sec-edgar-api 0.0.5 → 0.0.7

package/build/services/DocumentParser/XMLNode/TableNode.js

@@ -0,0 +1,132 @@
+"use strict";
+var __extends = (this && this.__extends) || (function () {
+    var extendStatics = function (d, b) {
+        extendStatics = Object.setPrototypeOf ||
+            ({ __proto__: [] } instanceof Array && function (d, b) { d.__proto__ = b; }) ||
+            function (d, b) { for (var p in b) if (Object.prototype.hasOwnProperty.call(b, p)) d[p] = b[p]; };
+        return extendStatics(d, b);
+    };
+    return function (d, b) {
+        if (typeof b !== "function" && b !== null)
+            throw new TypeError("Class extends value " + String(b) + " is not a constructor or null");
+        extendStatics(d, b);
+        function __() { this.constructor = d; }
+        d.prototype = b === null ? Object.create(b) : (__.prototype = b.prototype, new __());
+    };
+})();
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.TableNode = void 0;
+var XMLNode_1 = require("./XMLNode");
+var TableNode = /** @class */ (function (_super) {
+    __extends(TableNode, _super);
+    function TableNode() {
+        var _this = _super !== null && _super.apply(this, arguments) || this;
+        _this.title = null;
+        _this.headerRow = null;
+        return _this;
+    }
+    TableNode.prototype.getTitle = function () {
+        var _a;
+        return (_a = this.title) !== null && _a !== void 0 ? _a : '';
+    };
+    TableNode.prototype.setTitle = function (title) {
+        this.title = title;
+    };
+    TableNode.prototype.getChildren = function () {
+        return _super.prototype.getChildren.call(this);
+    };
+    TableNode.prototype.removeTopChild = function () {
+        this.removeChild(this.getChildren()[0]);
+    };
+    TableNode.prototype.removeEmptyTopRows = function () {
+        var _a;
+        while ((_a = this.getChildren()[0]) === null || _a === void 0 ? void 0 : _a.getIsEmpty()) {
+            this.removeChild(this.getChildren()[0]);
+        }
+    };
+    TableNode.prototype.prependChild = function (node) {
+        var prevTopChild = this.getChildren()[0];
+        this.getChildren().unshift(node);
+        if (node.getParent() !== this)
+            node.setParent(this);
+        prevTopChild === null || prevTopChild === void 0 ? void 0 : prevTopChild.setPreviousSibling(node);
+        var colArrTop = [];
+        var colArrBottom = [];
+        node.getChildren().forEach(function (col) {
+            colArrTop.push(col);
+            Array.from({ length: col.getColSpan() - 1 }).forEach(function () { return colArrTop.push(col); });
+        });
+        prevTopChild === null || prevTopChild === void 0 ? void 0 : prevTopChild.getChildren().forEach(function (col, i) {
+            colArrBottom.push(col);
+            Array.from({ length: col.getColSpan() - 1 }).forEach(function () { return colArrBottom.push(col); });
+            if (!col.getTopSiblings().includes(colArrTop[i])) {
+                col.addTopSibling(colArrTop[i]);
+            }
+        });
+    };
+    TableNode.prototype.toArray = function (parseValues) {
+        if (parseValues === void 0) { parseValues = true; }
+        return this.getChildren().map(function (row) { return row.toArray(parseValues); });
+    };
+    TableNode.prototype.setHeaderRow = function (row) {
+        this.headerRow = row;
+    };
+    TableNode.prototype.getHeaderRowIndex = function () {
+        var rows = this.getChildren();
+        // assume body index starts with row that has a non-bold number in it.
+        var bodyIndex = rows.findIndex(function (row) {
+            return row.getChildren().some(function (col) { return typeof col.parseValue() === 'number' && !col.getText().includes('}}'); });
+        });
+        var getRowData = function (row) { return row.getChildren().map(function (col) { return col.parseValue(); }); };
+        // if the header row is only one value, or empty, it's likely a label in the table body, so keep moving up.
