officeparser 6.1.0 → 7.0.0

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>;

package/dist/parsers/PowerPointParser.js

@@ -24,6 +24,8 @@
  */
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parsePowerPoint = 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");
@@ -392,117 +394,129 @@ const parsePowerPoint = async (buffer, config) => {
                 if (config.includeRawContent) {
                     pNode.rawContent = (0, xmlUtils_js_1.getRawContent)(p, xmlContentString, config);
                 }
-                const runs = (0, xmlUtils_js_1.getElementsByTagName)(p, "a:r");
-                for (let j = 0; j < runs.length; j++) {
-                    const r = runs[j];
-                    const t = (0, xmlUtils_js_1.getFirstElementByTagName)(r, "a:t");
-                    if (t && t.childNodes[0]) {
-                        const textContent = t.childNodes[0].nodeValue || '';
-                        pNode.text += textContent;
-                        const rPr = (0, xmlUtils_js_1.getFirstElementByTagName)(r, "a:rPr");
-                        const formatting = {};
-                        if (rPr) {
-                            if (rPr.getAttribute("b") === "1")
-                                formatting.bold = true;
-                            if (rPr.getAttribute("i") === "1")
-                                formatting.italic = true;
-                            if (rPr.getAttribute("u") === "sng")
-                                formatting.underline = true;
-                            if (rPr.getAttribute("strike") === "sngStrike")
-                                formatting.strikethrough = true;
-                            const sz = rPr.getAttribute("sz");
-                            if (sz)
-                                formatting.size = (parseInt(sz) / 100).toString() + 'pt';
-                            // Color extraction
-                            const solidFill = (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "a:solidFill");
-                            if (solidFill) {
-                                const srgbClr = (0, xmlUtils_js_1.getFirstElementByTagName)(solidFill, "a:srgbClr");
-                                if (srgbClr) {
-                                    const val = srgbClr.getAttribute("val");
-                                    if (val)
-                                        formatting.color = '#' + val;
+                // Process all children of <a:p> in order (runs, breaks, fields)
+                const children = Array.from(p.childNodes);
+                let activeNode = pNode;
+                nodes.push(activeNode);
+                for (const childNode of children) {
+                    if (!(0, xmlUtils_js_1.isElement)(childNode))
+                        continue;
+                    const element = childNode;
+                    const tag = element.tagName;
+                    if (tag === "a:r" || tag === "a:fld") {
+                        const t = (0, xmlUtils_js_1.getFirstElementByTagName)(element, "a:t");
+                        if (t && t.childNodes[0]) {
+                            const textContent = t.childNodes[0].nodeValue || "";
+                            activeNode.text += textContent;
+                            const rPr = (0, xmlUtils_js_1.getFirstElementByTagName)(element, "a:rPr");
+                            const formatting = {};
+                            if (rPr) {
+                                if (rPr.getAttribute("b") === "1")
+                                    formatting.bold = true;
+                                if (rPr.getAttribute("i") === "1")
+                                    formatting.italic = true;
+                                if (rPr.getAttribute("u") === "sng")
+                                    formatting.underline = true;
+                                if (rPr.getAttribute("strike") === "sngStrike")
+                                    formatting.strikethrough = true;
+                                const sz = rPr.getAttribute("sz");
+                                if (sz)
+                                    formatting.size = (parseInt(sz) / 100).toString() + "pt";
+                                // Color extraction
+                                const solidFill = (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "a:solidFill");
+                                if (solidFill) {
+                                    const srgbClr = (0, xmlUtils_js_1.getFirstElementByTagName)(solidFill, "a:srgbClr");
+                                    if (srgbClr) {
+                                        const val = srgbClr.getAttribute("val");
+                                        if (val)
+                                            formatting.color = "#" + val;
+                                    }
                                 }
-                            }
-                            // Highlight extraction
-                            const highlight = (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "a:highlight");
-                            if (highlight) {
-                                const srgbClr = (0, xmlUtils_js_1.getFirstElementByTagName)(highlight, "a:srgbClr");
-                                if (srgbClr) {
-                                    const val = srgbClr.getAttribute("val");
-                                    if (val)
-                                        formatting.backgroundColor = '#' + val;
+                                // Highlight extraction
+                                const highlight = (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "a:highlight");
+                                if (highlight) {
+                                    const srgbClr = (0, xmlUtils_js_1.getFirstElementByTagName)(highlight, "a:srgbClr");
+                                    if (srgbClr) {
+                                        const val = srgbClr.getAttribute("val");
+                                        if (val)
+                                            formatting.backgroundColor = "#" + val;
+                                    }
+                                }
+                                // Font family
+                                const latin = (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "a:latin");
+                                if (latin) {
+                                    const typeface = latin.getAttribute("typeface");
+                                    if (typeface)
+                                        formatting.font = typeface;
+                                }
+                                // Subscript/Superscript
+                                const baseline = rPr.getAttribute("baseline");
+                                if (baseline) {
+                                    const baselineVal = parseInt(baseline);
+                                    if (baselineVal < 0)
+                                        formatting.subscript = true;
+                                    if (baselineVal > 0)
+                                        formatting.superscript = true;
                                 }
                             }
-                            // Font family
-                            const latin = (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "a:latin");
-                            if (latin) {
-                                const typeface = latin.getAttribute("typeface");
-                                if (typeface)
-                                    formatting.font = typeface;
-                            }
-                            // Subscript/Superscript
-                            const baseline = rPr.getAttribute("baseline");
-                            if (baseline) {
-                                const baselineVal = parseInt(baseline);
-                                if (baselineVal < 0)
-                                    formatting.subscript = true;
-                                if (baselineVal > 0)
-                                    formatting.superscript = true;
+                            const textNode = {
+                                type: 'text',
+                                text: textContent,
+                                formatting: formatting
+                            };
+                            // Check for Hyperlinks
