officeparser 6.0.7 → 6.1.1

package/dist/parsers/PowerPointParser.js

@@ -24,13 +24,12 @@
  */
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parsePowerPoint = void 0;
-const xmldom_1 = require("@xmldom/xmldom");
-const chartUtils_1 = require("../utils/chartUtils");
-const errorUtils_1 = require("../utils/errorUtils");
-const imageUtils_1 = require("../utils/imageUtils");
-const ocrUtils_1 = require("../utils/ocrUtils");
-const xmlUtils_1 = require("../utils/xmlUtils");
-const zipUtils_1 = require("../utils/zipUtils");
+const chartUtils_js_1 = require("../utils/chartUtils.js");
+const errorUtils_js_1 = require("../utils/errorUtils.js");
+const imageUtils_js_1 = require("../utils/imageUtils.js");
+const ocrUtils_js_1 = require("../utils/ocrUtils.js");
+const xmlUtils_js_1 = require("../utils/xmlUtils.js");
+const zipUtils_js_1 = require("../utils/zipUtils.js");
 /**
  * Parses a PowerPoint presentation (.pptx) and extracts slides and notes.
  *
@@ -46,14 +45,21 @@ const parsePowerPoint = async (buffer, config) => {
     const mediaFileRegex = /ppt\/media\/.*/;
     const chartFileRegex = /ppt\/charts\/chart\d+\.xml/;
     const corePropsFileRegex = /docProps\/core\.xml/;
-    const xmlSerializer = new xmldom_1.XMLSerializer();
-    const files = await (0, zipUtils_1.extractFiles)(buffer, x => !!x.match(config.ignoreNotes ? slidesRegex : allFilesRegex) ||
+    const customPropsFileRegex = /docProps\/custom\.xml/;
+    const files = await (0, zipUtils_js_1.extractFiles)(buffer, x => !!x.match(config.ignoreNotes ? slidesRegex : allFilesRegex) ||
         !!x.match(corePropsFileRegex) ||
+        !!x.match(customPropsFileRegex) ||
         !!x.match(slideRelsRegex) ||
         (!!config.extractAttachments && (!!x.match(mediaFileRegex) || !!x.match(chartFileRegex))));
     // Extract metadata
     const corePropsFile = files.find(f => f.path.match(corePropsFileRegex));
-    const metadata = corePropsFile ? (0, xmlUtils_1.parseOfficeMetadata)(corePropsFile.content.toString()) : {};
+    const metadata = corePropsFile ? (0, xmlUtils_js_1.parseOfficeMetadata)(corePropsFile.content.toString()) : {};
+    const customPropsFile = files.find(f => f.path.match(customPropsFileRegex));
+    if (customPropsFile) {
+        const customProperties = (0, xmlUtils_js_1.parseOOXMLCustomProperties)(customPropsFile.content.toString());
+        if (Object.keys(customProperties).length > 0)
+            metadata.customProperties = customProperties;
+    }
     // Sort files
     files.sort((a, b) => {
         const aMatch = a.path.match(slideNumberRegex);
@@ -73,21 +79,21 @@ const parsePowerPoint = async (buffer, config) => {
     // per indent counters (for nested lists)
     const levelCounters = {};
     // Helper to parse a table node
-    const parseTable = (tblNode) => {
+    const parseTable = (tblNode, xmlContentString) => {
         const rows = [];
-        const trNodes = (0, xmlUtils_1.getElementsByTagName)(tblNode, "a:tr");
+        const trNodes = (0, xmlUtils_js_1.getElementsByTagName)(tblNode, "a:tr");
         for (let rIndex = 0; rIndex < trNodes.length; rIndex++) {
             const trNode = trNodes[rIndex];
             const cells = [];
-            const tcNodes = (0, xmlUtils_1.getElementsByTagName)(trNode, "a:tc");
+            const tcNodes = (0, xmlUtils_js_1.getElementsByTagName)(trNode, "a:tc");
             for (let cIndex = 0; cIndex < tcNodes.length; cIndex++) {
                 const tcNode = tcNodes[cIndex];
                 const cellChildren = [];
                 let cellText = '';
                 // Cells contain text bodies (txBody) which contain paragraphs
-                const txBody = (0, xmlUtils_1.getElementsByTagName)(tcNode, "a:txBody")[0];
+                const txBody = (0, xmlUtils_js_1.getFirstElementByTagName)(tcNode, "a:txBody");
                 if (txBody) {
-                    const paragraphs = (0, xmlUtils_1.getElementsByTagName)(txBody, "a:p");
+                    const paragraphs = (0, xmlUtils_js_1.getElementsByTagName)(txBody, "a:p");
                     for (const p of paragraphs) {
                         // Reuse paragraph parsing logic if possible, or duplicate for now