+        var headerIndex = bodyIndex - 1;
+        while (rows[headerIndex] && getRowData(rows[headerIndex]).filter(Boolean).length <= 1 && headerIndex >= 0) {
+            headerIndex--;
+        }
+        return headerIndex >= 0 ? headerIndex : null;
+    };
+    TableNode.prototype.mergeHeader = function (removeMergedChildren) {
+        var _a, _b, _c;
+        if (removeMergedChildren === void 0) { removeMergedChildren = true; }
+        var headerRowIndex = (_a = this.getHeaderRowIndex()) !== null && _a !== void 0 ? _a : -1;
+        var headerRow = (_b = this.getChildren()[headerRowIndex]) !== null && _b !== void 0 ? _b : null;
+        if (!headerRow)
+            return;
+        var table = headerRow.toTable(false);
+        var headerRowCols = headerRow.getChildren();
+        // start from the row above the header row to merge.
+        for (var rowIndex = headerRowIndex - 1; rowIndex >= 0; rowIndex--) {
+            var curRow = table[rowIndex];
+            // go through each header column to merge with the one above.
+            for (var colIndex = 0; colIndex < curRow.length; colIndex++) {
+                if (table[headerRowIndex][colIndex] === null)
+                    continue;
+                // if prev header col is empty, get nearest to the left.
+                var colIndexCur = colIndex;
+                while (!curRow[colIndexCur] && colIndexCur >= 0) {
+                    colIndexCur--;
+                }
+                // if the value is empty, continue.
+                var colValue = (_c = curRow[colIndexCur]) !== null && _c !== void 0 ? _c : null;
+                if (!colValue || !this.parseValue("".concat(colValue)))
+                    continue;
+                headerRowCols[colIndex].setText("".concat(colValue, " ").concat(headerRowCols[colIndex].getText()).trim());
+            }
+        }
+        if (removeMergedChildren) {
+            while (this.getChildren()[0] !== headerRow && this.getChildren().length > 0) {
+                this.removeTopChild();
+            }
+        }
+        return headerRow;
+    };
+    /**
+     * If header row is not set, this will try to find it.
+     */
+    TableNode.prototype.getHeaderRow = function () {
+        var _a, _b, _c;
+        return (_c = (_a = this.headerRow) !== null && _a !== void 0 ? _a : this.getChildren()[(_b = this.getHeaderRowIndex()) !== null && _b !== void 0 ? _b : -1]) !== null && _c !== void 0 ? _c : null;
+    };
+    return TableNode;
+}(XMLNode_1.XMLNode));
+exports.TableNode = TableNode;

package/build/services/DocumentParser/XMLNode/XMLNode.d.ts

@@ -0,0 +1,37 @@
+interface XMLNodeArgs {
+    path?: string;
+    attributesStr?: string;
+}
+interface GetSiblingsParams {
+    dir: 'previous' | 'next';
+    stopAtType?: new () => XMLNode;
+    includeStopAtType?: boolean;
+}
+export declare class XMLNode {
+    private parent;
+    private previousSibling;
+    private nextSibling;
+    private children;
+    private text;
+    private attributesStr;
+    private path;
+    constructor(args?: XMLNodeArgs);
+    getSiblings(params: GetSiblingsParams): XMLNode[];
+    extractBold(str?: string): string;
+    parseValue(str?: string | null): string | number | null;
+    setPreviousSibling(node: XMLNode | null): void;
+    setNextSibling(node: XMLNode | null): void;
+    removeChild(node: XMLNode): void;
+    setParent(node: XMLNode | null): void;
+    getNextSibling(): XMLNode | null;
+    getPreviousSibling(): XMLNode | null;
+    getParent(): XMLNode | null;
+    getAttributes(): Record<string, string>;
+    getAttributesStr(): string;
+    getChildren(): XMLNode[];
+    addChild(node: XMLNode): void;
+    getText(): string;
+    setText(text: string): void;
+    getPath(): string;
+}
+export {};

package/build/services/DocumentParser/XMLNode/XMLNode.js

