officeparser 7.5.0 → 7.5.1

package/dist/parsers/EpubParser.js

@@ -37,7 +37,7 @@ const parseEpub = async (buffer, config) => {
     const files = await (0, zipUtils_js_1.extractFiles)(buffer, (path) => /META-INF\/container\.xml$/i.test(path)
         || /\.opf$/i.test(path)
         || /\.(xhtml|html|htm)$/i.test(path)
-        || (!!config.extractAttachments && /\.(png|jpe?g|gif|svg|webp)$/i.test(path)), config.decompressionLimits);
+        || (!!config.extractAttachments && /\.(png|jpe?g|gif|svg|webp)$/i.test(path)), config.decompressionLimits, config);
     // The OPF path is authoritative via META-INF/container.xml; fall back to scanning
     // for any .opf file for malformed archives that skip the container manifest.
     let opfPath;
@@ -49,7 +49,7 @@ const parseEpub = async (buffer, config) => {
     }
     const opfFile = (opfPath && files.find(f => f.path === opfPath)) || files.find(f => /\.opf$/i.test(f.path));
     if (!opfFile) {
-        throw (0, errorUtils_js_1.getOfficeError)(types_js_1.OfficeErrorType.FILE_CORRUPTED, config, 'epub (no OPF manifest found)');
+        throw (0, errorUtils_js_1.getOfficeError)(types_js_1.OfficeErrorType.REQUIRED_PART_MISSING, config, { fileType: 'epub', part: 'OPF package document (.opf)' });
     }
     const opfDir = opfFile.path.includes('/') ? opfFile.path.substring(0, opfFile.path.lastIndexOf('/') + 1) : '';
     const opfXml = (0, xmlUtils_js_1.parseXmlString)(opfFile.content.toString('utf-8'));

package/dist/parsers/ExcelParser.js

@@ -67,7 +67,16 @@ const parseExcel = async (buffer, config) => {
         !!x.match(customPropsFileRegex) ||
         !!x.match(appPropsFileRegex) ||
         (!!config.extractAttachments && (!!x.match(mediaFileRegex) || !!x.match(drawingRelsRegex))) ||
-        ((!!config.extractAttachments || !config.ignoreComments) && !!x.match(relsRegex)), config.decompressionLimits);
+        ((!!config.extractAttachments || !config.ignoreComments) && !!x.match(relsRegex)), config.decompressionLimits, config);
+    // Every workbook has xl/workbook.xml; without it the archive is not a spreadsheet.
+    // Resolved up front so a file that cannot be a workbook fails before any of the parsing
+    // work below, and read again further down for the sheet-name map.
+    const workbookFile = (0, zipUtils_js_1.findRequiredPart)(files, path => path === 'xl/workbook.xml', config, { fileType: 'xlsx', part: 'xl/workbook.xml' });
+    // Worksheets, by contrast, are not guaranteed: a workbook holding only chartsheets is
+    // valid and simply has no cell text to extract. Warn rather than fail, so the caller can
+    // tell "nothing to read here" from "we read nothing".
+    if (!files.some(file => !!file.path.match(sheetsRegex)))
+        (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.NO_WORKSHEETS_FOUND, config);
     const sharedStringsFile = files.find(f => f.path === stringsFilePath);
     // Updated to store structured content (rich text runs) or simple string
     const sharedStrings = [];
@@ -372,9 +381,8 @@ const parseExcel = async (buffer, config) => {
     }
     // Parse workbook.xml to get sheet names and map them to sheet files
     const sheetNameMap = {};
-    const workbookFile = files.find(f => f.path === 'xl/workbook.xml');
     const workbookRelsFile = files.find(f => f.path === 'xl/_rels/workbook.xml.rels');
-    if (workbookFile && workbookRelsFile) {
+    if (workbookRelsFile) {
         // Parse rels to get rId -> file mapping
         const relsXml = (0, xmlUtils_js_1.parseXmlString)(workbookRelsFile.content.toString());
         const relationships = (0, xmlUtils_js_1.getElementsByTagName)(relsXml, "Relationship");
@@ -401,7 +409,6 @@ const parseExcel = async (buffer, config) => {
         }
     }
     const content = [];
-    const rawContents = [];
     for (const file of files) {
         if (file.path.match(mediaFileRegex))
             continue;
@@ -418,9 +425,6 @@ const parseExcel = async (buffer, config) => {
         if (file.path.match(drawingRelsRegex))
             continue;
         if (file.path.match(sheetsRegex)) {
-            if (config.includeRawContent) {
-                rawContents.push(file.content.toString());
-            }
             const sheetFilename = file.path.split('/').pop() || '';
             const relsFilename = `xl/worksheets/_rels/${sheetFilename}.rels`;
             const relsFile = files.find(f => f.path === relsFilename);

