officeparser 6.1.1 → 7.0.0

package/dist/parsers/OpenOfficeParser.js

@@ -23,6 +23,8 @@
  */
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parseOpenOffice = void 0;
+const types_js_1 = require("../types.js");
+const astUtils_js_1 = require("../utils/astUtils.js");
 const chartUtils_js_1 = require("../utils/chartUtils.js");
 const errorUtils_js_1 = require("../utils/errorUtils.js");
 const imageUtils_js_1 = require("../utils/imageUtils.js");
@@ -63,6 +65,7 @@ const parseOpenOffice = async (buffer, config) => {
     }
     const mainContentFile = files.find(f => f.path === 'content.xml') || files.find(f => f.path.match(contentFileRegex));
     const stylesFile = files.find(f => f.path === 'styles.xml');
+    const stylesDom = stylesFile ? (0, xmlUtils_js_1.parseXmlString)(stylesFile.content.toString()) : undefined;
     const content = [];
     const notes = [];
     // Style Map: styleName -> TextFormatting
@@ -76,8 +79,7 @@ const parseOpenOffice = async (buffer, config) => {
     let listIdCounter = 0;
     let lastWasList = false;
     // Helper to parse styles
-    const parseStyles = (xmlString) => {
-        const xml = (0, xmlUtils_js_1.parseXmlString)(xmlString);
+    const parseStyles = (xml) => {
         const styles = (0, xmlUtils_js_1.getElementsByTagName)(xml, "style:style");
         for (const style of styles) {
             const name = style.getAttribute("style:name");
@@ -156,8 +158,8 @@ const parseOpenOffice = async (buffer, config) => {
             }
         }
     };
-    if (stylesFile) {
-        parseStyles(stylesFile.content.toString());
+    if (stylesDom) {
+        parseStyles(stylesDom);
     }
     /**
      * Helper to parse a paragraph node (text:p or text:h) and extract its content.
@@ -177,9 +179,10 @@ const parseOpenOffice = async (buffer, config) => {
      */
     const parseInlineContent = (node, styleMap, config, notes, paragraphStyleMap, parentFormatting = {}, linkMetadata, sourceXml = '') => {
         const children = [];
+        const anchorIds = [];
         let fullText = '';
         if (!node.childNodes)
-            return { text: '', children: [] };
+            return { text: '', children: [], anchorIds: [] };
         for (let i = 0; i < node.childNodes.length; i++) {
             const child = node.childNodes[i];
             if (child.nodeType === 3) { // Text node
@@ -197,7 +200,12 @@ const parseOpenOffice = async (buffer, config) => {
             else if ((0, xmlUtils_js_1.isElement)(child)) {
                 const element = child;
                 const tagName = element.tagName;
-                if (tagName === 'text:s') {
+                if (tagName === 'text:bookmark' || tagName === 'text:bookmark-start') {
+                    const name = element.getAttribute('text:name');
+                    if (name)
+                        anchorIds.push(name);
+                }
+                else if (tagName === 'text:s') {
                     // Space
                     const count = parseInt(element.getAttribute('text:c') || '1');
                     const spaces = ' '.repeat(count);
@@ -236,15 +244,34 @@ const parseOpenOffice = async (buffer, config) => {
                     const spanContent = parseInlineContent(element, styleMap, config, notes, paragraphStyleMap, formatting, linkMetadata, sourceXml);
                     fullText += spanContent.text;
                     children.push(...spanContent.children);
+                    anchorIds.push(...spanContent.anchorIds);
                 }
                 else if (tagName === 'text:a') {
                     // Hyperlink
-                    const href = element.getAttribute('xlink:href') || '';
-                    const linkType = href.startsWith('#') ? 'internal' : 'external';
-                    const newLinkMetadata = { link: href, linkType: linkType };
+                    let href = element.getAttribute('xlink:href') || '';
+                    const isInternal = href.startsWith('#');
+                    const linkType = isInternal ? 'internal' : 'external';
+                    if (isInternal) {
+                        // ODT internal links can be encoded and might have suffixes like |outline
+                        try {
+                            href = decodeURIComponent(href).split('|')[0];
+                        }
+                        catch (e) {
+                            href = href.split('|')[0];
+                        }
+                        // Normalize internal link: if it contains #, keep only from # onwards
+                        if (href.includes('#')) {
+                            href = '#' + href.split('#').pop();
+                        }
+                    }
+                    let newLinkMetadata;
+                    if (!isInternal || !config.ignoreInternalLinks) {
+                        newLinkMetadata = { link: href, linkType: linkType };
+                    }
                     const linkContent = parseInlineContent(element, styleMap, config, notes, paragraphStyleMap, parentFormatting, newLinkMetadata, sourceXml);
                     fullText += linkContent.text;
                     children.push(...linkContent.children);
