officeparser 6.0.7 → 6.1.1

package/dist/parsers/OpenOfficeParser.js

@@ -23,12 +23,12 @@
  */
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parseOpenOffice = void 0;
-const chartUtils_1 = require("../utils/chartUtils");
-const errorUtils_1 = require("../utils/errorUtils");
-const imageUtils_1 = require("../utils/imageUtils");
-const ocrUtils_1 = require("../utils/ocrUtils");
-const xmlUtils_1 = require("../utils/xmlUtils");
-const zipUtils_1 = require("../utils/zipUtils");
+const chartUtils_js_1 = require("../utils/chartUtils.js");
+const errorUtils_js_1 = require("../utils/errorUtils.js");
+const imageUtils_js_1 = require("../utils/imageUtils.js");
+const ocrUtils_js_1 = require("../utils/ocrUtils.js");
+const xmlUtils_js_1 = require("../utils/xmlUtils.js");
+const zipUtils_js_1 = require("../utils/zipUtils.js");
 /**
  * Parses an OpenOffice document (.odt, .odp, .ods) and extracts content.
  *
@@ -43,7 +43,7 @@ const parseOpenOffice = async (buffer, config) => {
     const metaFileRegex = /meta\.xml/;
     const stylesFileRegex = /styles\.xml/;
     const mimetypeFileRegex = /mimetype/;
-    const files = await (0, zipUtils_1.extractFiles)(buffer, x => !!x.match(contentFileRegex) ||
+    const files = await (0, zipUtils_js_1.extractFiles)(buffer, x => !!x.match(contentFileRegex) ||
         !!x.match(objectContentFileRegex) ||
         !!x.match(metaFileRegex) ||
         !!x.match(stylesFileRegex) ||
@@ -70,17 +70,22 @@ const parseOpenOffice = async (buffer, config) => {
     const styleMap = {};
     const paragraphStyleMap = {};
     const listCounters = {}; // Track item index per listId/level
+    let currentListId = null;
+    let lastListType = null;
+    let lastListStyle = null;
+    let listIdCounter = 0;
+    let lastWasList = false;
     // Helper to parse styles
     const parseStyles = (xmlString) => {
-        const xml = (0, xmlUtils_1.parseXmlString)(xmlString);
-        const styles = (0, xmlUtils_1.getElementsByTagName)(xml, "style:style");
+        const xml = (0, xmlUtils_js_1.parseXmlString)(xmlString);
+        const styles = (0, xmlUtils_js_1.getElementsByTagName)(xml, "style:style");
         for (const style of styles) {
             const name = style.getAttribute("style:name");
             if (!name)
                 continue;
             const styleInfo = {};
             // Parse paragraph properties for alignment and drop caps
-            const paraProps = (0, xmlUtils_1.getElementsByTagName)(style, "style:paragraph-properties")[0];
+            const paraProps = (0, xmlUtils_js_1.getFirstElementByTagName)(style, "style:paragraph-properties");
             if (paraProps) {
                 const textAlign = paraProps.getAttribute("fo:text-align");
                 if (textAlign) {
@@ -97,7 +102,7 @@ const parseOpenOffice = async (buffer, config) => {
                     }
                 }
                 // Detect Drop Caps
-                const dropCap = (0, xmlUtils_1.getElementsByTagName)(paraProps, "style:drop-cap")[0];
+                const dropCap = (0, xmlUtils_js_1.getFirstElementByTagName)(paraProps, "style:drop-cap");
                 if (dropCap) {
                     styleInfo.dropCap = true;
                 }
@@ -106,9 +111,9 @@ const parseOpenOffice = async (buffer, config) => {
                 paragraphStyleMap[name] = styleInfo;
             }
             // Parse text properties
-            const textProps = (0, xmlUtils_1.getElementsByTagName)(style, "style:text-properties")[0];
+            const textProps = (0, xmlUtils_js_1.getFirstElementByTagName)(style, "style:text-properties");
             // Parse table cell properties (for ODS background)
-            const cellProps = (0, xmlUtils_1.getElementsByTagName)(style, "style:table-cell-properties")[0];
+            const cellProps = (0, xmlUtils_js_1.getFirstElementByTagName)(style, "style:table-cell-properties");
             const formatting = {};
             if (cellProps) {
                 const bgColor = cellProps.getAttribute("fo:background-color");
@@ -170,7 +175,7 @@ const parseOpenOffice = async (buffer, config) => {
      * @param linkMetadata - Metadata inherited from parent link
      * @returns Object containing text and children
      */
-    const parseInlineContent = (node, styleMap, config, notes, paragraphStyleMap, parentFormatting = {}, linkMetadata) => {
+    const parseInlineContent = (node, styleMap, config, notes, paragraphStyleMap, parentFormatting = {}, linkMetadata, sourceXml = '') => {
         const children = [];
         let fullText = '';
         if (!node.childNodes)
@@ -189,7 +194,7 @@ const parseOpenOffice = async (buffer, config) => {
                     });
                 }
             }
-            else if (child.nodeType === 1) {
+            else if ((0, xmlUtils_js_1.isElement)(child)) {
                 const element = child;
                 const tagName = element.tagName;
                 if (tagName === 'text:s') {
@@ -221,14 +226,14 @@ const parseOpenOffice = async (buffer, config) => {
                         type: 'text',
                         text: '\n',
                         formatting: parentFormatting,
-                        metadata: linkMetadata ? { ...linkMetadata } : undefined
+                        metadata: { ...(linkMetadata || {}), isLineBreak: true }
                     });
                 }
                 else if (tagName === 'text:span') {
                     // Formatted text span
                     const styleName = element.getAttribute("text:style-name");
                     const formatting = styleName ? { ...parentFormatting, ...styleMap[styleName] } : parentFormatting;
-                    const spanContent = parseInlineContent(element, styleMap, config, notes, paragraphStyleMap, formatting, linkMetadata);
+                    const spanContent = parseInlineContent(element, styleMap, config, notes, paragraphStyleMap, formatting, linkMetadata, sourceXml);
                     fullText += spanContent.text;
                     children.push(...spanContent.children);
                 }
@@ -237,7 +242,7 @@ const parseOpenOffice = async (buffer, config) => {