                         // For simplicity, duplicating basic logic here as the main loop one is tied to shapes
@@ -98,15 +104,15 @@ const parsePowerPoint = async (buffer, config) => {
                             metadata: {}
                         };
                         if (config.includeRawContent) {
-                            pNode.rawContent = p.toString();
+                            pNode.rawContent = (0, xmlUtils_js_1.getRawContent)(p, xmlContentString, config);
                         }
-                        const runs = (0, xmlUtils_1.getElementsByTagName)(p, "a:r");
+                        const runs = (0, xmlUtils_js_1.getElementsByTagName)(p, "a:r");
                         for (const r of runs) {
-                            const t = (0, xmlUtils_1.getElementsByTagName)(r, "a:t")[0];
+                            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_1.getElementsByTagName)(r, "a:rPr")[0];
+                                const rPr = (0, xmlUtils_js_1.getFirstElementByTagName)(r, "a:rPr");
                                 const formatting = {};
                                 if (rPr) {
                                     if (rPr.getAttribute("b") === "1")
@@ -120,16 +126,16 @@ const parsePowerPoint = async (buffer, config) => {
                                     const sz = rPr.getAttribute("sz");
                                     if (sz)
                                         formatting.size = (parseInt(sz) / 100).toString() + 'pt';
-                                    const solidFill = (0, xmlUtils_1.getElementsByTagName)(rPr, "a:solidFill")[0];
+                                    const solidFill = (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "a:solidFill");
                                     if (solidFill) {
-                                        const srgbClr = (0, xmlUtils_1.getElementsByTagName)(solidFill, "a:srgbClr")[0];
+                                        const srgbClr = (0, xmlUtils_js_1.getFirstElementByTagName)(solidFill, "a:srgbClr");
                                         if (srgbClr) {
                                             const val = srgbClr.getAttribute("val");
                                             if (val)
                                                 formatting.color = '#' + val;
                                         }
                                     }
-                                    const latin = (0, xmlUtils_1.getElementsByTagName)(rPr, "a:latin")[0];
+                                    const latin = (0, xmlUtils_js_1.getFirstElementByTagName)(rPr, "a:latin");
                                     if (latin) {
                                         const typeface = latin.getAttribute("typeface");
                                         if (typeface)
@@ -167,8 +173,8 @@ const parsePowerPoint = async (buffer, config) => {
         };
     };
     /** Extract an AST node for p:pic */
-    const extractImageNode = (imageNode, slideNumber) => {
-        const blip = (0, xmlUtils_1.getElementsByTagName)(imageNode, "a:blip")[0];
+    const extractImageNode = (imageNode, slideNumber, xmlContentString) => {
+        const blip = (0, xmlUtils_js_1.getFirstElementByTagName)(imageNode, "a:blip");
         if (!blip)
             return null;
         const rId = blip.getAttribute("r:embed");
@@ -178,8 +184,8 @@ const parsePowerPoint = async (buffer, config) => {
         if (!rel || rel.type !== "image")
             return null;
         const attachmentName = rel.target;
-        const nvPicPr = (0, xmlUtils_1.getElementsByTagName)(imageNode, "p:nvPicPr")[0];
-        const cNvPr = nvPicPr ? (0, xmlUtils_1.getElementsByTagName)(nvPicPr, "p:cNvPr")[0] : null;
+        const nvPicPr = (0, xmlUtils_js_1.getFirstElementByTagName)(imageNode, "p:nvPicPr");
+        const cNvPr = nvPicPr ? (0, xmlUtils_js_1.getFirstElementByTagName)(nvPicPr, "p:cNvPr") : null;
         const altText = cNvPr?.getAttribute("descr") || undefined;
         return {
             type: "image",
@@ -192,23 +198,11 @@ const parsePowerPoint = async (buffer, config) => {
     };
     /**
      * Extract an AST node for p:graphicFrame that contains a chart.