+                    anchorIds.push(...linkContent.anchorIds);
                 }
                 else if (tagName === 'text:note' && !config.ignoreNotes) {
                     // Footnote or endnote
@@ -263,7 +290,10 @@ const parseOpenOffice = async (buffer, config) => {
                                 type: 'paragraph',
                                 text: npContent.text,
                                 children: npContent.children,
-                                metadata: npContent.alignment ? { alignment: npContent.alignment } : undefined
+                                metadata: {
+                                    ...(npContent.alignment ? { alignment: npContent.alignment } : {}),
+                                    ...(npContent.anchorIds?.length ? { anchorIds: npContent.anchorIds } : {})
+                                }
                             };
                             noteChildren.push(npNode);
                         }
@@ -323,7 +353,7 @@ const parseOpenOffice = async (buffer, config) => {
                 }
             }
         }
-        return { text: fullText, children };
+        return { text: fullText, children, anchorIds };
     };
     /**
      * Helper to parse a paragraph node (text:p or text:h) and extract its content.
@@ -394,7 +424,7 @@ const parseOpenOffice = async (buffer, config) => {
                 }
             }
         }
-        return { text: content.text, children: content.children, alignment, style: paraStyle || undefined };
+        return { text: content.text, children: content.children, alignment, style: paraStyle || undefined, anchorIds: content.anchorIds };
     };
     /**
      * Splits paragraph content into multiple segments based on line breaks.
@@ -633,8 +663,9 @@ const parseOpenOffice = async (buffer, config) => {
                 (0, xmlUtils_js_1.getDirectChildren)(officeBody, "office:spreadsheet")[0];
             if (bodyContent) {
                 const bodyChildren = (0, xmlUtils_js_1.getDirectChildren)(bodyContent, "*");
+                const isSpreadsheet = bodyContent.tagName === "office:spreadsheet";
                 for (const child of bodyChildren) {
-                    traverse(child, content, false, xmlString);
+                    traverse(child, content, false, xmlString, isSpreadsheet);
                 }
             }
         }
@@ -646,19 +677,28 @@ const parseOpenOffice = async (buffer, config) => {
          * @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
+         * @param asSheet - If true, treats tables as sheets (for ODS)
          */
-        function traverse(node, targetArray, forceHeading = false, sourceXml) {
+        function 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';
+                const metadata = {
+                    ...(pContent.alignment ? { alignment: pContent.alignment } : {}),
+                    ...(pContent.style ? { style: pContent.style } : {}),
+                    ...(pContent.anchorIds?.length ? { anchorIds: pContent.anchorIds } : {})
+                };
+                const nodeId = node.getAttribute("xml:id") || node.getAttribute("text:id");
+                if (nodeId) {
+                    if (!metadata.anchorIds)
+                        metadata.anchorIds = [];
+                    metadata.anchorIds.push(nodeId);
+                }
                 const pNode = {
                     type,
                     text: pContent.text,
                     children: pContent.children,
-                    metadata: {
-                        ...(pContent.alignment ? { alignment: pContent.alignment } : {}),
-                        ...(pContent.style ? { style: pContent.style } : {})
-                    }
+                    metadata
                 };
                 if (type === 'heading' && pNode.metadata) {
                     pNode.metadata.level = pNode.metadata.level || 1;
@@ -675,15 +715,23 @@ const parseOpenOffice = async (buffer, config) => {
             else if (node.tagName === "text:h") {
                 const level = parseInt(node.getAttribute("text:outline-level") || "1");
                 const hContent = parseParagraphContent(node, paragraphStyleMap, styleMap, config, sourceXml);
+                const metadata = {
+                    level,
+                    ...(hContent.alignment ? { alignment: hContent.alignment } : {}),
+                    ...(hContent.style ? { style: hContent.style } : {}),
+                    ...(hContent.anchorIds?.length ? { anchorIds: hContent.anchorIds } : {})
+                };
+                const nodeId = node.getAttribute("xml:id") || node.getAttribute("text:id");
+                if (nodeId) {
+                    if (!metadata.anchorIds)
+                        metadata.anchorIds = [];
+                    metadata.anchorIds.push(nodeId);
+                }
                 const hNode = {
                     type: 'heading',
                     text: hContent.text,
                     children: hContent.children,
-                    metadata: {
-                        level,
-                        ...(hContent.alignment ? { alignment: hContent.alignment } : {}),