@@ -0,0 +1,154 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.XMLNode = void 0;
+var XMLNode = /** @class */ (function () {
+    function XMLNode(args) {
+        this.parent = null;
+        this.previousSibling = null;
+        this.nextSibling = null;
+        this.children = [];
+        this.text = '';
+        this.attributesStr = '';
+        this.path = '';
+        var _a = args !== null && args !== void 0 ? args : {}, _b = _a.attributesStr, attributesStr = _b === void 0 ? '' : _b, _c = _a.path, path = _c === void 0 ? '' : _c;
+        var pathParts = path.split('.').filter(function (str) { return !str.includes(':'); });
+        this.attributesStr = attributesStr;
+        this.path = pathParts.join('.');
+    }
+    XMLNode.prototype.getSiblings = function (params) {
+        var _a = params.stopAtType, stopAtType = _a === void 0 ? XMLNode : _a, dir = params.dir, _b = params.includeStopAtType, includeStopAtType = _b === void 0 ? false : _b;
+        var isPrevious = dir === 'previous';
+        var siblings = [];
+        var nextSibling = isPrevious ? this.getPreviousSibling() : this.getNextSibling();
+        var i = 0;
+        while (nextSibling && !(nextSibling instanceof stopAtType)) {
+            siblings.push(nextSibling);
+            nextSibling = isPrevious
+                ? nextSibling.getPreviousSibling()
+                : nextSibling.getNextSibling();
+            i++;
+            if (i > 1000)
+                throw new Error('infinite loop');
+        }
+        if (includeStopAtType && nextSibling instanceof stopAtType) {
+            siblings.push(nextSibling);
+        }
+        return siblings;
+    };
+    XMLNode.prototype.extractBold = function (str) {
+        var _a, _b;
+        if (str === void 0) { str = this.text; }
+        var boldText = (_b = (_a = str
+            .replace(/\n/g, '')
+            .replace(/\}\}\{\{/g, ' | ')
+            .replace(/ | /g, ' ')
+            .replace(/–/g, '- ')
+            .match(/(?<=\{\{).*?(?=\}\})/g)) === null || _a === void 0 ? void 0 : _a.join(' || ')) !== null && _b !== void 0 ? _b : '';
+        return boldText.replace(/\s+/g, ' ').trim();
+    };
+    XMLNode.prototype.parseValue = function (str) {
+        if (str === void 0) { str = this.text; }
+        if (str === null)
+            return null;
+        var text = str
+            .replace(/\n|&#160;|&nbsp;/g, ' ')
+            .replace(/&#174;/g, '')
+            .replace(/&#8211;|&#8212;|&#x2014;/g, '-')
+            .replace(/&#8217;|&#8220;|&#8221;/g, "'")
+            .replace(/\}\}\{\{/g, ' ')
+            .replace(/\{\{|\}\}/g, '')
+            .replace(/\s+/, ' ')
+            .trim();
+        if (str.replace(/&#8211;|&#8212;|&#x2014;/g, '-') === '-')
+            return '-';
+        if (text === '')
+            return null;
+        var colNum = text.replace(/,|\(|\)|\%/g, '').trim();
+        if (colNum === '-' || colNum === '$')
+            return null;
+        colNum = colNum.replace(/\-|\$/g, '');
+        if (!isNaN(Number(colNum))) {
+            if (text.includes('%'))
+                return text.replace(/[^a-zA-Z\d\s:]/g, '') === '' ? null : text;
+            return text.includes('(') || text.includes('-') ? Number(colNum) * -1 : Number(colNum);
+        }
+        return text;
+    };
+    XMLNode.prototype.setPreviousSibling = function (node) {
+        var prevPreviousSibling = this.previousSibling;
+        this.previousSibling = node;
+        if ((prevPreviousSibling === null || prevPreviousSibling === void 0 ? void 0 : prevPreviousSibling.getNextSibling()) === this) {
+            prevPreviousSibling.setNextSibling(null);
+        }
+        if ((node === null || node === void 0 ? void 0 : node.getNextSibling()) !== this) {
+            node === null || node === void 0 ? void 0 : node.setNextSibling(this);
+        }
+    };
+    XMLNode.prototype.setNextSibling = function (node) {
