officeparser 7.2.2 → 7.2.3

package/dist/parsers/OpenOfficeParser.js

@@ -31,6 +31,16 @@ 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");
+/**
+ * Helper to clean and extract attachment name from xlink:href or paths.
+ * Handles trailing slashes, leading "./", and subdirectories.
+ */
+const cleanAttachmentName = (href) => {
+    if (!href)
+        return '';
+    const cleaned = href.replace(/^\.\//, '').replace(/\/$/, '');
+    return cleaned.split('/').pop() || '';
+};
 /**
  * Parses an OpenOffice document (.odt, .odp, .ods) and extracts content.
  *
@@ -82,6 +92,7 @@ const parseOpenOffice = async (buffer, config) => {
     let lastListStyle = null;
     let listIdCounter = 0;
     let lastWasList = false;
+    let traverse;
     // Helper to parse styles
     const parseStyles = (xml) => {
         const styles = (0, xmlUtils_js_1.getElementsByTagName)(xml, "style:style");
@@ -170,6 +181,64 @@ const parseOpenOffice = async (buffer, config) => {
      * Returns the paragraph content without creating a content node.
      *
      * @param node - The paragraph element to parse
+     */
+    const parseMathML = (node) => {
+        if (!node)
+            return '';
+        if (node.nodeType === 3) { // Text node
+            return node.textContent || '';
+        }
+        if (node.nodeType !== 1) { // Not an element
+            return '';
+        }
+        const element = node;
+        const tagName = element.tagName.toLowerCase().replace(/^.*:/, ''); // strip namespace prefix
+        switch (tagName) {
+            case 'math':
+            case 'mrow':
+            case 'semantics':
+                return Array.from(element.childNodes).map(parseMathML).join('');
+            case 'mfrac': {
+                const children = Array.from(element.childNodes).filter((n) => n.nodeType === 1);
+                if (children.length >= 2) {
+                    return `(${parseMathML(children[0])})/(${parseMathML(children[1])})`;
+                }
+                return Array.from(element.childNodes).map(parseMathML).join('');
+            }
+            case 'msub': {
+                const children = Array.from(element.childNodes).filter((n) => n.nodeType === 1);
+                if (children.length >= 2) {
+                    return `${parseMathML(children[0])}_${parseMathML(children[1])}`;
+                }
+                return Array.from(element.childNodes).map(parseMathML).join('');
+            }
+            case 'msup': {
+                const children = Array.from(element.childNodes).filter((n) => n.nodeType === 1);
+                if (children.length >= 2) {
+                    return `${parseMathML(children[0])}^${parseMathML(children[1])}`;
+                }
+                return Array.from(element.childNodes).map(parseMathML).join('');
+            }
+            case 'msubsup': {
+                const children = Array.from(element.childNodes).filter((n) => n.nodeType === 1);
+                if (children.length >= 3) {
+                    return `${parseMathML(children[0])}_${parseMathML(children[1])}^${parseMathML(children[2])}`;
+                }
+                return Array.from(element.childNodes).map(parseMathML).join('');
+            }
+            case 'mi':
+            case 'mn':
+            case 'mo':
+            case 'mtext':
+            case 'ms':
+                return element.textContent || '';
+            case 'annotation':
+                return '';
+            default:
+                return Array.from(element.childNodes).map(parseMathML).join('');
+        }
+    };
+    /**
      * Helper to parse inline content (text, spans, links, notes, etc.) recursively.