-                        ...(hContent.style ? { style: hContent.style } : {})
-                    }
+                    metadata
                 };
                 if (config.includeRawContent) {
                     hNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
@@ -694,6 +742,20 @@ const parseOpenOffice = async (buffer, config) => {
             else if (node.tagName === "table:table") {
                 // Parse table with proper structure
                 const tableNode = parseTable(node, paragraphStyleMap, styleMap, config, sourceXml);
+                if (asSheet) {
+                    tableNode.type = 'sheet';
+                    const sheetName = node.getAttribute("table:name");
+                    if (sheetName) {
+                        tableNode.metadata = { ...tableNode.metadata, sheetName };
+                    }
+                }
+                const tableId = node.getAttribute("xml:id") || node.getAttribute("table:name");
+                if (tableId) {
+                    if (!tableNode.metadata)
+                        tableNode.metadata = {};
+                    tableNode.metadata.anchorIds = tableNode.metadata.anchorIds || [];
+                    tableNode.metadata.anchorIds.push(tableId);
+                }
                 if (config.includeRawContent) {
                     tableNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
                 }
@@ -720,9 +782,8 @@ const parseOpenOffice = async (buffer, config) => {
                         parentNode = parentNode.parentNode;
                     }
                 }
-                // Try to find list style in automatic styles to determine type and visibility
+                // Try to find list style in automatic styles or styles.xml to determine type and visibility
                 if (styleNameToCheck) {
-                    const automaticStyles = (0, xmlUtils_js_1.getFirstElementByTagName)((0, xmlUtils_js_1.parseXmlString)(mainContentFile?.content.toString() || ''), "office:automatic-styles");
                     if (automaticStyles) {
                         const listStyles = (0, xmlUtils_js_1.getElementsByTagName)(automaticStyles, "text:list-style");
                         for (const listStyle of listStyles) {
@@ -739,32 +800,32 @@ const parseOpenOffice = async (buffer, config) => {
                                     listType = 'unordered';
                                     isVisible = bulletLevels.some(l => !!l.getAttribute("text:bullet-char"));
                                 }
-                                if (imageLevels.length > 0)
+                                else if (imageLevels.length > 0) {
+                                    listType = 'unordered';
                                     isVisible = true;
+                                }
                                 break;
                             }
                         }
                     }
-                    // Also check in styles.xml if still unordered and hidden
-                    if (stylesFile && !isVisible) {
-                        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_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"));
-                                }
-                                else if (bulletLevels.length > 0) {
-                                    listType = 'unordered';
-                                    isVisible = bulletLevels.some(l => !!l.getAttribute("text:bullet-char"));
+                    if (!isVisible && stylesDom) {
+                        const officeStyles = (0, xmlUtils_js_1.getFirstElementByTagName)(stylesDom, "office:styles");
+                        if (officeStyles) {
+                            const listStyles = (0, xmlUtils_js_1.getElementsByTagName)(officeStyles, "text:list-style");
+                            for (const listStyle of listStyles) {
+                                if (listStyle.getAttribute("style:name") === styleNameToCheck) {
+                                    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");
+                                    if (numberLevels.length > 0) {
+                                        listType = 'ordered';
+                                        isVisible = numberLevels.some(l => !!l.getAttribute("style:num-format"));
+                                    }
+                                    else if (bulletLevels.length > 0) {
+                                        listType = 'unordered';
+                                        isVisible = bulletLevels.some(l => !!l.getAttribute("text:bullet-char"));
+                                    }
+                                    break;
                                 }
-                                if (imageLevels.length > 0)
-                                    isVisible = true;
-                                break;
                             }
                         }
                     }
@@ -919,14 +980,19 @@ const parseOpenOffice = async (buffer, config) => {
                         const parts = imageHref.split('/');
                         imageHref = parts[parts.length - 1];
                     }