                     const href = element.getAttribute('xlink:href') || '';
                     const linkType = href.startsWith('#') ? 'internal' : 'external';
                     const newLinkMetadata = { link: href, linkType: linkType };
-                    const linkContent = parseInlineContent(element, styleMap, config, notes, paragraphStyleMap, parentFormatting, newLinkMetadata);
+                    const linkContent = parseInlineContent(element, styleMap, config, notes, paragraphStyleMap, parentFormatting, newLinkMetadata, sourceXml);
                     fullText += linkContent.text;
                     children.push(...linkContent.children);
                 }
@@ -245,14 +250,14 @@ const parseOpenOffice = async (buffer, config) => {
                     // Footnote or endnote
                     const noteClass = (element.getAttribute('text:note-class') || 'footnote');
                     const noteId = element.getAttribute('text:id') || element.getAttribute('xml:id') || undefined;
-                    const noteBody = (0, xmlUtils_1.getElementsByTagName)(element, "text:note-body")[0];
+                    const noteBody = (0, xmlUtils_js_1.getFirstElementByTagName)(element, "text:note-body");
                     if (noteBody) {
                         // Extract note content recursively
-                        const notePs = (0, xmlUtils_1.getElementsByTagName)(noteBody, "text:p");
+                        const notePs = (0, xmlUtils_js_1.getElementsByTagName)(noteBody, "text:p");
                         const noteChildren = [];
                         let noteText = '';
                         for (const np of notePs) {
-                            const npContent = parseParagraphContent(np, paragraphStyleMap, styleMap, config);
+                            const npContent = parseParagraphContent(np, paragraphStyleMap, styleMap, config, sourceXml);
                             noteText += (noteText ? ' ' : '') + npContent.text;
                             const npNode = {
                                 type: 'paragraph',
@@ -284,8 +289,8 @@ const parseOpenOffice = async (buffer, config) => {
                     const frame = element;
                     // Extract alt text
                     let altText = '';
-                    const svgTitle = (0, xmlUtils_1.getElementsByTagName)(frame, "svg:title")[0];
-                    const svgDesc = (0, xmlUtils_1.getElementsByTagName)(frame, "svg:desc")[0];
+                    const svgTitle = (0, xmlUtils_js_1.getFirstElementByTagName)(frame, "svg:title");
+                    const svgDesc = (0, xmlUtils_js_1.getFirstElementByTagName)(frame, "svg:desc");
                     if (svgTitle && svgTitle.textContent) {
                         altText = svgTitle.textContent;
                     }
@@ -294,7 +299,7 @@ const parseOpenOffice = async (buffer, config) => {
                     }
                     // Extract image href
                     let imageHref = '';
-                    const drawImages = (0, xmlUtils_1.getElementsByTagName)(frame, "draw:image");
+                    const drawImages = (0, xmlUtils_js_1.getElementsByTagName)(frame, "draw:image");
                     if (drawImages.length > 0) {
                         imageHref = drawImages[0].getAttribute("xlink:href") || '';
                         if (imageHref) {
@@ -312,7 +317,7 @@ const parseOpenOffice = async (buffer, config) => {
                         }
                     };
                     if (config.includeRawContent) {
-                        imageNode.rawContent = frame.toString();
+                        imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frame, sourceXml, config);
                     }
                     children.push(imageNode);
                 }
@@ -330,14 +335,14 @@ const parseOpenOffice = async (buffer, config) => {
      * @param config - Parser configuration
      * @returns Object containing text, children, alignment, and style info
      */
-    const parseParagraphContent = (node, paraStyleMap, styleMap, config) => {
+    const parseParagraphContent = (node, paraStyleMap, styleMap, config, sourceXml) => {
         // Get paragraph style for alignment and drop caps
         const paraStyle = node.getAttribute("text:style-name");
         const styleInfo = paraStyle ? paraStyleMap[paraStyle] : undefined;
         const alignment = styleInfo?.alignment;
         const dropCap = styleInfo?.dropCap;
         // Parse content recursively using the new helper
-        const content = parseInlineContent(node, styleMap, config, notes, paraStyleMap);
+        const content = parseInlineContent(node, styleMap, config, notes, paraStyleMap, {}, undefined, sourceXml);
         // Add style name to metadata of children if they don't have one
         if (paraStyle) {
             content.children.forEach(child => {
@@ -391,6 +396,31 @@ const parseOpenOffice = async (buffer, config) => {
         }
         return { text: content.text, children: content.children, alignment, style: paraStyle || undefined };
     };
+    /**
+     * Splits paragraph content into multiple segments based on line breaks.
+     * Used to handle soft line breaks within list items.
+     *
+     * @param pContent - The content of a single paragraph
+     * @returns Array of content segments
+     */
+    const splitParagraphByBreaks = (pContent) => {
+        const segments = [];
+        let currentText = "";
+        let currentChildren = [];
+        for (const child of pContent.children) {
+            if (child.type === "text" && child.metadata?.isLineBreak) {
+                segments.push({ text: currentText, children: currentChildren });
+                currentText = "";
+                currentChildren = [];
+            }
+            else {
+                currentText += child.text || "";
+                currentChildren.push(child);
+            }
+        }
+        segments.push({ text: currentText, children: currentChildren });
+        return segments;
+    };
     /**
      * Helper to parse a table node and extract its structure.
      * Properly creates table → row → cell hierarchy with metadata.