-     * This mirrors extractImageNode, but for charts instead of images.
-     *
-     * Steps:
-     * 1. Find <a:graphicData> inside the graphicFrame.
-     * 2. Check if uri is the chart namespace.
-     * 3. Extract <c:chart> and read r:id.
-     * 4. Resolve the relationship using slideRelsMap.
-     * 5. Produce a chart node with attachmentName (like images).
-     * 6. Chart text/data will be injected later in the pipeline.
-     *
-     * @param frameNode p:graphicFrame element
-     * @param slideNumber Current slide number for relationship resolution
-     * @returns A chart OfficeContentNode or null if not a chart.
+     * ... (comments omitted for brevity) ...
      */
-    const extractChartNode = (frameNode, slideNumber) => {
+    const extractChartNode = (frameNode, slideNumber, xmlContentString) => {
         // Step 1: Find <a:graphicData>
-        const graphicData = (0, xmlUtils_1.getElementsByTagName)(frameNode, "a:graphicData")[0];
+        const graphicData = (0, xmlUtils_js_1.getFirstElementByTagName)(frameNode, "a:graphicData");
         if (!graphicData) {
             return null;
         }
@@ -220,7 +214,7 @@ const parsePowerPoint = async (buffer, config) => {
             return null;
         }
         // Step 3: Find <c:chart>
-        const cChart = (0, xmlUtils_1.getElementsByTagName)(graphicData, "c:chart")[0];
+        const cChart = (0, xmlUtils_js_1.getFirstElementByTagName)(graphicData, "c:chart");
         if (!cChart) {
             return null;
         }
@@ -246,24 +240,24 @@ const parsePowerPoint = async (buffer, config) => {
         };
         // Optional: include raw XML of the whole frame
         if (config.includeRawContent) {
-            chartNode.rawContent = xmlSerializer.serializeToString(frameNode);
+            chartNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frameNode, xmlContentString, config);
         }
         return chartNode;
     };
     /** Extract an AST node for p:graphicFrame */
-    const extractGraphicFrameNode = (frameNode, slideNumber) => {
-        const tbl = (0, xmlUtils_1.getElementsByTagName)(frameNode, "a:tbl")[0];
+    const extractGraphicFrameNode = (frameNode, slideNumber, xmlContentString) => {
+        const tbl = (0, xmlUtils_js_1.getFirstElementByTagName)(frameNode, "a:tbl");
         if (tbl) {
-            const tableNode = parseTable(tbl);
+            const tableNode = parseTable(tbl, xmlContentString);
             if (config.includeRawContent) {
-                tableNode.rawContent = xmlSerializer.serializeToString(frameNode);
+                tableNode.rawContent = (0, xmlUtils_js_1.getRawContent)(frameNode, xmlContentString, config);
             }
             if (tableNode.children && tableNode.children.length > 0) {
                 return tableNode;
             }
         }
         if (frameNode.getElementsByTagName("c:chart").length > 0) {
-            const chartNode = extractChartNode(frameNode, slideNumber);
+            const chartNode = extractChartNode(frameNode, slideNumber, xmlContentString);
             if (chartNode) {
                 return chartNode;
             }
@@ -271,17 +265,17 @@ const parsePowerPoint = async (buffer, config) => {
         return null;
     };
     /** Extract text and hyperlinks from a p:sp shape. */
-    const extractShapeNodes = (spNode, slideNumber) => {
+    const extractShapeNodes = (spNode, slideNumber, xmlContentString) => {
         const nodes = [];
         // Check for placeholder type (title, body, etc.)
