officeparser 6.0.7 → 6.1.1

package/dist/parsers/WordParser.js

@@ -40,6 +40,7 @@
  * - `<w:p>` - Paragraph
  * - `<w:r>` - Run (contiguous text with same formatting)
  * - `<w:t>` - Text content
+ * - `<w:br>` - Line or page break
  * - `<w:b>`, `<w:i>`, `<w:u>` - Bold, italic, underline
  * - `<w:pStyle>` - Paragraph style (for headings)
  * - `<w:numPr>` - List numbering properties
@@ -61,12 +62,11 @@
  */
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parseWord = void 0;
-const xmldom_1 = require("@xmldom/xmldom");
-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 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 a Word document (.docx) and extracts content, formatting, and metadata.
  *
@@ -90,13 +90,13 @@ const parseWord = async (buffer, config) => {
     const numberingFileRegex = /word\/numbering[\d+]?.xml/;
     const mediaFileRegex = /(word\/)?media\/.*/;
     const corePropsFileRegex = /docProps\/core[\d+]?.xml/;
+    const customPropsFileRegex = /docProps\/custom\.xml/;
     const relsFileRegex = /word\/_rels\/document[\d+]?.xml\.rels/;
     const stylesFileRegex = /word\/styles[\d+]?.xml/;
-    const xmlSerializer = new xmldom_1.XMLSerializer();
     // Helper to extract formatting from run properties XML string
     const extractFormattingFromXml = (rPr) => {
         const formatting = {};
-        const rPrString = xmlSerializer.serializeToString(rPr);
+        const rPrString = (0, xmlUtils_js_1.serializeXml)(rPr);
         // Helper to check boolean properties
         const getBoolVal = (xmlSnippet, tagName) => {
             const regex = new RegExp(`<${tagName}(?:\\s+w:val="([^"]+)")?\\s*\\/?>`);
@@ -133,7 +133,7 @@ const parseWord = async (buffer, config) => {
         // Font size
         const szMatch = rPrString.match(/<w:sz w:val="(\d+)"/);
         if (szMatch)
-            formatting.size = (parseInt(szMatch[1]) / 2).toString() + 'pt';
+            formatting.size = (parseInt(szMatch[1], 10) / 2).toString() + 'pt';
         // Color
         const colorMatch = rPrString.match(/<w:color w:val="([^"]+)"/);
         if (colorMatch && colorMatch[1] !== 'auto')
@@ -173,17 +173,45 @@ const parseWord = async (buffer, config) => {
         }
         return formatting;
     };
-    const files = await (0, zipUtils_1.extractFiles)(buffer, x => !!x.match(documentFileRegex) ||
+    // Helper to extract indentation from paragraph properties XML string
+    const extractIndentationFromXml = (pPr) => {
+        const ind = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "w:ind");
+        if (ind) {
+            const indentation = {};
+            const left = ind.getAttribute("w:left") || ind.getAttribute("w:start");
+            const right = ind.getAttribute("w:right") || ind.getAttribute("w:end");
+            const firstLine = ind.getAttribute("w:firstLine");
+            const hanging = ind.getAttribute("w:hanging");
+            if (left)
+                indentation.left = parseInt(left, 10);
+            if (right)
+                indentation.right = parseInt(right, 10);
+            if (firstLine)
+                indentation.firstLine = parseInt(firstLine, 10);
+            if (hanging)
+                indentation.hanging = parseInt(hanging, 10);
+            return Object.keys(indentation).length > 0 ? indentation : undefined;
+        }
+        return undefined;
+    };
+    const files = await (0, zipUtils_js_1.extractFiles)(buffer, x => !!x.match(documentFileRegex) ||
         !!x.match(footnotesFileRegex) ||
         !!x.match(endnotesFileRegex) ||
         !!x.match(numberingFileRegex) ||
         !!x.match(corePropsFileRegex) ||
+        !!x.match(customPropsFileRegex) ||
         !!x.match(relsFileRegex) ||
         !!x.match(stylesFileRegex) ||
         (!!config.extractAttachments && !!x.match(mediaFileRegex)));
     // Extract metadata
     const corePropsFile = files.find(f => f.path.match(corePropsFileRegex));
-    const metadata = corePropsFile ? (0, xmlUtils_1.parseOfficeMetadata)(corePropsFile.content.toString()) : {};
+    const metadata = corePropsFile ? (0, xmlUtils_js_1.parseOfficeMetadata)(corePropsFile.content.toString()) : {};
+    const customPropsFile = files.find(f => f.path.match(customPropsFileRegex));