package/dist/parsers/OpenOfficeParser.js

@@ -125,6 +125,8 @@ const cleanAttachmentName = (href) => {
     const cleaned = href.replace(/^\.\//, '').replace(/\/$/, '');
     return cleaned.split('/').pop() || '';
 };
+/** The ODF document types this parser handles, used to validate a caller-supplied file type. */
+const ODF_FILE_TYPES = ['odt', 'odp', 'ods'];
 /**
  * Parses an OpenOffice document (.odt, .odp, .ods) and extracts content.
  *
@@ -148,10 +150,17 @@ const parseOpenOffice = async (buffer, config) => {
         !!x.match(metaFileRegex) ||
         !!x.match(stylesFileRegex) ||
         !!x.match(mimetypeFileRegex) ||
-        (!!config.extractAttachments && !!x.match(mediaFileRegex)), config.decompressionLimits);
+        (!!config.extractAttachments && !!x.match(mediaFileRegex)), config.decompressionLimits, config);
     // 1. Determine File Type
     const mimetypeFile = files.find(f => f.path === 'mimetype');
-    let fileType = 'odt'; // Default
+    // The archive's own mimetype entry is authoritative when present. When it is missing,
+    // fall back to the type the caller asked for (or that was derived from the extension)
+    // rather than assuming text: guessing 'odt' for a spreadsheet sends the parser down the
+    // office:text branch, which finds nothing in an office:spreadsheet body and yields an
+    // empty document for a perfectly valid file.
+    let fileType = ODF_FILE_TYPES.includes(config.fileType)
+        ? config.fileType
+        : 'odt';
     if (mimetypeFile) {
         const mime = mimetypeFile.content.toString().trim();
         if (mime.includes('spreadsheet'))
@@ -161,7 +170,12 @@ const parseOpenOffice = async (buffer, config) => {
         else if (mime.includes('text'))
             fileType = 'odt';
     }
-    const mainContentFile = files.find(f => f.path === 'content.xml') || files.find(f => f.path.match(contentFileRegex));
+    // The document body is the content.xml at the archive root. The fallback stays anchored
+    // and excludes embedded objects: an ODF file can carry Object N/content.xml for a chart
+    // or formula, and an unanchored match would promote one of those to the document body
+    // when the real one is missing, silently parsing a chart as if it were the whole file.
+    const mainContentFile = files.find(f => f.path === 'content.xml')
+        || (0, zipUtils_js_1.findRequiredPart)(files, path => /(^|\/)content\.xml$/.test(path) && !objectContentFileRegex.test(path), config, { fileType, part: 'content.xml' });
     const stylesFile = files.find(f => f.path === 'styles.xml');
     const stylesDom = stylesFile ? (0, xmlUtils_js_1.parseXmlString)(stylesFile.content.toString()) : undefined;
     const content = [];

package/dist/parsers/PowerPointParser.js

@@ -57,6 +57,7 @@ const parsePowerPoint = async (buffer, config) => {
     const commentsFileRegex = /ppt\/comments\/comment\d+\.xml/;
     const commentAuthorsRegex = /ppt\/commentAuthors\.xml/;
     const slideMastersRegex = /ppt\/slideMasters\/slideMaster\d+\.xml/;
+    const presentationFileRegex = /ppt\/presentation\.xml/;
     const files = await (0, zipUtils_js_1.extractFiles)(buffer, x => !!x.match(config.ignoreNotes ? slidesRegex : allFilesRegex) ||
         !!x.match(corePropsFileRegex) ||
         !!x.match(customPropsFileRegex) ||
@@ -64,7 +65,14 @@ const parsePowerPoint = async (buffer, config) => {
         !!x.match(slideRelsRegex) ||
         (!config.ignoreComments && (!!x.match(commentsFileRegex) || !!x.match(commentAuthorsRegex))) ||
         (!config.ignoreSlideMasters && !!x.match(slideMastersRegex)) ||
-        (!!config.extractAttachments && (!!x.match(mediaFileRegex) || !!x.match(chartFileRegex))), config.decompressionLimits);
+        !!x.match(presentationFileRegex) ||
+        (!!config.extractAttachments && (!!x.match(mediaFileRegex) || !!x.match(chartFileRegex))), config.decompressionLimits, config);
+    // ppt/presentation.xml is the part that makes an archive a presentation, and unlike the
+    // slides it is always present: PowerPoint can save a deck with no slides at all, so an
+    // empty ppt/slides/ is a warning rather than a failure.
+    (0, zipUtils_js_1.findRequiredPart)(files, path => !!path.match(presentationFileRegex), config, { fileType: 'pptx', part: 'ppt/presentation.xml' });
+    if (!files.some(file => !!file.path.match(slidesRegex)))
+        (0, errorUtils_js_1.logWarning)(types_js_1.OfficeWarningType.NO_SLIDES_FOUND, config);
     // Extract metadata
     const corePropsFile = files.find(f => f.path.match(corePropsFileRegex));
     const metadata = corePropsFile ? (0, xmlUtils_js_1.parseOfficeMetadata)(corePropsFile.content.toString()) : {};
@@ -89,7 +97,6 @@ const parsePowerPoint = async (buffer, config) => {
         return aNum - bNum;
     });
     const content = [];
-    const rawContents = [];
     const slideRelsMap = {};
     const authorMap = {};
     if (!config.ignoreComments) {
@@ -752,15 +759,24 @@ const parsePowerPoint = async (buffer, config) => {
             continue;
         if (file.path.match(slideRelsRegex))
             continue;
+        // All three document-property parts are extracted for metadata and were read earlier;
+        // only the core one was skipped here, leaving the other two to be parsed again as if
+        // they might be slides.
         if (file.path.match(corePropsFileRegex))
             continue;
+        if (file.path.match(appPropsFileRegex))
+            continue;
+        if (file.path.match(customPropsFileRegex))
+            continue;
         if (file.path.includes("comment"))
             continue;
+        // This loop treats every remaining file as a slide or note, so the presentation part
+        // has to be skipped explicitly: it carries no slide number and would otherwise be
+        // added to the deck as an empty slide.
+        if (file.path.match(presentationFileRegex))
+            continue;
         const xmlContentString = file.content.toString();
         const xml = (0, xmlUtils_js_1.parseXmlString)(xmlContentString, { locator: config.includeRawContent });
-        if (config.includeRawContent) {
-            rawContents.push(xmlContentString);
-        }
         const slideMatch = file.path.match(slideNumberRegex);
         const slideNumber = slideMatch ? parseInt(slideMatch[1]) : 0;
         const masterMatch = file.path.match(/slideMaster(\d+)\.xml/);

