,GAAG,IAAI,CAACtE,oBAAoB,CAAE,CAAEkB,SAAS,EAAEC,iBAAiB,CAAG,CAAC;;MAE9E;MACAkD,eAAe,GAAGC,cAAc,CAACjD,MAAM,IAAI,CAAC;MAC5C4C,iBAAiB,GAAGI,eAAe,GAAGC,cAAc,CAAE,CAAC,CAAE,GAAG,EAAE;MAE9D,QAASzD,KAAK,CAACC,IAAI;QAClB,KAAK,YAAY;QACjB,KAAK,UAAU;UACd,IAAK,IAAI,CAAC1B,UAAU,CAAEyB,KAAK,CAACT,GAAI,CAAC,EAAG;YACnCmB,cAAc,CAACW,IAAI,CAAEV,eAAgB,CAAC;YACtCA,eAAe,GAAG,EAAE;UACrB,CAAC,MAAM;YACNA,eAAe,IAAIX,KAAK,CAACT,GAAG;UAC7B;UACA;QAED,KAAK,2BAA2B;UAC/BmE,eAAe,GAAG,IAAI,CAACjD,0BAA0B,CAAET,KAAK,EAAEU,cAAc,EAAEC,eAAgB,CAAC;UAC3FD,cAAc,GAAGgD,eAAe,CAAChD,cAAc;UAC/CC,eAAe,GAAG+C,eAAe,CAAC/C,eAAe;UACjD;QACD,KAAK,UAAU;UACdA,eAAe,IAAIX,KAAK,CAACT,GAAG;UAC5B;QACD,KAAK,oBAAoB;UACxBoB,eAAe,IAAIX,KAAK,CAACT,GAAG;;UAE5B;AACL;AACA;AACA;AACA;AACA;UACK,IAAK,CAAE,IAAAD,mBAAW,EAAEe,SAAU,CAAC,IAC9B,WAAW,KAAKA,SAAS,CAACJ,IAAI,IAC9B,oBAAoB,KAAKI,SAAS,CAACJ,IAAI,IACvC,IAAI,CAAC0D,iBAAiB,CAAEtD,SAAS,CAACd,GAAG,CAAE,CAAC,CAAG,CAAC,EAAG;YAC/C;YACA,IAAK,IAAI,CAACb,WAAW,CAAEsB,KAAK,CAACT,GAAI,CAAC,IAAIa,aAAa,IAAIA,aAAa,CAACb,GAAG,KAAK,GAAG,EAAG;cAClF;YACD;YACA;AACN;AACA;AACA;AACA;YACM,IAAK,IAAI,CAACb,WAAW,CAAEsB,KAAK,CAACT,GAAI,CAAC,IAAIS,KAAK,CAACT,GAAG,KAAK,GAAG,EAAG;cACzDoB,eAAe,GAAG,IAAI,CAACiD,gBAAgB,CAAEJ,eAAe,EACvDJ,iBAAiB,EACjBK,cAAc,EACdpD,SAAS,EACTK,cAAc,EACdC,eAAgB,CAAC;YACnB,CAAC,MAAM;cACND,cAAc,CAACW,IAAI,CAAEV,eAAgB,CAAC;cACtCA,eAAe,GAAG,EAAE;YACrB;UACD;UACA;QAED,KAAK,WAAW;UACfA,eAAe,IAAIX,KAAK,CAACT,GAAG;UAC5BkE,cAAc,GAAG,IAAI,CAACtE,oBAAoB,CAAE,CAAEkB,SAAS,EAAEC,iBAAiB,CAAG,CAAC;;UAE9E;UACAkD,eAAe,GAAGC,cAAc,CAACjD,MAAM,IAAI,CAAC;UAC5C4C,iBAAiB,GAAGI,eAAe,GAAGC,cAAc,CAAE,CAAC,CAAE,GAAG,EAAE;;UAE9D;UACA,IAAK,IAAI,CAACnB,oBAAoB,CAAE3B,eAAgB,CAAC,EAAG;YACnD;UACD;;UAEA;UACA,IAAK6C,eAAe,IAAI,IAAI,CAACrF,QAAQ,CAAEsF,cAAc,CAAE,CAAC,CAAG,CAAC,EAAG;YAC9D;UACD;;UAEA;UACA,IAAK,IAAI,CAACtD,qBAAqB,CAAEH,KAAK,EAAEI,aAAa,EAAEC,SAAS,EAAEC,iBAAkB,CAAC,EAAG;YACvF;UACD;;UAEA;UACA;UACA;UACA,IAAK,IAAI,CAAC1B,kBAAkB,CAAE+B,eAAgB,CAAC,EAAG;YACjD;UACD;;UAEA;AACL;AACA;AACA;AACA;UACKA,eAAe,GAAG,IAAI,CAACiD,gBAAgB,CAAEJ,eAAe,EACvDJ,iBAAiB,EACjBK,cAAc,EACdpD,SAAS,EACTK,cAAc,EACdC,eAAgB,CAAC;UAElB;QAED,KAAK,aAAa;UACjBA,eAAe,IAAIX,KAAK,CAACT,GAAG;UAC5B;QAED,KAAK,WAAW;UACfoB,eAAe,IAAIX,KAAK,CAACT,GAAG;UAE5BkE,cAAc,GAAG,IAAI,CAACtE,oBAAoB,CAAE,CAAEkB,SAAS,EAAEC,iBAAiB,CAAG,CAAC;;UAE9E;UACAkD,eAAe,GAAGC,cAAc,CAACjD,MAAM,IAAI,CAAC;UAC5C4C,iBAAiB,GAAGI,eAAe,GAAGC,cAAc,CAAE,CAAC,CAAE,GAAG,EAAE;;UAE9D;AACL;AACA;AACA;UACK,IACCD,eAAe,IAAI,IAAI,CAACrF,QAAQ,CAAEsF,cAAc,CAAE,CAAC,CAAG,CAAC,IACrD,IAAI,CAACvD,gBAAgB,CAAEE,aAAc,CAAC,IACrC,EAAI,IAAI,CAACP,wBAAwB,CAAEuD,iBAAkB,CAAC,IAAI,IAAI,CAACrD,eAAe,CAAEM,SAAU,CAAC,CAAM,EACnG;YACD;UACD;;UAEA;AACL;AACA;AACA;UACK,IACC,IAAI,CAACH,gBAAgB,CAAEE,aAAc,CAAC,KACpC,IAAI,CAACL,eAAe,CAAEM,SAAU,CAAC,IAAI,IAAI,CAACR,wBAAwB,CAAEuD,iBAAkB,CAAC,CAAE,EAC1F;YACD1C,cAAc,CAACW,IAAI,CAAEV,eAAgB,CAAC;YACtCA,eAAe,GAAG,EAAE;UACrB;UACA;MACF;IACD,CAAE,CAAC;IAEH,IAAK,EAAE,KAAKA,eAAe,EAAG;MAC7BD,cAAc,CAACW,IAAI,CAAEV,eAAgB,CAAC;IACvC;IAEA,IAAKwC,aAAa,EAAG;MACpBzC,cAAc,GAAG,IAAAvD,WAAG,EAAEuD,cAAc,EAAE,UAAUgB,QAAQ,EAAG;QAC1D,OAAOA,QAAQ,CAACnB,IAAI,CAAC,CAAC;MACvB,CAAE,CAAC;IACJ;IAEA,OAAOG,cAAc;EACtB;;EAEA;AACD;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;EACCkD,gBAAgBA,CAAEJ,eAAe,EAAEJ,iBAAiB,EAAEK,cAAc,EAAEpD,SAAS,EAAEK,cAAc,EAAEC,eAAe,EAAG;IAClH,IAAO6C,eAAe,IAAI,IAAI,CAAC3D,wBAAwB,CAAEuD,iBAAkB,CAAC,IAAI,IAAI,CAACO,iBAAiB,CAAEF,cAAc,CAAE,CAAC,CAAG,CAAC,IAC5H,IAAI,CAAC1D,eAAe,CAAEM,SAAU,CAAC,EAAG;MACpCK,cAAc,CAACW,IAAI,CAAEV,eAAgB,CAAC;MACtCA,eAAe,GAAG,EAAE;IACrB;IACA,OAAOA,eAAe;EACvB;;EAEA;AACD;AACA;AACA;AACA;AACA;EACCgD,iBAAiBA,CAAEvF,SAAS,EAAG;IAC9B,OAAO,IAAI,CAACK,IAAI,CAAEL,SAAU,CAAC;EAC9B;AACD;AAACyF,OAAA,CAAArH,OAAA,GAAAuB,iBAAA","ignoreList":[]}