@@ -401,15 +431,15 @@ const parseOpenOffice = async (buffer, config) => {
      * @param config - Parser configuration
      * @returns Table content node with proper structure
      */
-    const parseTable = (tableNode, paraStyleMap, styleMap, config) => {
+    const parseTable = (tableNode, paraStyleMap, styleMap, config, sourceXml) => {
         const rows = [];
         // Use getDirectChildren to avoid nested table rows
-        const tableRows = (0, xmlUtils_1.getDirectChildren)(tableNode, "table:table-row");
+        const tableRows = (0, xmlUtils_js_1.getDirectChildren)(tableNode, "table:table-row");
         let rowIndex = 0;
         for (const row of tableRows) {
             const cells = [];
             // Use getDirectChildren to avoid nested table cells
-            const tableCells = (0, xmlUtils_1.getDirectChildren)(row, "table:table-cell");
+            const tableCells = (0, xmlUtils_js_1.getDirectChildren)(row, "table:table-cell");
             const rowsRepeated = parseInt(row.getAttribute("table:number-rows-repeated") || "1");
             let colIndex = 0;
             for (const cell of tableCells) {
@@ -424,10 +454,10 @@ const parseOpenOffice = async (buffer, config) => {
                         return;
                     for (let i = 0; i < node.childNodes.length; i++) {
                         const child = node.childNodes[i];
-                        if (child.nodeType === 1) { // Element
+                        if ((0, xmlUtils_js_1.isElement)(child)) { // Element
                             const element = child;
                             if (element.tagName === "text:p" || element.tagName === "text:h") {
-                                const pContent = parseParagraphContent(element, paraStyleMap, styleMap, config);
+                                const pContent = parseParagraphContent(element, paraStyleMap, styleMap, config, sourceXml);
                                 const pNode = {
                                     type: element.tagName === "text:h" ? 'heading' : 'paragraph',
                                     text: pContent.text,
@@ -446,7 +476,7 @@ const parseOpenOffice = async (buffer, config) => {
                                     pNode.metadata.level = parseInt(element.getAttribute("text:outline-level") || "1");
                                 }
                                 if (config.includeRawContent) {
-                                    pNode.rawContent = element.toString();
+                                    pNode.rawContent = (0, xmlUtils_js_1.getRawContent)(element, sourceXml, config);
                                 }
                                 cellChildren.push(pNode);
                                 cellTextRef.value += pContent.text;
@@ -457,7 +487,7 @@ const parseOpenOffice = async (buffer, config) => {
                             }
                             else if (element.tagName === "table:table") {
                                 // Recursive call for nested table
-                                const nestedTableNode = parseTable(element, paraStyleMap, styleMap, config);
+                                const nestedTableNode = parseTable(element, paraStyleMap, styleMap, config, sourceXml);
                                 cellChildren.push(nestedTableNode);
                             }
                             else if (element.tagName === "draw:frame" || element.tagName === "draw:text-box") {
@@ -487,7 +517,7 @@ const parseOpenOffice = async (buffer, config) => {
                     if (rowSpan > 1)
                         cellMetadata.rowSpan = rowSpan;
                     if (config.includeRawContent) {
-                        cellNode.rawContent = cell.toString();
+                        cellNode.rawContent = (0, xmlUtils_js_1.getRawContent)(cell, sourceXml, config);
                     }
                     cells.push(cellNode);
                     colIndex++;
@@ -508,7 +538,7 @@ const parseOpenOffice = async (buffer, config) => {
                     });
                 }
                 if (config.includeRawContent) {
-                    rowNode.rawContent = row.toString();
+                    rowNode.rawContent = (0, xmlUtils_js_1.getRawContent)(row, sourceXml, config);
                 }
                 rows.push(rowNode);
                 rowIndex++;
@@ -520,18 +550,20 @@ const parseOpenOffice = async (buffer, config) => {
         };
     };
     const parseContentXml = (xmlString) => {
-        const xml = (0, xmlUtils_1.parseXmlString)(xmlString);
-        const body = (0, xmlUtils_1.getElementsByTagName)(xml, "office:body")[0];
+        const xml = (0, xmlUtils_js_1.parseXmlString)(xmlString, { locator: config.includeRawContent });
+        const body = (0, xmlUtils_js_1.getFirstElementByTagName)(xml, "office:body");
+        if (!body)
+            return;
         // Parse automatic styles (local to content.xml)
-        const automaticStyles = (0, xmlUtils_1.getElementsByTagName)(xml, "office:automatic-styles")[0];
+        const automaticStyles = (0, xmlUtils_js_1.getFirstElementByTagName)(xml, "office:automatic-styles");
         if (automaticStyles) {
-            const styles = (0, xmlUtils_1.getElementsByTagName)(automaticStyles, "style:style");
+            const styles = (0, xmlUtils_js_1.getElementsByTagName)(automaticStyles, "style:style");
             for (const style of styles) {
                 const name = style.getAttribute("style:name");
                 if (!name)
                     continue;
                 // Parse paragraph properties for alignment
-                const paraProps = (0, xmlUtils_1.getElementsByTagName)(style, "style:paragraph-properties")[0];
+                const paraProps = (0, xmlUtils_js_1.getFirstElementByTagName)(style, "style:paragraph-properties");
                 const styleInfo = {};
                 if (paraProps) {
                     const textAlign = paraProps.getAttribute("fo:text-align");
@@ -548,14 +580,14 @@ const parseOpenOffice = async (buffer, config) => {
                             styleInfo.alignment = alignMap[textAlign];
                         }
                     }
-                    const dropCap = (0, xmlUtils_1.getElementsByTagName)(paraProps, "style:drop-cap")[0];
+                    const dropCap = (0, xmlUtils_js_1.getFirstElementByTagName)(paraProps, "style:drop-cap");
                     if (dropCap)
                         styleInfo.dropCap = true;
                 }
                 if (Object.keys(styleInfo).length > 0) {
                     paragraphStyleMap[name] = styleInfo;
                 }
-                const textProps = (0, xmlUtils_1.getElementsByTagName)(style, "style:text-properties")[0];
+                const textProps = (0, xmlUtils_js_1.getFirstElementByTagName)(style, "style:text-properties");
                 if (textProps) {
                     const formatting = {};
                     if (textProps.getAttribute("fo:font-weight") === "bold" || textProps.getAttribute("style:font-weight-asian") === "bold")
@@ -593,6 +625,19 @@ const parseOpenOffice = async (buffer, config) => {
                 }
             }
         }
+        // Start traversal
+        const officeBody = (0, xmlUtils_js_1.getFirstElementByTagName)(xml, "office:body");
+        if (officeBody) {
+            const bodyContent = (0, xmlUtils_js_1.getDirectChildren)(officeBody, "office:text")[0] ||
+                (0, xmlUtils_js_1.getDirectChildren)(officeBody, "office:presentation")[0] ||
+                (0, xmlUtils_js_1.getDirectChildren)(officeBody, "office:spreadsheet")[0];
+            if (bodyContent) {
+                const bodyChildren = (0, xmlUtils_js_1.getDirectChildren)(bodyContent, "*");
+                for (const child of bodyChildren) {
+                    traverse(child, content, false, xmlString);
+                }
+            }
+        }
         /**
          * Recursively traverses a node and its children to extract content.