+    if (customPropsFile) {
+        const customProperties = (0, xmlUtils_js_1.parseOOXMLCustomProperties)(customPropsFile.content.toString());
+        if (Object.keys(customProperties).length > 0)
+            metadata.customProperties = customProperties;
+    }
     const footnoteMap = new Map();
     const endnoteMap = new Map();
     const collectedNotes = [];
@@ -193,11 +221,11 @@ const parseWord = async (buffer, config) => {
     const relsFile = files.find(f => f.path.match(relsFileRegex));
     const relsMap = {};
     if (relsFile) {
-        const relsXml = (0, xmlUtils_1.parseXmlString)(relsFile.content.toString());
-        const relationships = (0, xmlUtils_1.getElementsByTagName)(relsXml, "Relationship");
-        for (let i = 0; i < relationships.length; i++) {
-            const id = relationships[i].getAttribute("Id");
-            const target = relationships[i].getAttribute("Target");
+        const relsXml = (0, xmlUtils_js_1.parseXmlString)(relsFile.content.toString());
+        const relationships = (0, xmlUtils_js_1.getElementsByTagName)(relsXml, "Relationship");
+        for (const relationship of relationships) {
+            const id = relationship.getAttribute("Id");
+            const target = relationship.getAttribute("Target");
             if (id && target) {
                 relsMap[id] = target;
             }
@@ -206,27 +234,27 @@ const parseWord = async (buffer, config) => {
     const numberingFile = files.find(f => f.path.match(numberingFileRegex));
     const numberingMap = {};
     if (numberingFile) {
-        const numberingXml = (0, xmlUtils_1.parseXmlString)(numberingFile.content.toString());
-        const nums = (0, xmlUtils_1.getElementsByTagName)(numberingXml, "w:num");
-        const abstractNums = (0, xmlUtils_1.getElementsByTagName)(numberingXml, "w:abstractNum");
+        const numberingXml = (0, xmlUtils_js_1.parseXmlString)(numberingFile.content.toString());
+        const nums = (0, xmlUtils_js_1.getElementsByTagName)(numberingXml, "w:num");
+        const abstractNums = (0, xmlUtils_js_1.getElementsByTagName)(numberingXml, "w:abstractNum");
         const abstractNumMap = {};
-        for (let i = 0; i < abstractNums.length; i++) {
-            const abstractNumId = abstractNums[i].getAttribute("w:abstractNumId");
+        for (const abstractNum of abstractNums) {
+            const abstractNumId = abstractNum.getAttribute("w:abstractNumId");
             if (abstractNumId) {
-                abstractNumMap[abstractNumId] = abstractNums[i];
+                abstractNumMap[abstractNumId] = abstractNum;
             }
         }
-        for (let i = 0; i < nums.length; i++) {
-            const numId = nums[i].getAttribute("w:numId");
-            const abstractNumIdNode = (0, xmlUtils_1.getElementsByTagName)(nums[i], "w:abstractNumId")[0];
+        for (const num of nums) {
+            const numId = num.getAttribute("w:numId");
+            const abstractNumIdNode = (0, xmlUtils_js_1.getFirstElementByTagName)(num, "w:abstractNumId");
             const abstractNumId = abstractNumIdNode?.getAttribute("w:val");
             if (numId && abstractNumId && abstractNumMap[abstractNumId]) {
                 numberingMap[numId] = {};
-                const lvls = (0, xmlUtils_1.getElementsByTagName)(abstractNumMap[abstractNumId], "w:lvl");
-                for (let j = 0; j < lvls.length; j++) {
-                    const ilvl = lvls[j].getAttribute("w:ilvl");
-                    const numFmtNode = (0, xmlUtils_1.getElementsByTagName)(lvls[j], "w:numFmt")[0];
-                    const lvlTextNode = (0, xmlUtils_1.getElementsByTagName)(lvls[j], "w:lvlText")[0];
+                const lvls = (0, xmlUtils_js_1.getElementsByTagName)(abstractNumMap[abstractNumId], "w:lvl");
+                for (const lvl of lvls) {
+                    const ilvl = lvl.getAttribute("w:ilvl");
+                    const numFmtNode = (0, xmlUtils_js_1.getFirstElementByTagName)(lvl, "w:numFmt");
+                    const lvlTextNode = (0, xmlUtils_js_1.getFirstElementByTagName)(lvl, "w:lvlText");
                     if (ilvl) {
                         numberingMap[numId][ilvl] = {