+        var prevNextSibling = this.nextSibling;
+        this.nextSibling = node;
+        if ((prevNextSibling === null || prevNextSibling === void 0 ? void 0 : prevNextSibling.getPreviousSibling()) === this) {
+            prevNextSibling.setPreviousSibling(null);
+        }
+        if ((node === null || node === void 0 ? void 0 : node.getPreviousSibling()) !== this) {
+            node === null || node === void 0 ? void 0 : node.setPreviousSibling(this);
+        }
+    };
+    XMLNode.prototype.removeChild = function (node) {
+        this.children.splice(this.children.indexOf(node), 1);
+        if (node.getParent() === this) {
+            node.setParent(null);
+        }
+    };
+    XMLNode.prototype.setParent = function (node) {
+        var prevParent = this.parent;
+        this.parent = node;
+        if (prevParent === null || prevParent === void 0 ? void 0 : prevParent.getChildren().includes(this)) {
+            prevParent.removeChild(this);
+        }
+        if (!(node === null || node === void 0 ? void 0 : node.getChildren().includes(this))) {
+            node === null || node === void 0 ? void 0 : node.addChild(this);
+        }
+    };
+    XMLNode.prototype.getNextSibling = function () {
+        return this.nextSibling;
+    };
+    XMLNode.prototype.getPreviousSibling = function () {
+        return this.previousSibling;
+    };
+    XMLNode.prototype.getParent = function () {
+        return this.parent;
+    };
+    XMLNode.prototype.getAttributes = function () {
+        var _a;
+        var attributesObj = {};
+        (_a = this.attributesStr.match(/(\w+)=("[^"]*")/g)) === null || _a === void 0 ? void 0 : _a.forEach(function (attributeStr) {
+            var _a = attributeStr.split('='), key = _a[0], value = _a[1];
+            attributesObj[key.toLowerCase()] = value.replace(/"/g, '');
+        });
+        return attributesObj;
+    };
+    XMLNode.prototype.getAttributesStr = function () {
+        return this.attributesStr;
+    };
+    XMLNode.prototype.getChildren = function () {
+        return this.children;
+    };
+    XMLNode.prototype.addChild = function (node) {
+        this.children.push(node);
+        if (node.getParent() !== this) {
+            node.setParent(this);
+        }
+    };
+    XMLNode.prototype.getText = function () {
+        return this.text;
+    };
+    XMLNode.prototype.setText = function (text) {
+        this.text = text;
+    };
+    XMLNode.prototype.getPath = function () {
+        return this.path;
+    };
+    return XMLNode;
+}());
+exports.XMLNode = XMLNode;

package/build/services/DocumentParser/XMLParser.d.ts

@@ -0,0 +1,40 @@
+import { DocumentNode } from './XMLNode/DocumentNode';
+interface OnCharacterData {
+    char: string;
+    index: number;
+    path: string;
+    pathOccurrenceCount: number;
+    attributesStr: string;
+}
+interface ParseTableNodesParams {
+    xml: string;
+}
+interface Parse2Params {
+    xml: string;
+    onCharacter?: (data: OnCharacterData) => void;
+    onOpenTag?: (data: OnCharacterData) => void;
+    onCloseTag?: (data: OnCharacterData) => void;
+}
+interface IterateTablesParams {
+    xml: string;
+    parentPath?: string;
+    trimSpaces?: boolean;
+    onCharacter?: (data: OnCharacterData & {
+        textMap: Map<string, string>;
+    }) => void;
+    onOpenTag?: (data: OnCharacterData & {
+        textMap: Map<string, string>;
+    }) => void;
+    onCloseTag?: (data: OnCharacterData & {
+        textMap: Map<string, string>;
+    }) => void;
+}
+export default class XMLParser {
+    iterateXML(params: Parse2Params): string[];
+    /**
+     * Returns text in each table cell mapped by `${table}.${row}.${col}`
+     */
+    getTableTextMap(params: IterateTablesParams): Map<string, string>;
+    getDocumentNode(params: ParseTableNodesParams): DocumentNode;