          * Properly handles paragraphs, headings, tables, lists, and frames.
@@ -600,10 +645,11 @@ const parseOpenOffice = async (buffer, config) => {
          * @param node - The element to traverse
          * @param targetArray - The array to push extracted content nodes to
          * @param forceHeading - If true, treats all paragraphs as headings (used for slide titles)
+         * @param sourceXml - The source XML string for raw content extraction
          */
-        const traverse = (node, targetArray, forceHeading = false) => {
+        function traverse(node, targetArray, forceHeading = false, sourceXml) {
             if (node.tagName === "text:p") {
-                const pContent = parseParagraphContent(node, paragraphStyleMap, styleMap, config);
+                const pContent = parseParagraphContent(node, paragraphStyleMap, styleMap, config, sourceXml);
                 const type = (forceHeading || (node.getAttribute("text:style-name") || '').toLowerCase().includes('title')) ? 'heading' : 'paragraph';
                 const pNode = {
                     type,
@@ -621,13 +667,14 @@ const parseOpenOffice = async (buffer, config) => {
                 if (Object.keys(pNode.metadata || {}).length === 0)
                     delete pNode.metadata;
                 if (config.includeRawContent) {
-                    pNode.rawContent = node.toString();
+                    pNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
                 }
                 targetArray.push(pNode);
+                lastWasList = false;
             }
             else if (node.tagName === "text:h") {
                 const level = parseInt(node.getAttribute("text:outline-level") || "1");
-                const hContent = parseParagraphContent(node, paragraphStyleMap, styleMap, config);
+                const hContent = parseParagraphContent(node, paragraphStyleMap, styleMap, config, sourceXml);
                 const hNode = {
                     type: 'heading',
                     text: hContent.text,
@@ -639,27 +686,27 @@ const parseOpenOffice = async (buffer, config) => {
                     }
                 };
                 if (config.includeRawContent) {
-                    hNode.rawContent = node.toString();
+                    hNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
                 }
                 targetArray.push(hNode);
+                lastWasList = false;
             }
             else if (node.tagName === "table:table") {
                 // Parse table with proper structure
-                const tableNode = parseTable(node, paragraphStyleMap, styleMap, config);
+                const tableNode = parseTable(node, paragraphStyleMap, styleMap, config, sourceXml);
                 if (config.includeRawContent) {
-                    tableNode.rawContent = node.toString();
+                    tableNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
                 }
                 targetArray.push(tableNode);
+                lastWasList = false;
             }
             else if (node.tagName === "text:list") {
                 // Parse list structure with proper listId tracking
-                const listItems = (0, xmlUtils_1.getDirectChildren)(node, "text:list-item");
-                // Get list style name to use as listId (or generate one)
-                const listStyleName = node.getAttribute("text:style-name") || node.getAttribute("xml:id");
-                const listId = listStyleName || `list-${targetArray.length}`;
+                const listItems = (0, xmlUtils_js_1.getDirectChildren)(node, "text:list-item");
                 // Determine list type by checking the list style definition
                 let listType = 'unordered';
                 let isVisible = false;
+                const listStyleName = node.getAttribute("text:style-name") || node.getAttribute("xml:id");
                 let styleNameToCheck = listStyleName;
                 // If no style name, check parent list for inherited style
                 if (!styleNameToCheck) {
@@ -675,15 +722,15 @@ const parseOpenOffice = async (buffer, config) => {
                 }
                 // Try to find list style in automatic styles to determine type and visibility
                 if (styleNameToCheck) {
-                    const automaticStyles = (0, xmlUtils_1.getElementsByTagName)((0, xmlUtils_1.parseXmlString)(mainContentFile?.content.toString() || ''), "office:automatic-styles")[0];
+                    const automaticStyles = (0, xmlUtils_js_1.getFirstElementByTagName)((0, xmlUtils_js_1.parseXmlString)(mainContentFile?.content.toString() || ''), "office:automatic-styles");
                     if (automaticStyles) {
-                        const listStyles = (0, xmlUtils_1.getElementsByTagName)(automaticStyles, "text:list-style");
+                        const listStyles = (0, xmlUtils_js_1.getElementsByTagName)(automaticStyles, "text:list-style");
                         for (const listStyle of listStyles) {
                             if (listStyle.getAttribute("style:name") === styleNameToCheck) {
                                 // Check if it has bullet or number level styles
-                                const bulletLevels = (0, xmlUtils_1.getElementsByTagName)(listStyle, "text:list-level-style-bullet");
-                                const numberLevels = (0, xmlUtils_1.getElementsByTagName)(listStyle, "text:list-level-style-number");
-                                const imageLevels = (0, xmlUtils_1.getElementsByTagName)(listStyle, "text:list-level-style-image");
+                                const bulletLevels = (0, xmlUtils_js_1.getElementsByTagName)(listStyle, "text:list-level-style-bullet");
+                                const numberLevels = (0, xmlUtils_js_1.getElementsByTagName)(listStyle, "text:list-level-style-number");
+                                const imageLevels = (0, xmlUtils_js_1.getElementsByTagName)(listStyle, "text:list-level-style-image");
                                 if (numberLevels.length > 0) {
                                     listType = 'ordered';
                                     isVisible = numberLevels.some(l => !!l.getAttribute("style:num-format"));
@@ -700,13 +747,13 @@ const parseOpenOffice = async (buffer, config) => {
                     }
                     // Also check in styles.xml if still unordered and hidden
                     if (stylesFile && !isVisible) {
-                        const stylesXml = (0, xmlUtils_1.parseXmlString)(stylesFile.content.toString());
-                        const listStyles = (0, xmlUtils_1.getElementsByTagName)(stylesXml, "text:list-style");
+                        const stylesXml = (0, xmlUtils_js_1.parseXmlString)(stylesFile.content.toString());
+                        const listStyles = (0, xmlUtils_js_1.getElementsByTagName)(stylesXml, "text:list-style");
                         for (const listStyle of listStyles) {
                             if (listStyle.getAttribute("style:name") === styleNameToCheck) {
-                                const bulletLevels = (0, xmlUtils_1.getElementsByTagName)(listStyle, "text:list-level-style-bullet");