-        const nvSpPr = (0, xmlUtils_1.getElementsByTagName)(spNode, "p:nvSpPr")[0];
-        const nvPr = nvSpPr ? (0, xmlUtils_1.getElementsByTagName)(nvSpPr, "p:nvPr")[0] : null;
-        const ph = nvPr ? (0, xmlUtils_1.getElementsByTagName)(nvPr, "p:ph")[0] : null;
+        const nvSpPr = (0, xmlUtils_js_1.getFirstElementByTagName)(spNode, "p:nvSpPr");
+        const nvPr = nvSpPr ? (0, xmlUtils_js_1.getFirstElementByTagName)(nvSpPr, "p:nvPr") : null;
+        const ph = nvPr ? (0, xmlUtils_js_1.getFirstElementByTagName)(nvPr, "p:ph") : null;
         const type = ph ? ph.getAttribute("type") : "body";
         const isTitle = type === "title" || type === "ctrTitle";
-        const txBody = (0, xmlUtils_1.getElementsByTagName)(spNode, "p:txBody")[0];
+        const txBody = (0, xmlUtils_js_1.getFirstElementByTagName)(spNode, "p:txBody");
         if (txBody) {
-            const paragraphs = (0, xmlUtils_1.getElementsByTagName)(txBody, "a:p");
+            const paragraphs = (0, xmlUtils_js_1.getElementsByTagName)(txBody, "a:p");
             for (let i = 0; i < paragraphs.length; i++) {
                 const p = paragraphs[i];
                 const pNode = {
@@ -291,7 +285,7 @@ const parsePowerPoint = async (buffer, config) => {
                     metadata: isTitle ? { level: 1 } : {}
                 };
                 // Paragraph Alignment and List Detection
-                const pPr = (0, xmlUtils_1.getElementsByTagName)(p, "a:pPr")[0];
+                const pPr = (0, xmlUtils_js_1.getFirstElementByTagName)(p, "a:pPr");
                 let isList = false;
                 let listType = 'unordered';
                 let lvl = 0;
@@ -299,10 +293,10 @@ const parsePowerPoint = async (buffer, config) => {
                     const lvlAttr = pPr.getAttribute("lvl");
                     if (lvlAttr)
                         lvl = parseInt(lvlAttr);
-                    const buAutoNum = (0, xmlUtils_1.getElementsByTagName)(pPr, "a:buAutoNum")[0];
-                    const buChar = (0, xmlUtils_1.getElementsByTagName)(pPr, "a:buChar")[0];
-                    const buBlip = (0, xmlUtils_1.getElementsByTagName)(pPr, "a:buBlip")[0];
-                    const buNode = (0, xmlUtils_1.getElementsByTagName)(pPr, "a:bu")[0];
+                    const buAutoNum = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "a:buAutoNum");
+                    const buChar = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "a:buChar");
+                    const buBlip = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "a:buBlip");
+                    const buNode = (0, xmlUtils_js_1.getFirstElementByTagName)(pPr, "a:bu");
                     if (buAutoNum) {
                         isList = true;
                         listType = 'ordered';
@@ -396,119 +390,131 @@ const parsePowerPoint = async (buffer, config) => {
                     pNode.metadata = { ...pNode.metadata, level: 1 };
                 }
                 if (config.includeRawContent) {
-                    pNode.rawContent = p.toString();
+                    pNode.rawContent = (0, xmlUtils_js_1.getRawContent)(p, xmlContentString, config);
                 }
-                const runs = (0, xmlUtils_1.getElementsByTagName)(p, "a:r");
-                for (let j = 0; j < runs.length; j++) {
-                    const r = runs[j];