package/dist/parsers/WordParser.js

@@ -95,8 +95,12 @@ const parseWord = async (buffer, config) => {
     const footnotesFileRegex = /word\/footnotes[\d+]?.xml/;
     const endnotesFileRegex = /word\/endnotes[\d+]?.xml/;
     const commentsFileRegex = /word\/comments[\d+]?.xml/;
-    const headerFileRegex = /word\/header[\d+]?.xml/;
-    const footerFileRegex = /word\/footer[\d+]?.xml/;
+    // Headers and footers are the only parts a document can have many of: Word writes up to
+    // three per section (default, first page, even pages), so a handful of sections is enough
+    // to reach header10.xml. The single-character form the other parts use stops matching at
+    // nine, which would drop those later files as silently as not extracting them at all.
+    const headerFileRegex = /word\/header\d*\.xml/;
+    const footerFileRegex = /word\/footer\d*\.xml/;
     const numberingFileRegex = /word\/numbering[\d+]?.xml/;
     const mediaFileRegex = /(word\/)?media\/.*/;
     const corePropsFileRegex = /docProps\/core[\d+]?.xml/;
@@ -244,7 +248,12 @@ const parseWord = async (buffer, config) => {
         !!x.match(appPropsFileRegex) ||
         !!x.match(relsFileRegex) ||
         !!x.match(stylesFileRegex) ||
-        (!!config.extractAttachments && !!x.match(mediaFileRegex)), config.decompressionLimits);
+        (!config.ignoreComments && !!x.match(commentsFileRegex)) ||
+        (!config.ignoreHeadersAndFooters && (!!x.match(headerFileRegex) || !!x.match(footerFileRegex))) ||
+        (!!config.extractAttachments && !!x.match(mediaFileRegex)), config.decompressionLimits, config);
+    // A DOCX without its main document part is not a DOCX. Checked with the same regex the
+    // parse loop below uses to recognize it, so the two cannot fall out of step.
+    (0, zipUtils_js_1.findRequiredPart)(files, path => !!path.match(documentFileRegex), config, { fileType: 'docx', part: 'word/document.xml' });
     // Extract metadata
     const corePropsFile = files.find(f => f.path.match(corePropsFileRegex));
     const metadata = corePropsFile ? (0, xmlUtils_js_1.parseOfficeMetadata)(corePropsFile.content.toString()) : {};
@@ -405,7 +414,6 @@ const parseWord = async (buffer, config) => {
         }
     }
     const content = [];
-    const rawContents = [];
     const numberingState = {};
     const listCounters = {}; // Track item index per listId/level
     // Helper to parse a paragraph node
@@ -1092,9 +1100,6 @@ const parseWord = async (buffer, config) => {
         if (file.path.match(footerFileRegex))
             continue;
         const documentContent = file.content.toString();
-        if (config.includeRawContent) {
-            rawContents.push(documentContent);
-        }
         const doc = (0, xmlUtils_js_1.parseXmlString)(documentContent, { locator: config.includeRawContent });
         const body = (0, xmlUtils_js_1.getFirstElementByTagName)(doc, "w:body");
         if (body) {