-                                const numberLevels = (0, xmlUtils_1.getElementsByTagName)(listStyle, "text:list-level-style-number");
-                                const imageLevels = (0, xmlUtils_1.getElementsByTagName)(listStyle, "text:list-level-style-image");
+                                const bulletLevels = (0, xmlUtils_js_1.getElementsByTagName)(listStyle, "text:list-level-style-bullet");
+                                const numberLevels = (0, xmlUtils_js_1.getElementsByTagName)(listStyle, "text:list-level-style-number");
+                                const imageLevels = (0, xmlUtils_js_1.getElementsByTagName)(listStyle, "text:list-level-style-image");
                                 if (numberLevels.length > 0) {
                                     listType = 'ordered';
                                     isVisible = numberLevels.some(l => !!l.getAttribute("style:num-format"));
@@ -725,19 +772,38 @@ const parseOpenOffice = async (buffer, config) => {
                 // If the list is not visible, it's likely a layout list used by Impress.
                 // We should traverse its items and treat their content as regular nodes.
                 if (!isVisible) {
+                    lastWasList = false;
                     for (let i = 0; i < listItems.length; i++) {
                         const item = listItems[i];
                         if (item.childNodes) {
                             for (let j = 0; j < item.childNodes.length; j++) {
                                 const child = item.childNodes[j];
-                                if (child.nodeType === 1) { // Element
-                                    traverse(child, targetArray, forceHeading);
+                                if ((0, xmlUtils_js_1.isElement)(child)) { // Element
+                                    traverse(child, targetArray, forceHeading, sourceXml);
                                 }
                             }
                         }
                     }
                     return;
                 }
+                // List Continuity Logic:
+                // If this list follows another list of the same type and style, or we are in ODP and it's sequential,
+                // we should reuse the previous listId to maintain numbering.
+                const isODP = fileType === 'odp';
+                const sameStyle = styleNameToCheck && styleNameToCheck === lastListStyle;
+                const sameType = listType === lastListType;
+                let listId;
+                if (lastWasList && (sameStyle || (isODP && sameType))) {
+                    listId = currentListId;
+                }
+                else {
+                    // New list
+                    listId = styleNameToCheck || `list-${++listIdCounter}`;
+                    currentListId = listId;
+                    lastListType = listType;
+                    lastListStyle = styleNameToCheck;
+                }
+                lastWasList = true;
                 // Calculate indentation level by counting parent text:list elements
                 let indentation = 0;
                 let parent = node.parentNode;
@@ -758,63 +824,61 @@ const parseOpenOffice = async (buffer, config) => {
                 // Process each list item
                 for (let i = 0; i < listItems.length; i++) {
                     const item = listItems[i];
-                    // Increment item index for this list/level
-                    listCounters[listId][indentKey]++;
-                    const itemIndex = listCounters[listId][indentKey];
-                    // Reset deeper levels when we encounter an item at this level
-                    for (let k = indentation + 1; k < 10; k++) {
-                        if (listCounters[listId][k.toString()] !== undefined) {
-                            listCounters[listId][k.toString()] = -1;
-                        }
-                    }
+                    let hasIndexedThisItem = false;
                     // Iterate over direct children of list item (paragraphs, headings, nested lists)
                     if (item.childNodes) {
                         for (let j = 0; j < item.childNodes.length; j++) {
                             const child = item.childNodes[j];
-                            if (child.nodeType === 1) { // Element
+                            if ((0, xmlUtils_js_1.isElement)(child)) { // Element
                                 const element = child;
-                                if (element.tagName === "text:p") {
-                                    const pContent = parseParagraphContent(element, paragraphStyleMap, styleMap, config);
-                                    const listNode = {
-                                        type: 'list',
-                                        text: pContent.text,
-                                        children: pContent.children,
-                                        metadata: {
-                                            listType,
-                                            indentation,
-                                            itemIndex,
-                                            listId,
-                                            alignment: pContent.alignment || 'left',
-                                            style: pContent.style
+                                if (element.tagName === "text:p" || element.tagName === "text:h") {
+                                    if (!hasIndexedThisItem) {
+                                        listCounters[listId][indentKey]++;
+                                        hasIndexedThisItem = true;
+                                        for (let k = indentation + 1; k < 10; k++) {
+                                            if (listCounters[listId][k.toString()] !== undefined) {
+                                                listCounters[listId][k.toString()] = -1;
+                                            }
                                         }
-                                    };
-                                    if (config.includeRawContent)
-                                        listNode.rawContent = element.toString();
-                                    targetArray.push(listNode);
-                                }
-                                else if (element.tagName === "text:h") {
-                                    const level = parseInt(element.getAttribute("text:outline-level") || "1");
-                                    const hContent = parseParagraphContent(element, paragraphStyleMap, styleMap, config);
-                                    const listNode = {
-                                        type: 'list',
-                                        text: hContent.text,
-                                        children: hContent.children,
-                                        metadata: {
-                                            listType,
-                                            indentation,
-                                            itemIndex,
-                                            listId,
-                                            ...(hContent.alignment ? { alignment: hContent.alignment } : {}),
-                                            style: hContent.style
+                                    }