                             numFmt: numFmtNode?.getAttribute("w:val") || 'decimal',
@@ -241,44 +269,48 @@ const parseWord = async (buffer, config) => {
     const stylesFile = files.find(f => f.path.match(stylesFileRegex));
     const styleMap = {};
     if (stylesFile) {
-        const stylesXml = (0, xmlUtils_1.parseXmlString)(stylesFile.content.toString());
-        const styles = (0, xmlUtils_1.getElementsByTagName)(stylesXml, "w:style");
-        for (let i = 0; i < styles.length; i++) {
-            const styleId = styles[i].getAttribute("w:styleId");
+        const stylesXml = (0, xmlUtils_js_1.parseXmlString)(stylesFile.content.toString());
+        const styles = (0, xmlUtils_js_1.getElementsByTagName)(stylesXml, "w:style");
+        for (const style of styles) {
+            const styleId = style.getAttribute("w:styleId");
             if (styleId) {
-                const rPr = (0, xmlUtils_1.getElementsByTagName)(styles[i], "w:rPr")[0];
-                const pPr = (0, xmlUtils_1.getElementsByTagName)(styles[i], "w:pPr")[0];
+                const rPr = (0, xmlUtils_js_1.getFirstElementByTagName)(style, "w:rPr");
+                const pPr = (0, xmlUtils_js_1.getFirstElementByTagName)(style, "w:pPr");
                 const formatting = rPr ? extractFormattingFromXml(rPr) : {};
                 let alignment = undefined;
                 let backgroundColor = undefined;
+                let paragraphIndentation = undefined;
                 if (pPr) {
-                    const jc = (0, xmlUtils_1.getElementsByTagName)(pPr, "w:jc")[0];
+                    const jc = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "w:jc");
                     if (jc) {
                         const val = jc.getAttribute("w:val");
                         if (val === 'left' || val === 'center' || val === 'right' || val === 'justify') {
                             alignment = val;
                         }
                     }
-                    const shd = (0, xmlUtils_1.getElementsByTagName)(pPr, "w:shd")[0];
+                    const shd = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "w:shd");
                     if (shd) {
                         const fill = shd.getAttribute("w:fill");
                         if (fill && fill !== 'auto')
                             backgroundColor = '#' + fill;
                     }
+                    const ind = extractIndentationFromXml(pPr);
+                    if (ind)
+                        paragraphIndentation = ind;
                 }
-                styleMap[styleId] = { formatting, alignment, backgroundColor };
+                styleMap[styleId] = { formatting, alignment, backgroundColor, paragraphIndentation };
             }
         }
     }
     // Extract document defaults
     let docDefaults = {};
     if (stylesFile) {
-        const stylesXml = (0, xmlUtils_1.parseXmlString)(stylesFile.content.toString());
-        const docDefaultsNode = (0, xmlUtils_1.getElementsByTagName)(stylesXml, "w:docDefaults")[0];
+        const stylesXml = (0, xmlUtils_js_1.parseXmlString)(stylesFile.content.toString());
+        const docDefaultsNode = (0, xmlUtils_js_1.getFirstElementByTagName)(stylesXml, "w:docDefaults");
         if (docDefaultsNode) {
-            const rPrDefaultNode = (0, xmlUtils_1.getElementsByTagName)(docDefaultsNode, "w:rPrDefault")[0];
+            const rPrDefaultNode = (0, xmlUtils_js_1.getFirstElementByTagName)(docDefaultsNode, "w:rPrDefault");
             if (rPrDefaultNode) {
-                const rPr = (0, xmlUtils_1.getElementsByTagName)(rPrDefaultNode, "w:rPr")[0];
+                const rPr = (0, xmlUtils_js_1.getFirstElementByTagName)(rPrDefaultNode, "w:rPr");
                 if (rPr) {
                     docDefaults = extractFormattingFromXml(rPr);
                 }
@@ -288,13 +320,13 @@ const parseWord = async (buffer, config) => {
     // Detect the default paragraph style (for international compatibility)
     let defaultParaStyleId = undefined;
     if (stylesFile) {
-        const stylesXml = (0, xmlUtils_1.parseXmlString)(stylesFile.content.toString());
-        const styles = (0, xmlUtils_1.getElementsByTagName)(stylesXml, "w:style");