-                    const t = (0, xmlUtils_1.getElementsByTagName)(r, "a:t")[0];
-                    if (t && t.childNodes[0]) {
-                        const textContent = t.childNodes[0].nodeValue || '';
-                        pNode.text += textContent;
-                        const rPr = (0, xmlUtils_1.getElementsByTagName)(r, "a:rPr")[0];
-                        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_1.getElementsByTagName)(rPr, "a:solidFill")[0];
-                            if (solidFill) {
-                                const srgbClr = (0, xmlUtils_1.getElementsByTagName)(solidFill, "a:srgbClr")[0];
-                                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_1.getElementsByTagName)(rPr, "a:highlight")[0];
-                            if (highlight) {
-                                const srgbClr = (0, xmlUtils_1.getElementsByTagName)(highlight, "a:srgbClr")[0];
-                                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_1.getElementsByTagName)(rPr, "a:latin")[0];
-                            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_1.getElementsByTagName)(r, "a:hlinkClick")[0];
-                        // 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),
@@ -520,31 +526,32 @@ const parsePowerPoint = async (buffer, config) => {
      * transforms so nested groups inherit positional transforms.
      *
      * @param treeNode The XML node representing <p:spTree>
-     * @param parentTransform Transform inherited from parent groups (defaults to identity)
+     * @param slideNumber Current slide number for relationship resolution
+     * @param xmlContentString The source XML string for raw content extraction
      */
-    function traverseSpTree(treeNode, slideNumber) {
+    function traverseSpTree(treeNode, slideNumber, xmlContentString) {
         const nodes = [];
         // Process children in XML order (this preserves Z-order)
         for (const child of Array.from(treeNode?.childNodes || [])) {
-            if (child.nodeType !== 1) {
+            if (!(0, xmlUtils_js_1.isElement)(child)) {
                 continue;
             }
             const element = child;
             const tag = element.tagName;
             // Case 1: Normal shape
             if (tag === "p:sp") {
-                nodes.push(...extractShapeNodes(element, slideNumber));
+                nodes.push(...extractShapeNodes(element, slideNumber, xmlContentString));
             }
             // Case 2: Inline picture
             else if (tag === "p:pic") {
-                const imageNode = extractImageNode(element, slideNumber);
+                const imageNode = extractImageNode(element, slideNumber, xmlContentString);
                 if (imageNode) {
                     nodes.push(imageNode);
                 }
             }
             // Case 3: Chart or other graphic frame
             else if (tag === "p:graphicFrame") {
-                const tableNode = extractGraphicFrameNode(element, slideNumber);
+                const tableNode = extractGraphicFrameNode(element, slideNumber, xmlContentString);
                 if (tableNode) {
                     nodes.push(tableNode);
                 }
@@ -552,9 +559,11 @@ const parsePowerPoint = async (buffer, config) => {
             // Case 4: Grouped shape (recursive!)