+                                    const itemIndex = listCounters[listId][indentKey];
+                                    const pContent = parseParagraphContent(element, paragraphStyleMap, styleMap, config, sourceXml);
+                                    const segments = splitParagraphByBreaks(pContent);
+                                    for (let k = 0; k < segments.length; k++) {
+                                        const segment = segments[k];
+                                        if (!segment.text.trim() && segment.children.length === 0)
+                                            continue;
+                                        const isFirst = k === 0;
+                                        const nodeType = isFirst ? 'list' : 'paragraph';
+                                        const node = {
+                                            type: nodeType,
+                                            text: segment.text,
+                                            children: segment.children,
+                                            metadata: isFirst ? {
+                                                listType,
+                                                indentation,
+                                                itemIndex,
+                                                listId,
+                                                alignment: pContent.alignment || 'left',
+                                                style: pContent.style
+                                            } : {
+                                                alignment: pContent.alignment || 'left',
+                                                style: pContent.style
+                                            }
+                                        };
+                                        // Special case for headings in lists
+                                        if (isFirst && element.tagName === "text:h") {
+                                            const level = parseInt(element.getAttribute("text:outline-level") || "1");
+                                            node.metadata.level = level;
                                         }
-                                    };
-                                    if (config.includeRawContent)
-                                        listNode.rawContent = element.toString();
-                                    targetArray.push(listNode);
+                                        if (config.includeRawContent)
+                                            node.rawContent = (0, xmlUtils_js_1.getRawContent)(element, sourceXml, config);
+                                        targetArray.push(node);
+                                    }
                                 }
                                 else if (element.tagName === "text:list") {
                                     // Recursive call for nested list
-                                    traverse(element, targetArray, forceHeading);
+                                    traverse(element, targetArray, forceHeading, sourceXml);
                                 }
                             }
                         }
@@ -825,24 +889,24 @@ const parseOpenOffice = async (buffer, config) => {
                 const presClass = node.getAttribute("presentation:class");
                 const isHeading = presClass === "title" || presClass === "sub-title";
                 // In presentations, frames often contain text-boxes, images, tables, or objects
-                const textBox = (0, xmlUtils_1.getElementsByTagName)(node, "draw:text-box")[0];
-                const image = (0, xmlUtils_1.getElementsByTagName)(node, "draw:image")[0];
-                const table = (0, xmlUtils_1.getElementsByTagName)(node, "table:table")[0];
-                const object = (0, xmlUtils_1.getElementsByTagName)(node, "draw:object")[0];
+                const textBox = (0, xmlUtils_js_1.getFirstElementByTagName)(node, "draw:text-box");
+                const image = (0, xmlUtils_js_1.getFirstElementByTagName)(node, "draw:image");
+                const table = (0, xmlUtils_js_1.getFirstElementByTagName)(node, "table:table");
+                const object = (0, xmlUtils_js_1.getFirstElementByTagName)(node, "draw:object");
                 if (textBox) {
-                    traverse(textBox, targetArray, isHeading || forceHeading);
+                    traverse(textBox, targetArray, isHeading || forceHeading, sourceXml);
                 }
                 else if (table) {
-                    const tableNode = parseTable(table, paragraphStyleMap, styleMap, config);
+                    const tableNode = parseTable(table, paragraphStyleMap, styleMap, config, sourceXml);
                     if (config.includeRawContent)
-                        tableNode.rawContent = table.toString();
+                        tableNode.rawContent = (0, xmlUtils_js_1.getRawContent)(table, sourceXml, config);
                     targetArray.push(tableNode);
                 }
                 else if (image) {
                     // Extract alt text from svg:title or svg:desc
                     let altText = '';
-                    const svgTitle = (0, xmlUtils_1.getElementsByTagName)(node, "svg:title")[0];
-                    const svgDesc = (0, xmlUtils_1.getElementsByTagName)(node, "svg:desc")[0];
+                    const svgTitle = (0, xmlUtils_js_1.getFirstElementByTagName)(node, "svg:title");
+                    const svgDesc = (0, xmlUtils_js_1.getFirstElementByTagName)(node, "svg:desc");
                     if (svgTitle && svgTitle.textContent) {
                         altText = svgTitle.textContent;
                     }
@@ -865,7 +929,7 @@ const parseOpenOffice = async (buffer, config) => {
                         }
                     };
                     if (config.includeRawContent) {
-                        imageNode.rawContent = node.toString();
+                        imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
                     }
                     targetArray.push(imageNode);
                 }
@@ -877,7 +941,7 @@ const parseOpenOffice = async (buffer, config) => {
                         const objectPath = `${attachmentName}/content.xml`;
                         const objectFile = files.find(f => f.path === objectPath || f.path.endsWith(objectPath));
                         if (objectFile) {
-                            const chartData = (0, chartUtils_1.extractChartData)(objectFile.content);
+                            const chartData = (0, chartUtils_js_1.extractChartData)(objectFile.content);
                             const chartNode = {
                                 type: 'chart',
                                 text: chartData.rawTexts.join(" "),
@@ -887,7 +951,7 @@ const parseOpenOffice = async (buffer, config) => {
                                 }
                             };
                             if (config.includeRawContent)
-                                chartNode.rawContent = node.toString();
+                                chartNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
                             targetArray.push(chartNode);
                         }