+        const stylesXml = (0, xmlUtils_js_1.parseXmlString)(stylesFile.content.toString());
+        const styles = (0, xmlUtils_js_1.getElementsByTagName)(stylesXml, "w:style");
         // Look for a style with w:type="paragraph" and w:default="1"
-        for (let i = 0; i < styles.length; i++) {
-            const styleType = styles[i].getAttribute("w:type");
-            const isDefault = styles[i].getAttribute("w:default");
-            const styleId = styles[i].getAttribute("w:styleId");
+        for (const style of styles) {
+            const styleType = style.getAttribute("w:type");
+            const isDefault = style.getAttribute("w:default");
+            const styleId = style.getAttribute("w:styleId");
             if (styleType === "paragraph" && isDefault === "1" && styleId) {
                 defaultParaStyleId = styleId;
                 break;
@@ -310,22 +342,22 @@ const parseWord = async (buffer, config) => {
     const numberingState = {};
     const listCounters = {}; // Track item index per listId/level
     // Helper to parse a paragraph node
-    const parseParagraph = (pNode) => {
-        const pXml = xmlSerializer.serializeToString(pNode);
+    const parseParagraph = (pNode, documentContent) => {
+        const pXml = pNode.toString();
         // Check if it's a list item
-        const numPr = (0, xmlUtils_1.getElementsByTagName)(pNode, "w:numPr")[0];
+        const numPr = (0, xmlUtils_js_1.getFirstElementByTagName)(pNode, "w:numPr");
         const isList = !!numPr;
         // Check if it's a heading
-        const pPr = (0, xmlUtils_1.getElementsByTagName)(pNode, "w:pPr")[0];
-        const pStyle = pPr ? (0, xmlUtils_1.getElementsByTagName)(pPr, "w:pStyle")[0] : null;
-        const pStyleVal = pStyle ? pStyle.getAttribute("w:val") : null;
+        const pPr = (0, xmlUtils_js_1.getFirstElementByTagName)(pNode, "w:pPr");
+        const pStyle = pPr ? (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "w:pStyle") : null;
+        const pStyleVal = pStyle?.getAttribute("w:val");
         const isHeading = pStyleVal ? (pStyleVal.startsWith("Heading") || pStyleVal === "Title") : false;
         // Extract Paragraph Style Properties
         const styleProps = pStyleVal && styleMap[pStyleVal] ? styleMap[pStyleVal] : { formatting: {} };
         // Extract Alignment
         let alignment = styleProps.alignment;
         if (pPr) {
-            const jc = (0, xmlUtils_1.getElementsByTagName)(pPr, "w:jc")[0];
+            const jc = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "w:jc");
             if (jc) {
                 const val = jc.getAttribute("w:val");
                 if (val === 'left' || val === 'center' || val === 'right' || val === 'justify') {
@@ -333,10 +365,18 @@ const parseWord = async (buffer, config) => {
                 }
             }
         }
+        // Extract Indentation
+        let paraIndentation = styleProps.paragraphIndentation;
+        if (pPr) {
+            const ind = extractIndentationFromXml(pPr);
+            if (ind) {
+                paraIndentation = { ...paraIndentation, ...ind };
+            }
+        }
         // Extract Paragraph Background
         let paraBackgroundColor = styleProps.backgroundColor;
         if (pPr) {
-            const shd = (0, xmlUtils_1.getElementsByTagName)(pPr, "w:shd")[0];
+            const shd = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "w:shd");
             if (shd) {
                 const fill = shd.getAttribute("w:fill");
                 if (fill && fill !== 'auto') {
@@ -347,7 +387,7 @@ const parseWord = async (buffer, config) => {
         // Extract paragraph-level run properties
         let paragraphRunFormatting = { ...styleProps.formatting };
         if (pPr) {
-            const pPrRPr = (0, xmlUtils_1.getElementsByTagName)(pPr, "w:rPr")[0];
+            const pPrRPr = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "w:rPr");
             if (pPrRPr) {
                 const pPrFormatting = extractFormattingFromXml(pPrRPr);
                 for (const key in pPrFormatting) {
@@ -366,9 +406,9 @@ const parseWord = async (buffer, config) => {
         const children = [];
         // Traverse children of paragraph (runs, hyperlinks, etc.)