+                    const metadata = {
+                        attachmentName: imageHref,
+                        ...(altText ? { altText } : {})
+                    };
+                    const frameId = node.getAttribute("xml:id") || node.getAttribute("draw:name");
+                    if (frameId) {
+                        metadata.anchorIds = [frameId];
+                    }
                     const imageNode = {
                         type: 'image',
                         text: '',
                         children: [],
-                        metadata: {
-                            attachmentName: imageHref,
-                            ...(altText ? { altText } : {})
-                        }
+                        metadata
                     };
                     if (config.includeRawContent) {
                         imageNode.rawContent = (0, xmlUtils_js_1.getRawContent)(node, sourceXml, config);
@@ -1259,10 +1325,10 @@ const parseOpenOffice = async (buffer, config) => {
             if (config.ocr) {
                 if (attachment.mimeType.startsWith('image/')) {
                     try {
-                        attachment.ocrText = (await (0, ocrUtils_js_1.performOcr)(media.content, { language: config.ocrLanguage, ...config.ocrConfig })).trim();
+                        attachment.ocrText = (await (0, ocrUtils_js_1.performOcr)(media.content, { ...config.ocrConfig })).trim();
                     }
                     catch (e) {
-                        (0, errorUtils_js_1.logWarning)(`OCR failed for ${attachment.name}:`, config, e);
+                        (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.OCR_FAILED, config, attachment.name, e);
                     }
                 }
             }
@@ -1399,7 +1465,7 @@ const parseOpenOffice = async (buffer, config) => {
                             node.text = attachment.ocrText;
                         }
                         if (attachment.chartData && node.type === 'chart') {
-                            node.text = attachment.chartData.rawTexts.join(config.newlineDelimiter ?? '\n');
+                            node.text = attachment.chartData.rawTexts.join(config.newlineDelimiter);
                         }
                     }
                 }
@@ -1433,32 +1499,27 @@ const parseOpenOffice = async (buffer, config) => {
     if (config.putNotesAtLast && notes.length > 0) {
         content.push(...notes);
     }
-    return {
-        type: fileType,
-        metadata: {
-            ...metadata,
-            styleMap: combinedStyleMap
-        },
-        content: content,
-        attachments: attachments,
-        toText: () => content.map(c => {
-            const getText = (node) => {
-                let t = '';
-                if (node.children && node.children.length > 0) {
-                    // Check if children have their own children (container vs leaf)
-                    // If children are leaf nodes (text/image), join with empty string
-                    // If children are container nodes (paragraphs/rows), join with newline
-                    const hasGrandChildren = node.children.some(child => child.children && child.children.length > 0);
-                    const separator = hasGrandChildren ? (config.newlineDelimiter ?? '\n') : '';
-                    t += node.children.map(getText).filter(t => t != '').join(separator);
-                }
-                else {
-                    t += node.text || '';
-                }
-                return t;
-            };
-            return getText(c);
-        }).filter(t => t != '').join(config.newlineDelimiter ?? '\n')
-    };
+    const toTextSync = () => content.map(c => {
+        const getText = (node) => {
+            let t = '';
+            if (node.children && node.children.length > 0) {
+                // Check if children have their own children (container vs leaf)
+                // If children are leaf nodes (text/image), join with empty string
+                // If children are container nodes (paragraphs/rows), join with newline
+                const hasGrandChildren = node.children.some(child => child.children && child.children.length > 0);
+                const separator = hasGrandChildren ? config.newlineDelimiter : '';
+                t += node.children.map(getText).filter(t => t != '').join(separator);
+            }
+            else {
+                t += node.text || '';
+            }
+            return t;
+        };
+        return getText(c);
+    }).filter(t => t != '').join(config.newlineDelimiter);
+    return (0, astUtils_js_1.createAST)(fileType, {
+        ...metadata,
+        styleMap: combinedStyleMap
+    }, content, attachments, config, toTextSync);
 };
 exports.parseOpenOffice = parseOpenOffice;

package/dist/parsers/PdfParser.d.ts

@@ -56,7 +56,7 @@
  * @see https://mozilla.github.io/pdf.js/ PDF.js documentation
  * @see https://www.adobe.com/devnet/pdf/pdf_reference.html PDF Reference
  */
-import { OfficeParserAST, OfficeParserConfig } from '../types.js';
+import { FullOfficeParserConfig, OfficeParserAST } from '../types.js';
 /**
  * Parses a PDF file and extracts content.