             else if (tag === "p:grpSp") {
                 // Extract the nested <p:spTree> inside the group
-                const nestedTree = element.getElementsByTagName("p:spTree")[0];
+                const nestedTree = (0, xmlUtils_js_1.getFirstElementByTagName)(element, "p:spTree");
                 // Recurse into the nested tree
-                nodes.push(...traverseSpTree(nestedTree, slideNumber));
+                if (nestedTree) {
+                    nodes.push(...traverseSpTree(nestedTree, slideNumber, xmlContentString));
+                }
             }
         }
         return nodes;
@@ -587,9 +596,9 @@ const parsePowerPoint = async (buffer, config) => {
                 // Prepare map for this slide
                 slideRelsMap[slideNum] = {};
                 // Parse the rels XML
-                const relsXml = (0, xmlUtils_1.parseXmlString)(file.content.toString());
+                const relsXml = (0, xmlUtils_js_1.parseXmlString)(file.content.toString());
                 // Get all Relationship nodes
-                const relationships = (0, xmlUtils_1.getElementsByTagName)(relsXml, "Relationship");
+                const relationships = (0, xmlUtils_js_1.getElementsByTagName)(relsXml, "Relationship");
                 // Loop through each relationship node
                 for (let i = 0; i < relationships.length; i++) {
                     // Relationship ID, Example: "rId2"
@@ -653,7 +662,7 @@ const parsePowerPoint = async (buffer, config) => {
         if (file.path.match(corePropsFileRegex))
             continue;
         const xmlContentString = file.content.toString();
-        const xml = (0, xmlUtils_1.parseXmlString)(xmlContentString);
+        const xml = (0, xmlUtils_js_1.parseXmlString)(xmlContentString, { locator: config.includeRawContent });
         if (config.includeRawContent) {
             rawContents.push(xmlContentString);
         }
@@ -669,15 +678,15 @@ const parsePowerPoint = async (buffer, config) => {
             }
         };
         if (config.includeRawContent) {
-            slideNode.rawContent = file.content.toString();
+            slideNode.rawContent = (0, xmlUtils_js_1.getRawContent)(xml, xmlContentString, config);
         }
         /**
          * Extract slide contents in correct document order by scanning p:spTree children.
          * This ensures p:pic, p:sp, p:graphicFrame appear in AST in the exact sequence.
          */
-        const spTree = (0, xmlUtils_1.getElementsByTagName)(xml, "p:spTree")[0];
+        const spTree = (0, xmlUtils_js_1.getFirstElementByTagName)(xml, "p:spTree");
         if (spTree) {
-            slideNode.children?.push(...traverseSpTree(spTree, slideNumber));
+            slideNode.children?.push(...traverseSpTree(spTree, slideNumber, xmlContentString));
         }
         if (slideNode.children && slideNode.children.length > 0) {
             content.push(slideNode);
@@ -690,15 +699,15 @@ const parsePowerPoint = async (buffer, config) => {
     if (config.extractAttachments) {
         // Extract media files as attachments
         for (const media of mediaFiles) {
-            const attachment = (0, imageUtils_1.createAttachment)(media.path.split('/').pop() || 'image', media.content);
+            const attachment = (0, imageUtils_js_1.createAttachment)(media.path.split('/').pop() || 'image', media.content);
             attachments.push(attachment);
             if (config.ocr) {
                 if (attachment.mimeType.startsWith('image/')) {
                     try {
-                        attachment.ocrText = (await (0, ocrUtils_1.performOcr)(media.content, config.ocrLanguage)).trim();
+                        attachment.ocrText = (await (0, ocrUtils_js_1.performOcr)(media.content, { language: config.ocrLanguage, ...config.ocrConfig })).trim();
                     }
                     catch (e) {
-                        (0, errorUtils_1.logWarning)(`OCR failed for ${attachment.name}:`, config, e);
+                        (0, errorUtils_js_1.logWarning)(`OCR failed for ${attachment.name}:`, config, e);
                     }
                 }
             }
@@ -715,12 +724,12 @@ const parsePowerPoint = async (buffer, config) => {
             attachments.push(attachment);
             // Extract text from chart XML
             try {
-                const chartData = await (0, chartUtils_1.extractChartData)(chart.content);
+                const chartData = await (0, chartUtils_js_1.extractChartData)(chart.content);
                 // Assign chartData to attachment
                 attachment.chartData = chartData;
             }
             catch (e) {
-                (0, errorUtils_1.logWarning)(`Failed to extract text from chart ${chart.path}:`, config, e);
+                (0, errorUtils_js_1.logWarning)(`Failed to extract text from chart ${chart.path}:`, config, e);
             }
         }
         // Loop through nodes to find images and charts and link their text and chartData