         const processChildNode = (node) => {
-            if (node.nodeName === 'w:r') {
+            if ((0, xmlUtils_js_1.isElement)(node) && node.nodeName === 'w:r') {
                 const runNode = node;
-                const rPr = (0, xmlUtils_1.getElementsByTagName)(runNode, "w:rPr")[0];
+                const rPr = (0, xmlUtils_js_1.getFirstElementByTagName)(runNode, "w:rPr");
                 // Formatting
                 let formatting = {};
                 // Apply paragraph-level formatting
@@ -376,7 +416,7 @@ const parseWord = async (buffer, config) => {
                     formatting[key] = paragraphRunFormatting[key];
                 }
                 // Check for run style
-                const rStyle = rPr ? (0, xmlUtils_1.getElementsByTagName)(rPr, "w:rStyle")[0] : null;
+                const rStyle = rPr ? (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "w:rStyle") : null;
                 const rStyleVal = rStyle ? rStyle.getAttribute("w:val") : pStyleVal;
                 if (rStyleVal && styleMap[rStyleVal]) {
                     for (const key in styleMap[rStyleVal].formatting) {
@@ -400,48 +440,93 @@ const parseWord = async (buffer, config) => {
                 if (!formatting.backgroundColor && paraBackgroundColor) {
                     formatting.backgroundColor = paraBackgroundColor;
                 }
-                // Text content
-                const tNodes = (0, xmlUtils_1.getElementsByTagName)(runNode, "w:t");
-                for (const tNode of tNodes) {
-                    const tContent = tNode.textContent || '';
-                    text += tContent;
-                    const textNode = {
-                        type: 'text',
-                        text: tContent,
-                        formatting: formatting
-                    };
-                    if (config.includeRawContent) {
-                        textNode.rawContent = xmlSerializer.serializeToString(tNode);
+                for (const child of runNode.childNodes) {
+                    if (!(0, xmlUtils_js_1.isElement)(child))
+                        continue;
+                    // also handle unprefixed version (mirroring the behaviour of getElementsByTagName)
+                    // Text content
+                    if (child.tagName === "w:t" || child.tagName === "t") {
+                        const tNode = child;
+                        const tContent = tNode.textContent || '';
+                        text += tContent;
+                        const textNode = {
+                            type: 'text',
+                            text: tContent,
+                            formatting: formatting
+                        };
+                        if (config.includeRawContent) {
+                            textNode.rawContent = (0, xmlUtils_js_1.getRawContent)(tNode, documentContent, config);
+                        }
+                        // Always set a style: run style > paragraph style > detected default
+                        // Use detected default style for international compatibility
+                        const nodeStyle = rStyleVal || pStyleVal || defaultParaStyleId;
+                        if (nodeStyle) {
+                            textNode.metadata = { style: nodeStyle };
+                        }
+                        children.push(textNode);
+                    }
+                    // Break nodes
+                    else if (config.includeBreakNodes &&
+                        (child.tagName === "w:br"
+                            || child.tagName === "br"
+                            || child.tagName === "w:cr"
+                            || child.tagName === "cr")) {
+                        const brNode = child;
+                        let breakType = 'textWrapping';
+                        if (child.tagName === "w:cr" || child.tagName === "cr") {
+                            breakType = 'carriageReturn';
+                        }
+                        else {
+                            const nodeBreakType = brNode.getAttribute("w:type") || brNode.getAttribute("type");
+                            if (nodeBreakType !== null) {
+                                breakType = nodeBreakType;
+                            }
+                        }
+                        let breakClear = undefined;
+                        if (breakType === 'textWrapping' && brNode.getAttribute("w:clear") !== null) {
+                            breakClear = brNode.getAttribute("w:clear");
+                        }
+                        const breakNode = {
+                            type: 'break',
+                            metadata: { breakType, clear: breakClear }
+                        };
+                        if (config.includeRawContent) {
+                            breakNode.rawContent = (0, xmlUtils_js_1.getRawContent)(brNode, documentContent, config);
+                        }
+                        children.push(breakNode);
                     }