  *
@@ -64,4 +64,4 @@ import { OfficeParserAST, OfficeParserConfig } from '../types.js';
  * @param config - Parser configuration
  * @returns Promise resolving to the parsed AST
  */
-export declare const parsePdf: (buffer: Buffer, config: OfficeParserConfig) => Promise<OfficeParserAST>;
+export declare const parsePdf: (buffer: Buffer, config: FullOfficeParserConfig) => Promise<OfficeParserAST>;

package/dist/parsers/PdfParser.js

@@ -59,12 +59,15 @@
  */
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parsePdf = void 0;
+const defaults_js_1 = require("../defaults.js");
+const types_js_1 = require("../types.js");
+const astUtils_js_1 = require("../utils/astUtils.js");
+const dateUtils_js_1 = require("../utils/dateUtils.js");
+const envUtils_js_1 = require("../utils/envUtils.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 moduleLoader_js_1 = require("../utils/moduleLoader.js");
-const dateUtils_js_1 = require("../utils/dateUtils.js");
-const envUtils_js_1 = require("../utils/envUtils.js");
+const ocrUtils_js_1 = require("../utils/ocrUtils.js");
 /** Type guard for TextItem in PDF.js 5.x */
 function isTextItem(item) {
     return item && typeof item.str === 'string' && Array.isArray(item.transform) && item.transform.length >= 6;
@@ -265,29 +268,38 @@ function convertToRgbaBuffer(data, width, height, kind) {
 const parsePdf = async (buffer, config) => {
     const pdfjs = await (0, moduleLoader_js_1.loadPdfJs)();
     // Configure worker
-    if (config.pdfWorkerSrc) {
-        pdfjs.GlobalWorkerOptions.workerSrc = config.pdfWorkerSrc;
+    const workerSrc = config.pdfWorkerSrc;
+    if (envUtils_js_1.isBrowser) {
+        pdfjs.GlobalWorkerOptions.workerSrc = workerSrc;
     }
     else {
-        // Fallbacks when no workerSrc is provided
-        if (envUtils_js_1.isBrowser) {
-            // Browser: Default to CDN
-            pdfjs.GlobalWorkerOptions.workerSrc = `https://unpkg.com/pdfjs-dist@${pdfjs.version}/build/pdf.worker.min.mjs`;
+        // Node.js: Try to auto-resolve local worker path to avoid remote fetch errors
+        (0, envUtils_js_1.assertNode)('pdf-worker-auto-resolution');
+        let resolved = false;
+        // If the user provided a custom path (not the default CDN one), use it.
+        // Otherwise, try to find it locally.
+        if (workerSrc !== defaults_js_1.DEFAULT_OFFICE_PARSER_CONFIG.pdfWorkerSrc && workerSrc !== '') {
+            pdfjs.GlobalWorkerOptions.workerSrc = workerSrc;
+            resolved = true;
         }
         else {
-            // Node.js: Try to auto-resolve local worker path to avoid remote fetch errors
-            (0, envUtils_js_1.assertNode)('pdf-worker-auto-resolution');
             try {
                 // We use require.resolve to find the exact path of the installed package.