+                            const hlinkClick = (0, xmlUtils_js_1.getFirstElementByTagName)(element, "a:hlinkClick");
+                            if (hlinkClick) {
+                                const rId = hlinkClick.getAttribute("r:id");
+                                const action = hlinkClick.getAttribute("action");
+                                let link;
+                                let linkType;
+                                if (rId && slideRelsMap[slideNumber] && slideRelsMap[slideNumber][rId] && slideRelsMap[slideNumber][rId].type === "hyperlink") {
+                                    link = slideRelsMap[slideNumber][rId].target;
+                                    linkType = "external";
+                                }
+                                else if (rId && slideRelsMap[slideNumber] && slideRelsMap[slideNumber][rId] && slideRelsMap[slideNumber][rId].type === "slide") {
+                                    link = slideRelsMap[slideNumber][rId].target;
+                                    linkType = "internal";
+                                }
+                                else if (action) {
+                                    link = action;
+                                    linkType = "internal";
+                                }
+                                if (link) {
+                                    textNode.metadata = { link, linkType };
+                                }
                             }
+                            activeNode.children?.push(textNode);
                         }
-                        const textNode = {
-                            type: 'text',
-                            text: textContent,
-                            formatting: formatting
-                        };
-                        // Check for Hyperlinks
-                        const hlinkClick = (0, xmlUtils_js_1.getFirstElementByTagName)(r, "a:hlinkClick");
-                        // Check if this run has a hyperlink click action
-                        if (hlinkClick) {
-                            // Relationship ID for the link
-                            const rId = hlinkClick.getAttribute("r:id");
-                            // Optional action attribute, often for internal jumps
-                            const action = hlinkClick.getAttribute("action");
-                            // Result placeholders
-                            let link;
-                            let linkType;
-                            // Case 1: Relationship exists in slideRelsMap and is a real hyperlink (external URL)
-                            if (rId
-                                && slideRelsMap[slideNumber]
-                                && slideRelsMap[slideNumber][rId]
-                                && slideRelsMap[slideNumber][rId].type === "hyperlink") {
-                                // External URL
-                                link = slideRelsMap[slideNumber][rId].target;
-                                linkType = "external";
-                            }
-                            // Case 2: Relationship exists and is an internal slide reference
-                            else if (rId
-                                && slideRelsMap[slideNumber]
-                                && slideRelsMap[slideNumber][rId]
-                                && slideRelsMap[slideNumber][rId].type === "slide") {
-                                // Example target: ppt/slides/slide3.xml
-                                link = slideRelsMap[slideNumber][rId].target;
-                                linkType = "internal";
-                            }
-                            // Case 3: action attribute like ppaction://hlinksldjump
-                            else if (action) {
-                                link = action;
-                                linkType = "internal";
-                            }
-                            // Assign metadata only if a link was actually discovered
-                            if (link) {
-                                textNode.metadata = { link, linkType };
+                    }
+                    else if (tag === "a:br") {
+                        if (isList) {
+                            // Split the list item on soft break into a paragraph node
+                            activeNode = {
+                                type: 'paragraph',
+                                text: '',
+                                children: [],
+                                metadata: {
+                                    indentation: lvl,
+                                    alignment: pNode.metadata?.alignment || 'left'
+                                }
+                            };
+                            if (config.includeRawContent) {
+                                activeNode.rawContent = (0, xmlUtils_js_1.getRawContent)(p, xmlContentString, config);
                             }
+                            nodes.push(activeNode);
+                        }
+                        else {
+                            // In a normal paragraph, just add a newline
+                            activeNode.text += "\n";
+                            activeNode.children?.push({ type: 'text', text: "\n" });
                         }
-                        pNode.children?.push(textNode);
                     }
                 }
-                if (pNode.text) {
-                    nodes.push(pNode);
-                }
             }
         }
-        return nodes;
+        return nodes.filter(n => n.text?.trim() || (n.children && n.children.length > 0));
     };
     /**
      * Recursively traverses a PowerPoint shape tree (p:spTree),
@@ -692,10 +706,10 @@ const parsePowerPoint = 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);
                     }
                 }
             }
@@ -717,7 +731,7 @@ const parsePowerPoint = async (buffer, config) => {
                 attachment.chartData = chartData;
             }
             catch (e) {
-                (0, errorUtils_js_1.logWarning)(`Failed to extract text from chart ${chart.path}:`, config, e);
+                (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.CHART_DATA_EXTRACTION_FAILED, config, chart.path, e);
             }
         }
         // Loop through nodes to find images and charts and link their text and chartData
@@ -733,7 +747,7 @@ const parsePowerPoint = async (buffer, config) => {
                                 node.text = attachment.ocrText;
                         }
                         if (node.type === 'chart') {
-                            node.text = attachment.chartData?.rawTexts.join(config.newlineDelimiter || '\n');
+                            node.text = attachment.chartData?.rawTexts.join(config.newlineDelimiter);
                         }
                     }
                 }
@@ -752,24 +766,19 @@ const parsePowerPoint = async (buffer, config) => {
             return aIsNote - bIsNote;
         });
     }
-    return {
-        type: 'pptx',
-        metadata: metadata,
-        content: content,
-        attachments: attachments,
-        toText: () => content.map(c => {
-            // Recursive text extraction
-            const getText = (node) => {
-                let t = '';
-                if (node.children) {
-                    t += node.children.map(getText).filter(t => t != '').join(!node.children[0]?.children ? '' : config.newlineDelimiter ?? '\n');
-                }
-                else
-                    t += node.text || '';
-                return t;
-            };
-            return getText(c);
-        }).filter(t => t != '').join(config.newlineDelimiter ?? '\n')
-    };
+    const toTextSync = () => content.map(c => {
+        // Recursive text extraction
+        const getText = (node) => {
+            let t = '';
+            if (node.children) {
+                t += node.children.map(getText).filter(t => t != '').join(!node.children[0]?.children ? '' : config.newlineDelimiter);
+            }
+            else
+                t += node.text || '';
+            return t;
+        };
+        return getText(c);
+    }).filter(t => t != '').join(config.newlineDelimiter);
+    return (0, astUtils_js_1.createAST)('pptx', metadata, content, attachments, config, toTextSync);
 };
 exports.parsePowerPoint = parsePowerPoint;