-                    // Always set a style: run style > paragraph style > detected default
-                    // Use detected default style for international compatibility
-                    const nodeStyle = rStyleVal || pStyleVal || defaultParaStyleId;
-                    if (nodeStyle) {
-                        textNode.metadata = { style: nodeStyle };
+                    else if (config.includeBreakNodes && (child.tagName === "w:lastRenderedPageBreak" || child.tagName === "lastRenderedPageBreak")) {
+                        const breakNode = {
+                            type: 'break',
+                            metadata: { breakType: 'lastRenderedPage' }
+                        };
+                        if (config.includeRawContent) {
+                            breakNode.rawContent = (0, xmlUtils_js_1.getRawContent)(child, documentContent, config);
+                        }
+                        children.push(breakNode);
                     }
-                    children.push(textNode);
                 }
                 // Images/Drawings
                 if (config.extractAttachments) {
-                    const drawings = (0, xmlUtils_1.getElementsByTagName)(runNode, "w:drawing");
-                    const picts = (0, xmlUtils_1.getElementsByTagName)(runNode, "w:pict");
+                    const drawings = (0, xmlUtils_js_1.getElementsByTagName)(runNode, "w:drawing");
+                    const picts = (0, xmlUtils_js_1.getElementsByTagName)(runNode, "w:pict");
                     const allImages = [...drawings, ...picts];
                     for (const imgNode of allImages) {
-                        const imgXml = xmlSerializer.serializeToString(imgNode);
+                        const imgXml = (0, xmlUtils_js_1.serializeXml)(imgNode);
                         // Extract Alt Text
                         let altText = '';
-                        const docPr = (0, xmlUtils_1.getElementsByTagName)(imgNode, "wp:docPr")[0];
+                        const docPr = (0, xmlUtils_js_1.getFirstElementByTagName)(imgNode, "wp:docPr");
                         if (docPr) {
                             altText = docPr.getAttribute("descr") || docPr.getAttribute("title") || '';
                         }
                         // Extract Relationship ID
                         let rId = '';
-                        const blip = (0, xmlUtils_1.getElementsByTagName)(imgNode, "a:blip")[0];
+                        const blip = (0, xmlUtils_js_1.getFirstElementByTagName)(imgNode, "a:blip");
                         if (blip) {
                             rId = blip.getAttribute("r:embed") || '';
                         }
                         else {
-                            const imagedata = (0, xmlUtils_1.getElementsByTagName)(imgNode, "v:imagedata")[0];
+                            const imagedata = (0, xmlUtils_js_1.getFirstElementByTagName)(imgNode, "v:imagedata");
                             if (imagedata) {
                                 rId = imagedata.getAttribute("r:id") || '';
                             }
@@ -456,7 +541,7 @@ const parseWord = async (buffer, config) => {
                                     metadata: { attachmentName: filename, altText: altText }
                                 };
                                 if (config.includeRawContent) {
-                                    imageNode.rawContent = imgXml;
+                                    imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(imgNode, documentContent, config);
                                 }
                                 children.push(imageNode);
                             }
@@ -467,7 +552,7 @@ const parseWord = async (buffer, config) => {
                                 text: '',
                             };
                             if (config.includeRawContent) {
-                                imageNode.rawContent = imgXml;
+                                imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(imgNode, documentContent, config);
                             }
                             children.push(imageNode);
                         }
@@ -475,7 +560,7 @@ const parseWord = async (buffer, config) => {
                 }
                 // Footnotes/Endnotes inside runs
                 if (!config.ignoreNotes) {
-                    const footnoteRef = (0, xmlUtils_1.getElementsByTagName)(runNode, "w:footnoteReference")[0];
+                    const footnoteRef = (0, xmlUtils_js_1.getFirstElementByTagName)(runNode, "w:footnoteReference");
                     if (footnoteRef) {
                         const id = footnoteRef.getAttribute("w:id");
                         if (id && footnoteMap.has(id)) {
@@ -494,7 +579,7 @@ const parseWord = async (buffer, config) => {
                             }
                         }
                     }
-                    const endnoteRef = (0, xmlUtils_1.getElementsByTagName)(runNode, "w:endnoteReference")[0];