      *
      * @param node - The element to parse (paragraph, span, or link)
@@ -325,41 +394,114 @@ const parseOpenOffice = async (buffer, config) => {
                     }
                 }
                 else if (tagName === 'draw:frame') {
-                    // Inline image
                     const frame = element;
-                    // Extract alt text
-                    let altText = '';
-                    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;
+                    const drawTextBox = (0, xmlUtils_js_1.getFirstElementByTagName)(frame, "draw:text-box");
+                    const drawObject = (0, xmlUtils_js_1.getFirstElementByTagName)(frame, "draw:object");
+                    if (drawTextBox) {
+                        const textBoxChildren = [];
+                        traverse(drawTextBox, textBoxChildren, false, sourceXml);
+                        children.push(...textBoxChildren);
+                        const textBoxText = textBoxChildren.map(c => c.text || '').join('\n');
+                        fullText += textBoxText;
                     }
-                    else if (svgDesc && svgDesc.textContent) {
-                        altText = svgDesc.textContent;
-                    }
-                    // Extract image href
-                    let imageHref = '';
-                    const drawImages = (0, xmlUtils_js_1.getElementsByTagName)(frame, "draw:image");
-                    if (drawImages.length > 0) {
-                        imageHref = drawImages[0].getAttribute("xlink:href") || '';
-                        if (imageHref) {
-                            const parts = imageHref.split('/');
-                            imageHref = parts[parts.length - 1];
+                    else if (drawObject) {
+                        const href = drawObject.getAttribute("xlink:href");
+                        let isFormula = false;
+                        let formulaText = '';
+                        let attachmentName = '';
+                        if (href) {
+                            attachmentName = cleanAttachmentName(href);
+                            const objectPath = `${attachmentName}/content.xml`;
+                            const objectFile = files.find(f => f.path === objectPath || f.path.endsWith(objectPath));
+                            if (objectFile) {
+                                const objXml = (0, xmlUtils_js_1.parseXmlString)(objectFile.content.toString());
+                                const mathNode = (0, xmlUtils_js_1.getFirstElementByTagName)(objXml, "math");
+                                if (mathNode) {
+                                    isFormula = true;
+                                    formulaText = parseMathML(mathNode).trim();
+                                }
+                            }
+                        }
+                        if (isFormula) {
+                            fullText += formulaText;
+                            const textNode = {
+                                type: 'text',
+                                text: formulaText,
+                                formatting: parentFormatting,
+                                metadata: linkMetadata ? { ...linkMetadata } : undefined
+                            };
+                            if (config.includeRawContent) {
+                                textNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frame, sourceXml, config);
+                            }
+                            children.push(textNode);
+                        }
+                        else {
+                            // Standard inline image extraction fallback if object is not a formula
+                            let altText = '';
+                            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;
+                            }
+                            else if (svgDesc && svgDesc.textContent) {
+                                altText = svgDesc.textContent;
+                            }
+                            let imageHref = '';
+                            const drawImages = (0, xmlUtils_js_1.getElementsByTagName)(frame, "draw:image");
+                            if (drawImages.length > 0) {
+                                imageHref = drawImages[0].getAttribute("xlink:href") || '';
+                                if (imageHref) {
+                                    imageHref = cleanAttachmentName(imageHref);
+                                }
+                            }
+                            const imageNode = {
+                                type: 'image',
+                                text: '',
+                                children: [],
+                                metadata: {
+                                    attachmentName: imageHref || attachmentName,
+                                    ...(altText ? { altText } : {})
+                                }
+                            };
+                            if (config.includeRawContent) {
+                                imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frame, sourceXml, config);
+                            }
+                            children.push(imageNode);
                         }
                     }
-                    const imageNode = {
-                        type: 'image',
-                        text: '',
-                        children: [],
-                        metadata: {
-                            attachmentName: imageHref,
-                            ...(altText ? { altText } : {})
+                    else {
+                        // Standard inline image extraction fallback
+                        let altText = '';
+                        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;
                         }
-                    };
-                    if (config.includeRawContent) {
-                        imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frame, sourceXml, config);
+                        else if (svgDesc && svgDesc.textContent) {
+                            altText = svgDesc.textContent;
+                        }
+                        let imageHref = '';
+                        const drawImages = (0, xmlUtils_js_1.getElementsByTagName)(frame, "draw:image");