package/dist/parsers/RtfParser.d.ts

@@ -39,7 +39,7 @@
  * @see https://www.biblioscape.com/rtf15_spec.htm RTF 1.5 Specification
  * @see https://latex2rtf.sourceforge.net/RTF-Spec-1.2.pdf RTF 1.2 Specification
  */
-import { OfficeParserAST, OfficeParserConfig } from '../types.js';
+import { FullOfficeParserConfig, OfficeParserAST } from '../types.js';
 /**
  * Represents an RTF group (content enclosed in braces).
  * Groups create formatting scopes and can contain other groups, control words, or text.
@@ -130,6 +130,12 @@ export declare class SimpleRtfParser {
     private index;
     /** The RTF content as a Buffer */
     private buffer;
+    /** Current code page for character decoding (default is Windows-1252) */
+    private codePage;
+    /** Cached TextDecoders for different code pages */
+    private decoders;
+    /** Buffer for consecutive text bytes to handle multi-byte encodings and UTF-8 detection */
+    private pendingBytes;
     /** Total length of the buffer */
     private length;
     /**
@@ -140,6 +146,20 @@ export declare class SimpleRtfParser {
     parse(): RtfGroup;
     private parseControl;
     private parseText;
+    /**
+     * Flushes the pending bytes buffer as a text node to the current group.
+     * @param group The group to append the text node to
+     */
+    private flushPendingText;
+    /**
+     * Decodes a byte array using a "UTF-8 first" strategy.
+     * If the bytes form valid UTF-8 and contain non-ASCII characters, UTF-8 is preferred.
+     * Otherwise, falls back to the specified code page.
+     * @param bytes The bytes to decode
+     * @param codePage The RTF code page ID
+     * @returns The decoded string
+     */
+    private decodeBytes;
 }
 /**
  * Parses an RTF file and returns the AST.
@@ -164,4 +184,4 @@ export declare class SimpleRtfParser {
  * @param config The parser configuration.
  * @returns The parsed AST.
  */
-export declare const parseRtf: (buffer: Buffer, config: OfficeParserConfig) => Promise<OfficeParserAST>;
+export declare const parseRtf: (buffer: Buffer, config: FullOfficeParserConfig) => Promise<OfficeParserAST>;