+                    const endnoteRef = (0, xmlUtils_js_1.getFirstElementByTagName)(runNode, "w:endnoteReference");
                     if (endnoteRef) {
                         const id = endnoteRef.getAttribute("w:id");
                         if (id && endnoteMap.has(id)) {
@@ -515,7 +600,7 @@ const parseWord = async (buffer, config) => {
                     }
                 }
             }
-            else if (node.nodeName === 'w:hyperlink') {
+            else if ((0, xmlUtils_js_1.isElement)(node) && node.nodeName === 'w:hyperlink') {
                 const hlNode = node;
                 const rId = hlNode.getAttribute("r:id");
                 const anchor = hlNode.getAttribute("w:anchor");
@@ -548,10 +633,10 @@ const parseWord = async (buffer, config) => {
             processChildNode(child);
         }
         if (isList) {
-            const numIdNode = (0, xmlUtils_1.getElementsByTagName)(numPr, "w:numId")[0];
-            const ilvlNode = (0, xmlUtils_1.getElementsByTagName)(numPr, "w:ilvl")[0];
+            const numIdNode = (0, xmlUtils_js_1.getFirstElementByTagName)(numPr, "w:numId");
+            const ilvlNode = (0, xmlUtils_js_1.getFirstElementByTagName)(numPr, "w:ilvl");
             const numId = numIdNode ? numIdNode.getAttribute("w:val") || '0' : '0';
-            const ilvl = ilvlNode ? parseInt(ilvlNode.getAttribute("w:val") || '0') : 0;
+            const ilvl = ilvlNode ? parseInt(ilvlNode.getAttribute("w:val") || '0', 10) : 0;
             let listType = 'ordered';
             let itemIndex = 0;
             if (numId && numberingMap[numId]) {
@@ -585,6 +670,7 @@ const parseWord = async (buffer, config) => {
                 metadata: {
                     listType,
                     indentation: ilvl,
+                    paragraphIndentation: paraIndentation,
                     alignment: (alignment || 'left'),
                     listId: numId,
                     itemIndex: itemIndex,
@@ -592,19 +678,19 @@ const parseWord = async (buffer, config) => {
                 }
             };
             if (config.includeRawContent)
-                listNode.rawContent = pXml;
+                listNode.rawContent = (0, xmlUtils_js_1.getRawContent)(pNode, documentContent, config);
             return listNode;
         }
         else if (isHeading) {
-            const level = pStyleVal ? parseInt(pStyleVal.replace("Heading", "")) || 1 : 1;
+            const level = pStyleVal ? parseInt(pStyleVal.replace("Heading", ""), 10) || 1 : 1;
             const headingNode = {
                 type: 'heading',
                 text: text,
                 children: children,
-                metadata: { level, alignment, style: pStyleVal ?? undefined }
+                metadata: { level, alignment, paragraphIndentation: paraIndentation, style: pStyleVal ?? undefined }
             };
             if (config.includeRawContent)
-                headingNode.rawContent = pXml;
+                headingNode.rawContent = (0, xmlUtils_js_1.getRawContent)(pNode, documentContent, config);
             return headingNode;
         }
         else {
@@ -612,23 +698,23 @@ const parseWord = async (buffer, config) => {
                 type: 'paragraph',
                 text: text,
                 children: children,
-                metadata: { alignment, style: pStyleVal ?? undefined }
+                metadata: { alignment, paragraphIndentation: paraIndentation, style: pStyleVal ?? undefined }
             };
             if (config.includeRawContent)
-                paraNode.rawContent = pXml;
+                paraNode.rawContent = (0, xmlUtils_js_1.getRawContent)(pNode, documentContent, config);
             return paraNode;
         }
     };
     // Helper to parse a table node
-    const parseTable = (tblNode) => {
+    const parseTable = (tblNode, documentContent) => {
         const rows = [];
         // Only get direct child rows, not nested table rows
-        const trNodes = (0, xmlUtils_1.getDirectChildren)(tblNode, "w:tr");
+        const trNodes = (0, xmlUtils_js_1.getDirectChildren)(tblNode, "w:tr");
         for (let rIndex = 0; rIndex < trNodes.length; rIndex++) {
             const trNode = trNodes[rIndex];
             const cells = [];
             // Only get direct child cells, not nested table cells
-            const tcNodes = (0, xmlUtils_1.getDirectChildren)(trNode, "w:tc");
+            const tcNodes = (0, xmlUtils_js_1.getDirectChildren)(trNode, "w:tc");
             for (let cIndex = 0; cIndex < tcNodes.length; cIndex++) {
                 const tcNode = tcNodes[cIndex];
                 const cellChildren = [];
@@ -636,14 +722,14 @@ const parseWord = async (buffer, config) => {
                 // Cells contain paragraphs (and other block-level elements)
                 const cellContentNodes = Array.from(tcNode.childNodes);
                 for (const child of cellContentNodes) {
-                    if (child.nodeName === 'w:p') {
-                        const pNode = parseParagraph(child);
+                    if ((0, xmlUtils_js_1.isElement)(child) && child.nodeName === 'w:p') {