                 // @ts-ignore - 'require' is available in Node.js/CommonJS environment
-                const workerPath = require.resolve('pdfjs-dist/legacy/build/pdf.worker.mjs');
-                pdfjs.GlobalWorkerOptions.workerSrc = workerPath;
+                const localWorkerPath = require.resolve('pdfjs-dist/legacy/build/pdf.worker.mjs');
+                // Use file:// URL for the worker source in Node.js to ensure compatibility with ESM-native PDF.js 5.x
+                // We use dynamic import for 'url' to avoid breaking browser bundles
+                const { pathToFileURL } = await import('url');
+                pdfjs.GlobalWorkerOptions.workerSrc = pathToFileURL(localWorkerPath).href;
+                resolved = true;
             }
             catch (e) {
-                if (config.outputErrorToConsole)
-                    console.warn("[PdfParser] Could not auto-resolve local worker path:", e);
+                (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.PDF_WORKER_FALLBACK, config, undefined, e);
             }
         }
+        if (!resolved) {
+            pdfjs.GlobalWorkerOptions.workerSrc = workerSrc;
+        }
     }
     const uint8Array = new Uint8Array(buffer);
     const loadingTask = pdfjs.getDocument({
@@ -302,7 +314,7 @@ const parsePdf = async (buffer, config) => {
     catch (e) {
         const message = e instanceof Error ? e.message : String(e);
         if (message.includes('workerSrc') || message.includes('No "GlobalWorkerOptions.workerSrc" specified')) {
-            throw (0, errorUtils_js_1.getOfficeError)(errorUtils_js_1.OfficeErrorType.PDF_WORKER_MISSING, config);
+            throw (0, errorUtils_js_1.getOfficeError)(types_js_1.OfficeErrorType.PDF_WORKER_MISSING, config);
         }
         throw e;
     }
@@ -382,8 +394,7 @@ const parsePdf = async (buffer, config) => {
         }
     }
     catch (e) {
-        if (config.outputErrorToConsole)
-            console.error("Error extracting embedded attachments:", e);
+        (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.ATTACHMENT_EXTRACTION_FAILED, config, undefined, e);
     }
     // --- First Pass: Collect all items for font statistics ---
     for (let i = 1; i <= numPages; i++) {
@@ -396,13 +407,13 @@ const parsePdf = async (buffer, config) => {
             textContent = await page.getTextContent();
         }
         catch (e) {
-            if (config.outputErrorToConsole)
-                console.warn(`[PdfParser] Error loading page ${i}:`, e);
+            (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.PAGE_LOAD_FAILED, config, i, e);
             // Push empty items to maintain index alignment for second pass
             allPageItems.push(pageItems);
             continue;
         }
         const commonObjs = page.commonObjs;
+        const fontCache = new Map();
         for (const item of textContent.items) {
             // PDF.js 5.x: textContent.items can contain TextMarkedContent which lack
             // 'str' and 'transform'. Skip these to avoid crashes and page skipping.
@@ -422,11 +433,15 @@ const parsePdf = async (buffer, config) => {
             if (textItem.fontName && commonObjs) {
                 try {
                     if (commonObjs.has(textItem.fontName)) {
-                        // Use callback-based get to ensure safe resolution
-                        const fontData = await new Promise((resolve) => {
-                            // @ts-ignore - commonObjs.get is callback-based in legacy builds
-                            commonObjs.get(textItem.fontName, (data) => resolve(data));
-                        });
+                        let fontData = fontCache.get(textItem.fontName);
+                        if (!fontData) {
+                            // Use callback-based get to ensure safe resolution
+                            fontData = await new Promise((resolve) => {
+                                // @ts-ignore - commonObjs.get is callback-based in legacy builds
+                                commonObjs.get(textItem.fontName, (data) => resolve(data));
+                            });
+                            fontCache.set(textItem.fontName, fontData);
+                        }
                         if (fontData?.name && typeof fontData.name === 'string') {
                             // Remove PDF subset prefix (6 uppercase letters + '+')
                             fontName = fontData.name.replace(/^[A-Z]{6}\+/, '');
@@ -485,9 +500,7 @@ const parsePdf = async (buffer, config) => {
                                 });
                             }
                             catch (e) {
-                                if (config.outputErrorToConsole) {
-                                    console.error(`[PdfParser] Failed to load dependency ${dep}:`, e);
-                                }
+                                (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.DEPENDENCY_LOAD_FAILED, config, dep, e);
                             }
                         }
                     }
@@ -507,7 +520,7 @@ const parsePdf = async (buffer, config) => {
                                     targetObjs.get(imgName, (data) => resolve(data));
                                 });
                                 // Browser-specific: Handle ImageBitmap if data is missing
-                                if (typeof window !== 'undefined' && !imgObj.data && imgObj.bitmap) {
+                                if (envUtils_js_1.isBrowser && !imgObj.data && imgObj.bitmap) {