+                        if (drawImages.length > 0) {
+                            imageHref = drawImages[0].getAttribute("xlink:href") || '';
+                            if (imageHref) {
+                                imageHref = cleanAttachmentName(imageHref);
+                            }
+                        }
+                        const imageNode = {
+                            type: 'image',
+                            text: '',
+                            children: [],
+                            metadata: {
+                                attachmentName: imageHref,
+                                ...(altText ? { altText } : {})
+                            }
+                        };
+                        if (config.includeRawContent) {
+                            imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frame, sourceXml, config);
+                        }
+                        children.push(imageNode);
                     }
-                    children.push(imageNode);
                 }
             }
         }
@@ -713,7 +855,7 @@ const parseOpenOffice = async (buffer, config) => {
          * @param sourceXml - The source XML string for raw content extraction
          * @param asSheet - If true, treats tables as sheets (for ODS)
          */
-        function traverse(node, targetArray, forceHeading = false, sourceXml, asSheet = false) {
+        traverse = (node, targetArray, forceHeading = false, sourceXml = '', asSheet = false) => {
             if (node.tagName === "text:p") {
                 const pContent = parseParagraphContent(node, paragraphStyleMap, styleMap, config, sourceXml);
                 const type = (forceHeading || (node.getAttribute("text:style-name") || '').toLowerCase().includes('title')) ? 'heading' : 'paragraph';
@@ -1011,8 +1153,7 @@ const parseOpenOffice = async (buffer, config) => {
                     // Extract image href to link to attachment
                     let imageHref = image.getAttribute("xlink:href") || '';
                     if (imageHref) {
-                        const parts = imageHref.split('/');
-                        imageHref = parts[parts.length - 1];
+                        imageHref = cleanAttachmentName(imageHref);
                     }
                     const metadata = {
                         attachmentName: imageHref,
@@ -1034,25 +1175,47 @@ const parseOpenOffice = async (buffer, config) => {
                     targetArray.push(imageNode);
                 }
                 else if (object) {
-                    // Handle embedded objects like charts
+                    // Handle embedded objects like charts or math formulas
                     const href = object.getAttribute("xlink:href");
                     if (href) {
-                        const attachmentName = href.split('/')[0];
+                        const attachmentName = cleanAttachmentName(href);
                         const objectPath = `${attachmentName}/content.xml`;
                         const objectFile = files.find(f => f.path === objectPath || f.path.endsWith(objectPath));
                         if (objectFile) {
-                            const chartData = (0, chartUtils_js_1.extractChartData)(objectFile.content);
-                            const chartNode = {
-                                type: 'chart',
-                                text: chartData.rawTexts.join(" "),
-                                metadata: {
-                                    attachmentName: attachmentName,
-                                    chartData
+                            const objXml = (0, xmlUtils_js_1.parseXmlString)(objectFile.content.toString());
+                            const mathNode = (0, xmlUtils_js_1.getFirstElementByTagName)(objXml, "math");
+                            if (mathNode) {
+                                // Math formula object at block level
+                                const formulaText = parseMathML(mathNode).trim();
+                                const formulaNode = {
+                                    type: 'paragraph',
+                                    text: formulaText,
+                                    children: [
+                                        {
+                                            type: 'text',
+                                            text: formulaText
+                                        }
+                                    ]
+                                };
+                                if (config.includeRawContent) {
+                                    formulaNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
                                 }
-                            };
-                            if (config.includeRawContent)
-                                chartNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
-                            targetArray.push(chartNode);
+                                targetArray.push(formulaNode);
+                            }
+                            else {
+                                const chartData = (0, chartUtils_js_1.extractChartData)(objectFile.content);
+                                const chartNode = {
+                                    type: 'chart',
+                                    text: chartData.rawTexts.join(" "),
+                                    metadata: {
+                                        attachmentName: attachmentName,
+                                        chartData
+                                    }
+                                };
+                                if (config.includeRawContent)
+                                    chartNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
+                                targetArray.push(chartNode);
+                            }
                         }
                         else {
                             const chartNode = {
@@ -1077,8 +1240,7 @@ const parseOpenOffice = async (buffer, config) => {
                     }
                 }
             }
-        }
-        ;
+        };
         // ODS: Spreadsheet
         if (fileType === 'ods') {