+                        const pNode = parseParagraph(child, documentContent);
                         cellChildren.push(pNode);
                         cellText += pNode.text;
                     }
-                    else if (child.nodeName === 'w:tbl') {
+                    else if ((0, xmlUtils_js_1.isElement)(child) && child.nodeName === 'w:tbl') {
                         // Nested table
-                        const nestedTable = parseTable(child);
+                        const nestedTable = parseTable(child, documentContent);
                         cellChildren.push(nestedTable);
                         // Don't add nested table text to cell text - it will be handled recursively
                     }
@@ -671,26 +757,28 @@ const parseWord = async (buffer, config) => {
     if (!config.ignoreNotes) {
         const footnotesFile = files.find(f => f.path.match(footnotesFileRegex));
         if (footnotesFile) {
-            const footnotesDoc = (0, xmlUtils_1.parseXmlString)(footnotesFile.content.toString());
-            const footnoteNodes = (0, xmlUtils_1.getElementsByTagName)(footnotesDoc, "w:footnote");
+            const footnotesDoc = (0, xmlUtils_js_1.parseXmlString)(footnotesFile.content.toString());
+            const footnoteXml = footnotesFile.content.toString();
+            const footnoteNodes = (0, xmlUtils_js_1.getElementsByTagName)(footnotesDoc, "w:footnote");
             for (const node of footnoteNodes) {
                 const id = node.getAttribute("w:id");
                 if (!id || id === "-1" || id === "0")
                     continue;
-                const pNodes = (0, xmlUtils_1.getElementsByTagName)(node, "w:p");
-                footnoteMap.set(id, pNodes.map(p => parseParagraph(p)));
+                const pNodes = (0, xmlUtils_js_1.getElementsByTagName)(node, "w:p");
+                footnoteMap.set(id, pNodes.map(p => parseParagraph(p, footnoteXml)));
             }
         }
         const endnotesFile = files.find(f => f.path.match(endnotesFileRegex));
         if (endnotesFile) {
-            const endnotesDoc = (0, xmlUtils_1.parseXmlString)(endnotesFile.content.toString());
-            const endnoteNodes = (0, xmlUtils_1.getElementsByTagName)(endnotesDoc, "w:endnote");
+            const endnotesDoc = (0, xmlUtils_js_1.parseXmlString)(endnotesFile.content.toString());
+            const endnoteXml = endnotesFile.content.toString();
+            const endnoteNodes = (0, xmlUtils_js_1.getElementsByTagName)(endnotesDoc, "w:endnote");
             for (const node of endnoteNodes) {
                 const id = node.getAttribute("w:id");
                 if (!id || id === "-1" || id === "0")
                     continue;
-                const pNodes = (0, xmlUtils_1.getElementsByTagName)(node, "w:p");
-                endnoteMap.set(id, pNodes.map(p => parseParagraph(p)));
+                const pNodes = (0, xmlUtils_js_1.getElementsByTagName)(node, "w:p");
+                endnoteMap.set(id, pNodes.map(p => parseParagraph(p, endnoteXml)));
             }
         }
     }
@@ -711,16 +799,16 @@ const parseWord = async (buffer, config) => {
         if (config.includeRawContent) {
             rawContents.push(documentContent);
         }
-        const doc = (0, xmlUtils_1.parseXmlString)(documentContent);
-        const body = (0, xmlUtils_1.getElementsByTagName)(doc, "w:body")[0];
+        const doc = (0, xmlUtils_js_1.parseXmlString)(documentContent, { locator: config.includeRawContent });
+        const body = (0, xmlUtils_js_1.getFirstElementByTagName)(doc, "w:body");
         if (body) {
             const bodyChildren = Array.from(body.childNodes);
             for (const child of bodyChildren) {
-                if (child.nodeName === 'w:p') {
-                    content.push(parseParagraph(child));
+                if ((0, xmlUtils_js_1.isElement)(child) && child.nodeName === 'w:p') {
+                    content.push(parseParagraph(child, documentContent));
                 }
-                else if (child.nodeName === 'w:tbl') {
-                    content.push(parseTable(child));
+                else if ((0, xmlUtils_js_1.isElement)(child) && child.nodeName === 'w:tbl') {
+                    content.push(parseTable(child, documentContent));
                 }
             }
         }
@@ -728,15 +816,15 @@ const parseWord = async (buffer, config) => {
     // Extract attachments
     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);
                     }
                 }
             }
@@ -776,6 +864,9 @@ const parseWord = async (buffer, config) => {
                 if (node.children) {
                     t += node.children.map(getText).filter(t => t != '').join(!node.children[0]?.children ? '' : config.newlineDelimiter ?? '\n');
                 }
+                else if (node.type === 'break') {
+                    t += config.newlineDelimiter ?? '\n';
+                }
                 else
                     t += node.text || '';
                 return t;