+}
+export {};

package/build/services/DocumentParser/XMLParser.js

@@ -0,0 +1,306 @@
+"use strict";
+var __assign = (this && this.__assign) || function () {
+    __assign = Object.assign || function(t) {
+        for (var s, i = 1, n = arguments.length; i < n; i++) {
+            s = arguments[i];
+            for (var p in s) if (Object.prototype.hasOwnProperty.call(s, p))
+                t[p] = s[p];
+        }
+        return t;
+    };
+    return __assign.apply(this, arguments);
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+var ColNode_1 = require("./XMLNode/ColNode");
+var DocumentNode_1 = require("./XMLNode/DocumentNode");
+var HRNode_1 = require("./XMLNode/HRNode");
+var NonTableNode_1 = require("./XMLNode/NonTableNode");
+var RowNode_1 = require("./XMLNode/RowNode");
+var TableNode_1 = require("./XMLNode/TableNode");
+var XMLParser = /** @class */ (function () {
+    function XMLParser() {
+    }
+    XMLParser.prototype.iterateXML = function (params) {
+        var _a, _b, _c;
+        var xml = params.xml, onCharacter = params.onCharacter, onCloseTag = params.onCloseTag, onOpenTag = params.onOpenTag;
+        var selfEnclosingTags = new Set([
+            'filename',
+            'description',
+            'br',
+            'meta',
+            'link',
+            'img',
+            'input',
+            'hr',
+            'area',
+            'base',
+            'col',
+            'command',
+            'embed',
+            'keygen',
+            'param',
+            'source',
+            'track',
+            'wbr',
+        ]);
+        var spaceChars = new Set(['\n', '\r', '\t', ' ']);
+        var pathOccurrenceCountMap = new Map();
+        var curPath = '';
+        var curTag = '';
+        var curAttributes = '';
+        var didStart = false;
+        var pathsArr = [];
+        for (var i = 0; i < xml.length; i++) {
+            var char = xml[i];
+            var isOpenTag = char === '<' && xml[i + 1] !== '/' && xml[i + 1] !== '?' && xml[i + 1] !== '!';
+            var isCloseTag = char === '<' && xml[i + 1] === '/';
+            var onCharacterData = {
+                char: char,
+                index: i,
+                path: curPath,
+                pathOccurrenceCount: (_a = pathOccurrenceCountMap.get(curPath)) !== null && _a !== void 0 ? _a : 0,
+                attributesStr: curAttributes,
+            };
+            if (isOpenTag) {
+                var didEndTagName = false;
+                var j = 0;
+                didStart = true;
+                i++;
+                while (xml[i] !== '>') {
+                    didEndTagName = didEndTagName || spaceChars.has(xml[i]) || xml[i] === '/';
+                    if (!didEndTagName) {
+                        curTag += xml[i].toLowerCase();
+                    }
+                    else if (xml[i] !== '/') {
+                        curAttributes += xml[i];
+                    }
+                    i++;
+                    j++;
+                    if (j > 1000000) {
+                        throw new Error('too many iterations');
+                    }
+                }
+                var pathNew = "".concat(curPath).concat(curPath.length > 0 ? '.' : '').concat(curTag).toLowerCase();
+                var countBefore = (_b = pathOccurrenceCountMap.get(pathNew)) !== null && _b !== void 0 ? _b : 0;
+                var pathOccurrenceCount = (_c = pathOccurrenceCountMap.set(pathNew, countBefore + 1).get(pathNew)) !== null && _c !== void 0 ? _c : 0;
+                onCharacterData.path = pathNew;
+                onCharacterData.pathOccurrenceCount = pathOccurrenceCount;
+                onCharacterData.attributesStr = curAttributes;
+                pathsArr.push(pathNew);
+                onOpenTag === null || onOpenTag === void 0 ? void 0 : onOpenTag(onCharacterData);