             const spreadsheet = (0, xmlUtils_js_1.getFirstElementByTagName)(body, "office:spreadsheet");
@@ -1156,28 +1318,50 @@ const parseOpenOffice = async (buffer, config) => {
                                 if (drawImages.length > 0) {
                                     const rawHref = drawImages[0].getAttribute("xlink:href");
                                     if (rawHref) {
-                                        const parts = rawHref.split('/');
-                                        imageHref = parts[parts.length - 1];
+                                        imageHref = cleanAttachmentName(rawHref);
                                     }
                                 }
-                                // Extract chart object href
+                                // Extract chart or math object href
                                 let chartHref = '';
+                                let isFormula = false;
+                                let formulaText = '';
                                 const drawObjects = (0, xmlUtils_js_1.getElementsByTagName)(frame, "draw:object");
                                 if (drawObjects.length > 0) {
                                     const href = drawObjects[0].getAttribute("xlink:href");
                                     if (href) {
-                                        // Object href is usually "./Object 1"
-                                        chartHref = href.split('/')[0];
+                                        chartHref = cleanAttachmentName(href);
+                                        const objectPath = `${chartHref}/content.xml`;
+                                        const objectFile = files.find(f => f.path === objectPath || f.path.endsWith(objectPath));
+                                        if (objectFile) {
+                                            const objXml = (0, xmlUtils_js_1.parseXmlString)(objectFile.content.toString());
+                                            const mathNode = (0, xmlUtils_js_1.getFirstElementByTagName)(objXml, "math");
+                                            if (mathNode) {
+                                                isFormula = true;
+                                                formulaText = parseMathML(mathNode).trim();
+                                            }
+                                        }
                                     }
                                 }
-                                if (drawImages.length > 0) {
+                                if (isFormula) {
+                                    cellText += formulaText;
+                                    const textNode = {
+                                        type: 'text',
+                                        text: formulaText,
+                                        formatting: {}
+                                    };
+                                    if (config.includeRawContent) {
+                                        textNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frame, xmlString, config);
+                                    }
+                                    children.push(textNode);
+                                }
+                                else if (drawImages.length > 0) {
                                     // logic for image node
                                     const imageNode = {
                                         type: 'image',
                                         text: '', // Will be populated by assignAttachmentData
                                         children: [],
                                         metadata: {
-                                            attachmentName: imageHref, // Might be empty, will resolve in assignAttachmentData
+                                            attachmentName: imageHref || chartHref, // Might be empty, will resolve in assignAttachmentData
                                             ...(altText ? { altText } : {})
                                         }
                                     };
@@ -1195,6 +1379,9 @@ const parseOpenOffice = async (buffer, config) => {
                                             attachmentName: chartHref
                                         }
                                     };
+                                    if (config.includeRawContent) {
+                                        chartNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frame, xmlString, config);
+                                    }
                                     children.push(chartNode);
                                 }
                             }

package/dist/parsers/WordParser.js

@@ -473,7 +473,7 @@ const parseWord = async (buffer, config) => {
         const comments = [];
         // Traverse children of paragraph (runs, hyperlinks, etc.)
         const processChildNode = (node) => {
-            if ((0, xmlUtils_js_1.isElement)(node) && node.nodeName === 'w:r') {
+            if ((0, xmlUtils_js_1.isElement)(node) && (node.nodeName === 'w:r' || node.nodeName === 'm:r')) {
                 const runNode = node;
                 const rPr = (0, xmlUtils_js_1.getFirstElementByTagName)(runNode, "w:rPr");
                 // Formatting
@@ -512,7 +512,7 @@ const parseWord = async (buffer, config) => {
                         continue;
                     // also handle unprefixed version (mirroring the behaviour of getElementsByTagName)
                     // Text content
-                    if (child.tagName === "w:t" || child.tagName === "t") {
+                    if (child.tagName === "w:t" || child.tagName === "t" || child.tagName === "m:t") {
                         const tNode = child;
                         const tContent = tNode.textContent || '';
                         text += tContent;

package/dist/types.d.ts

@@ -317,7 +317,7 @@ export interface OfficeParserConfig {
      * The URL/path to the PDF.js worker script.
      *
      * **Mandatory** when using PDF parsing in browser environments to avoid worker configuration errors.
-     * If not provided, it defaults to `https://cdn.jsdelivr.net/npm/pdfjs-dist@5.6.205/build/pdf.worker.min.mjs`.
+     * If not provided, it defaults to `https://cdn.jsdelivr.net/npm/pdfjs-dist@6.1.200/build/pdf.worker.min.mjs`.