                                     try {
                                         const canvas = document.createElement('canvas');
                                         canvas.width = imgObj.width;
@@ -520,8 +533,7 @@ const parsePdf = async (buffer, config) => {
                                         }
                                     }
                                     catch (e) {
-                                        if (config.outputErrorToConsole)
-                                            console.error(`[PdfParser] Failed to extract from ImageBitmap:`, e);
+                                        (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.IMAGE_PROCESSING_FAILED, config, undefined, e);
                                     }
                                 }
                                 if (imgObj?.data && imgObj.width > 0 && imgObj.height > 0) {
@@ -554,8 +566,7 @@ const parsePdf = async (buffer, config) => {
                 }
             }
             catch (e) {
-                if (config.outputErrorToConsole)
-                    console.error(`Error extracting images from page ${i}:`, e);
+                (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.IMAGE_EXTRACTION_FAILED, config, `from page ${i}`, e);
             }
         }
         allPageItems.push(pageItems);
@@ -570,8 +581,7 @@ const parsePdf = async (buffer, config) => {
             page = await pdfDocument.getPage(pageNum);
         }
         catch (e) {
-            if (config.outputErrorToConsole)
-                console.warn(`[PdfParser] Error loading page ${pageNum} in second pass:`, e);
+            (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.PAGE_LOAD_FAILED, config, pageNum, e);
             continue;
         }
         const pageItems = allPageItems[i];
@@ -594,8 +604,7 @@ const parsePdf = async (buffer, config) => {
             }
         }
         catch (e) {
-            if (config.outputErrorToConsole)
-                console.error(`Error extracting annotations from page ${pageNum}:`, e);
+            (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.ANNOTATION_EXTRACTION_FAILED, config, pageNum, e);
         }
         // Sort items: Y descending (top to bottom), then X ascending (left to right)
         pageItems.sort((a, b) => {
@@ -744,12 +753,11 @@ const parsePdf = async (buffer, config) => {
                             try {
                                 // Skip OCR for very small images/artifacts (e.g. < 10px) to avoid Tesseract warnings
                                 if (item.width >= 10 && item.height >= 10) {
-                                    attachment.ocrText = (await (0, ocrUtils_js_1.performOcr)(bmpBuffer, { language: config.ocrLanguage, ...config.ocrConfig })).trim();
+                                    attachment.ocrText = (await (0, ocrUtils_js_1.performOcr)(bmpBuffer, { ...config.ocrConfig })).trim();
                                 }
                             }
                             catch (e) {
-                                if (config.outputErrorToConsole)
-                                    console.error(`OCR failed for ${attachmentName}:`, e);
+                                (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.OCR_FAILED, config, attachmentName, e);
                             }
                         }
                         attachments.push(attachment);
@@ -764,7 +772,7 @@ const parsePdf = async (buffer, config) => {
                         });
                     }
                     catch (e) {
-                        (0, errorUtils_js_1.logWarning)(`Failed to process image ${attachmentName}:`, config, e);
+                        (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.IMAGE_EXTRACTION_FAILED, config, attachmentName, e);
                     }
                 }
             }
@@ -777,17 +785,12 @@ const parsePdf = async (buffer, config) => {
         content.push({
             type: 'page',
             children: pageContent,
-            text: pageContent.map(node => node.text).join(config.newlineDelimiter ?? '\n\n'),
+            text: pageContent.map(node => node.text).join(config.newlineDelimiter),
             metadata: { pageNumber: pageNum }
         });
     }
-    return {
-        type: 'pdf',
-        metadata: metadata,
-        content: content,
-        attachments: attachments,
-        toText: () => content.map(c => c.text).join(config.newlineDelimiter ?? '\n\n')
-    };
+    const toTextSync = () => content.map(c => c.text).join(config.newlineDelimiter);
+    return (0, astUtils_js_1.createAST)('pdf', metadata, content, attachments, config, toTextSync);
 };
 exports.parsePdf = parsePdf;
 /**

package/dist/parsers/PowerPointParser.d.ts

@@ -21,7 +21,7 @@
  * @module PowerPointParser
  * @see https://www.ecma-international.org/publications-and-standards/standards/ecma-376/
  */
-import { OfficeParserAST, OfficeParserConfig } from '../types.js';
+import { FullOfficeParserConfig, OfficeParserAST } from '../types.js';
 /**
  * Parses a PowerPoint presentation (.pptx) and extracts slides and notes.
  *
@@ -29,4 +29,4 @@ import { OfficeParserAST, OfficeParserConfig } from '../types.js';
  * @param config - Parser configuration
  * @returns A promise resolving to the parsed AST
  */
-export declare const parsePowerPoint: (buffer: Buffer, config: OfficeParserConfig) => Promise<OfficeParserAST>;
+export declare const parsePowerPoint: (buffer: Buffer, config: FullOfficeParserConfig) => Promise<OfficeParserAST>;