+                if (selfEnclosingTags.has(curTag)) {
+                    onCloseTag === null || onCloseTag === void 0 ? void 0 : onCloseTag(onCharacterData);
+                }
+                else {
+                    curPath = pathNew;
+                }
+                curTag = '';
+            }
+            else if (isCloseTag) {
+                while (xml[i] !== '>') {
+                    i++;
+                }
+                onCloseTag === null || onCloseTag === void 0 ? void 0 : onCloseTag(onCharacterData);
+                curPath = curPath.slice(0, curPath.lastIndexOf('.'));
+                curAttributes = '';
+            }
+            else if (didStart) {
+                onCharacter === null || onCharacter === void 0 ? void 0 : onCharacter(onCharacterData);
+            }
+        }
+        return pathsArr;
+    };
+    /**
+     * Returns text in each table cell mapped by `${table}.${row}.${col}`
+     */
+    XMLParser.prototype.getTableTextMap = function (params) {
+        var xml = params.xml, parentPath = params.parentPath, onCharacter = params.onCharacter, onCloseTag = params.onCloseTag, onOpenTag = params.onOpenTag, _a = params.trimSpaces, trimSpaces = _a === void 0 ? true : _a;
+        var rowPaths = new Set([
+            "".concat(parentPath, ".table.tbody.tr"),
+            "".concat(parentPath, ".table.thead.tr"),
+            "".concat(parentPath, ".table.tfoot.tr"),
+            "".concat(parentPath, ".table.tr"),
+        ]);
+        var colPaths = new Set([
+            "".concat(parentPath, ".table.tbody.tr.td"),
+            "".concat(parentPath, ".table.thead.tr.td"),
+            "".concat(parentPath, ".table.tfoot.tr.td"),
+            "".concat(parentPath, ".table.tr.td"),
+            "".concat(parentPath, ".table.tbody.tr.th"),
+            "".concat(parentPath, ".table.thead.tr.th"),
+            "".concat(parentPath, ".table.tfoot.tr.th"),
+            "".concat(parentPath, ".table.tr.th"),
+        ]);
+        var table = 0;
+        var row = 0;
+        var col = 0;
+        var textByColKey = new Map();
+        var spaceChars = new Set(['\n', '\r', '\t']);
+        this.iterateXML({
+            xml: xml,
+            onOpenTag: function (data) {
+                var _a;
+                var path = data.path;
+                var colKey = "".concat(table, ".").concat(row, ".").concat(col);
+                var textCur = (_a = textByColKey.get(colKey)) !== null && _a !== void 0 ? _a : '';
+                var pathLower = path.toLowerCase();
+                if (textCur.trim().length === 0 && col === 0) {
+                    textByColKey.delete(colKey);
+                }
+                var isTable = parentPath ? pathLower === "".concat(parentPath, ".table") : pathLower.endsWith('table');
+                var isRow = parentPath ? rowPaths.has(pathLower) : pathLower.endsWith('tr');
+                var isCol = parentPath
+                    ? colPaths.has(pathLower)
+                    : pathLower.endsWith('td') || pathLower.endsWith('th');
+                if (isTable) {
+                    table++;
+                    col = 0;
+                    row = 0;
+                }
+                else if (isRow) {
+                    row++;
+                    col = 0;
+                }
+                else if (isCol) {
+                    col++;
+                }
+                onOpenTag === null || onOpenTag === void 0 ? void 0 : onOpenTag(__assign(__assign({}, data), { textMap: textByColKey }));
+            },
+            onCharacter: function (data) {
+                var _a;
+                var char = spaceChars.has(data.char) ? ' ' : data.char;
+                var colKey = "".concat(table, ".").concat(row, ".").concat(col);
+                var textCur = (_a = textByColKey.get(colKey)) !== null && _a !== void 0 ? _a : '';