      * You can override this with your own local path or a different CDN link.
      */
     pdfWorkerSrc?: string;
@@ -1872,4 +1872,7 @@ export interface OfficeParserAST {
      */
     to<T extends this, D extends SupportedDestination<T['type']>>(this: T, destination: D, config?: GeneratorConfig<D>): Promise<ConversionResult<D>>;
 }
+declare global {
+    const __SLIM__: boolean | undefined;
+}
 export {};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "officeparser",
-  "version": "7.2.2",
+  "version": "7.2.3",
   "description": "A robust, strictly-typed Node.js and Browser library for parsing office files (.docx, .pptx, .xlsx, .odt, .odp, .ods, .pdf, .rtf, .csv, .md, .html) and generating high-fidelity outputs in Markdown, HTML, CSV, RTF, and RAG-focused chunks.",
   "funding": "https://github.com/sponsors/harshankur",
   "main": "dist/index.js",
@@ -13,6 +13,11 @@
       "browser": "./dist/officeparser.browser.mjs",
       "import": "./dist/index.mjs",
       "require": "./dist/index.js"
+    },
+    "./slim": {
+      "types": "./dist/officeparser.browser.slim.d.ts",
+      "browser": "./dist/officeparser.browser.slim.mjs",
+      "import": "./dist/officeparser.browser.slim.mjs"
     }
   },
   "sideEffects": false,
@@ -26,10 +31,10 @@
     "build": "npm run sync:versions && npm run build:node && npm run build:esm-wrapper && npm run build:browser:types && npm run build:browser",
     "build:node": "tsc",
     "build:esm-wrapper": "node scripts/generate-esm-wrapper.js",
-    "build:browser:types": "dts-bundle-generator --no-check -o dist/officeparser.browser.d.ts src/index.ts",
+    "build:browser:types": "dts-bundle-generator --no-check -o dist/officeparser.browser.d.ts src/index.ts && cp dist/officeparser.browser.d.ts dist/officeparser.browser.slim.d.ts",
     "build:browser": "node build_browser.js && npm run sync:docs",
     "sync:versions": "node scripts/sync-pdfjs-versions.js",
-    "sync:docs": "mkdir -p docs/dist && cp dist/officeparser.browser.iife.js docs/dist/ && cp dist/officeparser.browser.mjs docs/dist/ && mkdir -p docs/test/files && cp test/files/* docs/test/files/",
+    "sync:docs": "mkdir -p docs/dist && cp dist/officeparser.browser.iife.js docs/dist/ && cp dist/officeparser.browser.mjs docs/dist/ && cp dist/officeparser.browser.slim.iife.js docs/dist/ && cp dist/officeparser.browser.slim.mjs docs/dist/ && mkdir -p docs/test/files && cp test/files/* docs/test/files/",
     "lint": "eslint src",
     "test": "npm run lint && npm run test:clean && npm run build && npm run test:license && npm run test:artifacts && npm run test:parser && npm run test:generator && npm run test:cli",
     "test:baseline": "npm run test:parser:baseline && npm run test:generator:baseline",
@@ -105,9 +110,9 @@
   "homepage": "https://officeparser.harshankur.com",
   "dependencies": {
     "@xmldom/xmldom": "^0.9.10",
-    "fflate": "^0.8.2",
+    "fflate": "^0.8.3",
     "file-type": "^22.0.1",
-    "pdfjs-dist": "5.6.205",
+    "pdfjs-dist": "6.1.200",
     "tesseract.js": "^7.0.0"
   },
   "peerDependenciesMeta": {
@@ -121,8 +126,8 @@
     "@typescript-eslint/parser": "^8.59.3",
     "buffer": "^6.0.3",
     "dts-bundle-generator": "^9.5.1",
-    "esbuild": "^0.27.4",
-    "esbuild-plugins-node-modules-polyfill": "^1.8.1",
+    "esbuild": "^0.28.1",
+    "esbuild-plugins-node-modules-polyfill": "^1.8.2",
     "eslint": "^10.3.0",
     "husky": "^9.1.7",
     "postject": "^1.0.0-alpha.6",