                         else {
@@ -897,7 +961,7 @@ const parseOpenOffice = async (buffer, config) => {
                                 metadata: { attachmentName: attachmentName }
                             };
                             if (config.includeRawContent)
-                                chartNode.rawContent = node.toString();
+                                chartNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
                             targetArray.push(chartNode);
                         }
                     }
@@ -907,28 +971,29 @@ const parseOpenOffice = async (buffer, config) => {
                 if (node.childNodes) {
                     for (let i = 0; i < node.childNodes.length; i++) {
                         const child = node.childNodes[i];
-                        if (child.nodeType === 1) { // Element
-                            traverse(child, targetArray, forceHeading);
+                        if ((0, xmlUtils_js_1.isElement)(child)) { // Element
+                            traverse(child, targetArray, forceHeading, sourceXml);
                         }
                     }
                 }
             }
-        };
+        }
+        ;
         // ODS: Spreadsheet
         if (fileType === 'ods') {
-            const spreadsheet = (0, xmlUtils_1.getElementsByTagName)(body, "office:spreadsheet")[0];
+            const spreadsheet = (0, xmlUtils_js_1.getFirstElementByTagName)(body, "office:spreadsheet");
             if (spreadsheet) {
-                const tables = (0, xmlUtils_1.getElementsByTagName)(spreadsheet, "table:table");
+                const tables = (0, xmlUtils_js_1.getElementsByTagName)(spreadsheet, "table:table");
                 for (let i = 0; i < tables.length; i++) {
                     const table = tables[i];
                     const sheetName = table.getAttribute("table:name") || `Sheet${i + 1}`;
                     const rows = [];
-                    const tableRows = (0, xmlUtils_1.getElementsByTagName)(table, "table:table-row");
+                    const tableRows = (0, xmlUtils_js_1.getElementsByTagName)(table, "table:table-row");
                     let rowIndex = 0;
                     for (let r = 0; r < tableRows.length; r++) {
                         const row = tableRows[r];
                         const cells = [];
-                        const tableCells = (0, xmlUtils_1.getElementsByTagName)(row, "table:table-cell");
+                        const tableCells = (0, xmlUtils_js_1.getElementsByTagName)(row, "table:table-cell");
                         let colIndex = 0;
                         const rowsRepeated = parseInt(row.getAttribute("table:number-rows-repeated") || "1");
                         for (let c = 0; c < tableCells.length; c++) {
@@ -937,11 +1002,11 @@ const parseOpenOffice = async (buffer, config) => {
                             // Extract text from cell (paragraphs inside cell)
                             let cellText = "";
                             const children = [];
-                            const ps = (0, xmlUtils_1.getElementsByTagName)(cell, "text:p");
+                            const ps = (0, xmlUtils_js_1.getElementsByTagName)(cell, "text:p");
                             for (let p = 0; p < ps.length; p++) {
                                 const para = ps[p];
                                 // Parse text:span elements for formatted text
-                                const spans = (0, xmlUtils_1.getElementsByTagName)(para, "text:span");
+                                const spans = (0, xmlUtils_js_1.getElementsByTagName)(para, "text:span");
                                 if (spans.length > 0) {
                                     for (const span of spans) {
                                         const styleName = span.getAttribute("text:style-name");
@@ -973,12 +1038,12 @@ const parseOpenOffice = async (buffer, config) => {
                                     cellText += "\n";
                             }
                             // Check for embedded draw:frame (images) in cell
-                            const drawFrames = (0, xmlUtils_1.getElementsByTagName)(cell, "draw:frame");
+                            const drawFrames = (0, xmlUtils_js_1.getElementsByTagName)(cell, "draw:frame");
                             for (const frame of drawFrames) {
                                 // Extract alt text from svg:title or svg:desc
                                 let altText = '';
-                                const svgTitle = (0, xmlUtils_1.getElementsByTagName)(frame, "svg:title")[0];
-                                const svgDesc = (0, xmlUtils_1.getElementsByTagName)(frame, "svg:desc")[0];
+                                const svgTitle = (0, xmlUtils_js_1.getFirstElementByTagName)(frame, "svg:title");
+                                const svgDesc = (0, xmlUtils_js_1.getFirstElementByTagName)(frame, "svg:desc");
                                 if (svgTitle && svgTitle.textContent) {
                                     altText = svgTitle.textContent;
                                 }
@@ -987,7 +1052,7 @@ const parseOpenOffice = async (buffer, config) => {
                                 }
                                 // Extract image href
                                 let imageHref = '';
-                                const drawImages = (0, xmlUtils_1.getElementsByTagName)(frame, "draw:image");
+                                const drawImages = (0, xmlUtils_js_1.getElementsByTagName)(frame, "draw:image");
                                 if (drawImages.length > 0) {
                                     const rawHref = drawImages[0].getAttribute("xlink:href");
                                     if (rawHref) {
@@ -997,7 +1062,7 @@ const parseOpenOffice = async (buffer, config) => {
                                 }
                                 // Extract chart object href
                                 let chartHref = '';
-                                const drawObjects = (0, xmlUtils_1.getElementsByTagName)(frame, "draw:object");
+                                const drawObjects = (0, xmlUtils_js_1.getElementsByTagName)(frame, "draw:object");
                                 if (drawObjects.length > 0) {
                                     const href = drawObjects[0].getAttribute("xlink:href");
                                     if (href) {
@@ -1017,7 +1082,7 @@ const parseOpenOffice = async (buffer, config) => {
                                         }
                                     };
                                     if (config.includeRawContent) {
-                                        imageNode.rawContent = frame.toString();
+                                        imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frame, xmlString, config);
                                     }
                                     children.push(imageNode);
                                 }
@@ -1046,7 +1111,7 @@ const parseOpenOffice = async (buffer, config) => {
                                         metadata: { row: rowIndex, col: colIndex }
                                     };
                                     if (config.includeRawContent) {