+                if (!(trimSpaces && char === ' ' && textCur.endsWith(' '))) {
+                    textByColKey.set(colKey, "".concat(textCur).concat(char));
+                }
+                onCharacter === null || onCharacter === void 0 ? void 0 : onCharacter(__assign(__assign({}, data), { textMap: textByColKey }));
+            },
+            onCloseTag: function (data) {
+                var _a;
+                var colKey = "".concat(table, ".").concat(row, ".").concat(col);
+                var textCur = (_a = textByColKey.get(colKey)) !== null && _a !== void 0 ? _a : '';
+                if (textCur.trim().length === 0 && col === 0) {
+                    textByColKey.delete(colKey);
+                }
+                else if (!textCur.endsWith(' ')) {
+                    textByColKey.set(colKey, "".concat(textCur, " "));
+                }
+                onCloseTag === null || onCloseTag === void 0 ? void 0 : onCloseTag(__assign(__assign({}, data), { textMap: textByColKey }));
+            },
+        });
+        return textByColKey;
+    };
+    XMLParser.prototype.getDocumentNode = function (params) {
+        var xml = params.xml;
+        var rowsArr = [];
+        var colsArr = [];
+        var documentNode = new DocumentNode_1.DocumentNode();
+        var curNode = null;
+        var prevRowCols = [];
+        var curRowCols = [];
+        var isBold = false;
+        var boldPath = null;
+        var pushColToRow = function (col) {
+            var _a;
+            var colSpan = col.getColSpan();
+            var colIndex = curRowCols.length;
+            col.setIndex(colIndex);
+            for (var i = 0; i < colSpan; i++) {
+                curRowCols.push(col);
+            }
+            var topSibling = (_a = prevRowCols[colIndex]) !== null && _a !== void 0 ? _a : null;
+            topSibling === null || topSibling === void 0 ? void 0 : topSibling.addBottomSibling(col);
+        };
+        this.iterateXML({
+            xml: xml,
+            onCloseTag: function () {
+                var _a;
+                if ((curNode === null || curNode === void 0 ? void 0 : curNode.getPath()) === boldPath) {
+                    curNode === null || curNode === void 0 ? void 0 : curNode.setText("".concat((_a = curNode === null || curNode === void 0 ? void 0 : curNode.getText()) !== null && _a !== void 0 ? _a : '', "}}"));
+                    boldPath = null;
+                }
+            },
+            onCharacter: function (_a) {
+                var _b;
+                var char = _a.char;
+                curNode === null || curNode === void 0 ? void 0 : curNode.setText(((_b = curNode === null || curNode === void 0 ? void 0 : curNode.getText()) !== null && _b !== void 0 ? _b : '') + char);
+            },
+            onOpenTag: function (_a) {
+                var _b;
+                var path = _a.path, attributesStr = _a.attributesStr;
+                // skip nested tables
+                if (path.split('.').reduce(function (acc, cur) { return (cur === 'table' ? acc + 1 : acc); }, 0) > 1)
+                    return;
+                var tag = path.split('.').pop();
+                var isInTable = path.includes('table');
+                var topLevelNodes = documentNode.getChildren();
+                var prevTopLevelNode = topLevelNodes[topLevelNodes.length - 1];
+                var wasHorizontalLine = prevTopLevelNode instanceof HRNode_1.HRNode;
+                var wasNonTableNode = prevTopLevelNode instanceof NonTableNode_1.NonTableNode;
+                var wasBold = isBold;
+                var attributesLower = attributesStr.toLowerCase().replace(/\s/g, '');
+                isBold =
+                    tag === 'b' ||
+                        tag === 'strong' ||
+                        attributesLower.includes('font-weight:bold') ||
+                        attributesLower.includes('font-weight:700') ||