-                                        cellNode.rawContent = cell.toString();
+                                        cellNode.rawContent = (0, xmlUtils_js_1.getRawContent)(cell, xmlString, config);
                                     }
                                     cells.push(cellNode);
                                 }
@@ -1070,7 +1135,7 @@ const parseOpenOffice = async (buffer, config) => {
                                     });
                                 }
                                 if (config.includeRawContent) {
-                                    rowNode.rawContent = row.toString();
+                                    rowNode.rawContent = (0, xmlUtils_js_1.getRawContent)(row, xmlString, config);
                                 }
                                 rows.push(rowNode);
                                 rowIndex++;
@@ -1086,7 +1151,7 @@ const parseOpenOffice = async (buffer, config) => {
                         metadata: { sheetName }
                     };
                     if (config.includeRawContent) {
-                        sheetNode.rawContent = table.toString();
+                        sheetNode.rawContent = (0, xmlUtils_js_1.getRawContent)(table, xmlString, config);
                     }
                     content.push(sheetNode);
                 }
@@ -1094,9 +1159,9 @@ const parseOpenOffice = async (buffer, config) => {
         }
         // ODP: Presentation
         else if (fileType === 'odp') {
-            const presentation = (0, xmlUtils_1.getElementsByTagName)(body, "office:presentation")[0];
+            const presentation = (0, xmlUtils_js_1.getFirstElementByTagName)(body, "office:presentation");
             if (presentation) {
-                const pages = (0, xmlUtils_1.getDirectChildren)(presentation, "draw:page");
+                const pages = (0, xmlUtils_js_1.getDirectChildren)(presentation, "draw:page");
                 const odpNotes = [];
                 for (let i = 0; i < pages.length; i++) {
                     const page = pages[i];
@@ -1111,7 +1176,7 @@ const parseOpenOffice = async (buffer, config) => {
                     if (pageChildren) {
                         for (let j = 0; j < pageChildren.length; j++) {
                             const child = pageChildren[j];
-                            if (child.nodeType === 1) { // Element
+                            if ((0, xmlUtils_js_1.isElement)(child)) { // Element
                                 const element = child;
                                 if (element.tagName === "presentation:notes") {
                                     if (!config.ignoreNotes) {
@@ -1123,16 +1188,16 @@ const parseOpenOffice = async (buffer, config) => {
                                                 noteId: `slide-note-${i + 1}`
                                             }
                                         };
-                                        traverse(element, noteNode.children);
+                                        traverse(element, noteNode.children, false, xmlString);
                                     }
                                     continue;
                                 }
-                                traverse(element, slideNode.children);
+                                traverse(element, slideNode.children, false, xmlString);
                             }
                         }
                     }
                     if (config.includeRawContent) {
-                        slideNode.rawContent = page.toString();
+                        slideNode.rawContent = (0, xmlUtils_js_1.getRawContent)(page, xmlString, config);
                     }
                     content.push(slideNode);
                     if (noteNode && noteNode.children && noteNode.children.length > 0) {
@@ -1151,9 +1216,9 @@ const parseOpenOffice = async (buffer, config) => {
         }
         // ODT: Text Document (and generic fallback)
         else {
-            const textDoc = (0, xmlUtils_1.getElementsByTagName)(body, "office:text")[0];
+            const textDoc = (0, xmlUtils_js_1.getFirstElementByTagName)(body, "office:text");
             if (textDoc) {
-                traverse(textDoc, content);
+                traverse(textDoc, content, false, xmlString);
             }
         }
     };
@@ -1167,8 +1232,8 @@ const parseOpenOffice = async (buffer, config) => {
     if (config.extractAttachments) {
         const objectFiles = files.filter(f => f.path.match(/Object \d+\/content\.xml/));
         for (const objFile of objectFiles) {
-            const objXml = (0, xmlUtils_1.parseXmlString)(objFile.content.toString());
-            const isChart = (0, xmlUtils_1.getElementsByTagName)(objXml, "chart:chart").length > 0;
+            const objXml = (0, xmlUtils_js_1.parseXmlString)(objFile.content.toString());
+            const isChart = (0, xmlUtils_js_1.getElementsByTagName)(objXml, "chart:chart").length > 0;
             if (isChart) {
                 const objectId = objFile.path.split('/')[0];
                 const attachment = {
@@ -1179,7 +1244,7 @@ const parseOpenOffice = async (buffer, config) => {
                     extension: 'xml'
                 };
                 // Extract data from chart XML
-                const chartData = (0, chartUtils_1.extractChartData)(objFile.content);
+                const chartData = (0, chartUtils_js_1.extractChartData)(objFile.content);
                 if (chartData.rawTexts.length > 0) {
                     attachment.chartData = chartData;
                 }
@@ -1189,22 +1254,22 @@ const parseOpenOffice = async (buffer, config) => {
     }
     if (config.extractAttachments) {
         for (const media of mediaFiles) {
-            const attachment = (0, imageUtils_1.createAttachment)(media.path.split('/').pop() || 'image', media.content);
+            const attachment = (0, imageUtils_js_1.createAttachment)(media.path.split('/').pop() || 'image', media.content);
             attachments.push(attachment);
             if (config.ocr) {
                 if (attachment.mimeType.startsWith('image/')) {
                     try {
-                        attachment.ocrText = (await (0, ocrUtils_1.performOcr)(media.content, config.ocrLanguage)).trim();
+                        attachment.ocrText = (await (0, ocrUtils_js_1.performOcr)(media.content, { language: config.ocrLanguage, ...config.ocrConfig })).trim();
                     }
                     catch (e) {
-                        (0, errorUtils_1.logWarning)(`OCR failed for ${attachment.name}:`, config, e);
+                        (0, errorUtils_js_1.logWarning)(`OCR failed for ${attachment.name}:`, config, e);
                     }
                 }
             }
         }
     }
     const metaFile = files.find(f => f.path.match(metaFileRegex));
-    const metadata = metaFile ? (0, xmlUtils_1.parseOfficeMetadata)(metaFile.content.toString()) : {};
+    const metadata = metaFile ? (0, xmlUtils_js_1.parseOfficeMetadata)(metaFile.content.toString()) : {};
     // Helper: Resolve ODS chart cell references to actual values
     // ODS charts often link to cell ranges (e.g., [Sheet1.$A$1:.$A$5]) instead of embedding values
     const resolveChartReferences = (chartData, nodes) => {