+                        attributesLower.includes('font-weight:800') ||
+                        attributesLower.includes('font-weight:900');
+                if (!isInTable) {
+                    prevRowCols = [];
+                    curRowCols = [];
+                }
+                if (tag === 'hr' && !isInTable) {
+                    var hr = new HRNode_1.HRNode({ attributesStr: attributesStr, path: path });
+                    hr.setPreviousSibling(prevTopLevelNode !== null && prevTopLevelNode !== void 0 ? prevTopLevelNode : null);
+                    topLevelNodes.push(hr);
+                    curNode = hr;
+                }
+                else if (tag === 'table') {
+                    var table = new TableNode_1.TableNode({ attributesStr: attributesStr, path: path });
+                    table.setPreviousSibling(prevTopLevelNode !== null && prevTopLevelNode !== void 0 ? prevTopLevelNode : null);
+                    topLevelNodes.push(table);
+                    curNode = table;
+                }
+                else if (tag === 'tr') {
+                    var row = new RowNode_1.RowNode({ attributesStr: attributesStr, path: path });
+                    var prevRow = rowsArr[rowsArr.length - 1];
+                    row.setParent(prevTopLevelNode);
+                    row.setPreviousSibling((prevRow === null || prevRow === void 0 ? void 0 : prevRow.getParent()) === row.getParent() ? prevRow : null);
+                    rowsArr.push(row);
+                    prevRowCols = curRowCols;
+                    curRowCols = [];
+                    curNode = row;
+                }
+                else if (tag === 'td' || tag === 'th') {
+                    var col = new ColNode_1.ColNode({ attributesStr: attributesStr, path: path });
+                    var prevCol = colsArr[colsArr.length - 1];
+                    col.setParent(rowsArr[rowsArr.length - 1]);
+                    col.setPreviousSibling((prevCol === null || prevCol === void 0 ? void 0 : prevCol.getParent()) === col.getParent() ? prevCol : null);
+                    colsArr.push(col);
+                    pushColToRow(col);
+                    curNode = col;
+                }
+                else if ((!isInTable && !wasNonTableNode) || (wasHorizontalLine && tag !== 'hr')) {
+                    var node = new NonTableNode_1.NonTableNode({ attributesStr: attributesStr, path: path });
+                    node.setPreviousSibling(prevTopLevelNode !== null && prevTopLevelNode !== void 0 ? prevTopLevelNode : null);
+                    topLevelNodes.push(node);
+                    curNode = node;
+                }
+                else if (curNode && !curNode.getText().endsWith('\n')) {
+                    curNode.setText("".concat(curNode.getText().trim(), "\n"));
+                }
+                if (isBold && !wasBold && !(curNode === null || curNode === void 0 ? void 0 : curNode.getText().endsWith('{{'))) {
+                    curNode === null || curNode === void 0 ? void 0 : curNode.setText("".concat(curNode === null || curNode === void 0 ? void 0 : curNode.getText().trim(), "{{"));
+                }
+                if (isBold) {
+                    boldPath = (_b = curNode === null || curNode === void 0 ? void 0 : curNode.getPath()) !== null && _b !== void 0 ? _b : null;
+                }
+            },
+        });
+        return documentNode;
+    };
+    return XMLParser;
+}());
+exports.default = XMLParser;

package/build/services/DocumentParser/index.d.ts

@@ -0,0 +1,2 @@
+import DocumentParser from './DocumentParser';
+export default DocumentParser;

package/build/services/DocumentParser/index.js

@@ -0,0 +1,4 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+var DocumentParser_1 = require("./DocumentParser");
+exports.default = DocumentParser_1.default;