officeparser 7.0.3 → 7.1.0

package/dist/parsers/CsvParser.js

@@ -2,6 +2,7 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parseCsv = void 0;
 const astUtils_js_1 = require("../utils/astUtils.js");
+const errorUtils_js_1 = require("../utils/errorUtils.js");
 /**
  * Parses a CSV file and extracts a single sheet with rows and cells.
  *
@@ -10,6 +11,10 @@ const astUtils_js_1 = require("../utils/astUtils.js");
  * @returns A promise resolving to the parsed AST
  */
 const parseCsv = async (buffer, config) => {
+    // Honour cancellation requests before the character-by-character parsing loop starts.
+    // CSV has no OCR or async I/O, but very large files can still occupy the thread for a
+    // noticeable duration, so short-circuiting on an aborted signal is still worthwhile.
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     const textStr = buffer.toString('utf-8');
     const delimiter = config.csvDelimiter;
     const records = [];

package/dist/parsers/ExcelParser.js

@@ -40,6 +40,10 @@ const zipUtils_js_1 = require("../utils/zipUtils.js");
  * @returns A promise resolving to the parsed AST
  */
 const parseExcel = async (buffer, config) => {
+    // Honour cancellation requests immediately — before extracting the ZIP archive.
+    // XLSX parsing involves decompressing multiple XML sheets and potentially running OCR
+    // on embedded chart images, so short-circuiting here saves significant work.
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     const sheetsRegex = /xl\/worksheets\/sheet\d+.xml/g;
     const drawingsRegex = /xl\/drawings\/drawing\d+.xml/g;
     const chartsRegex = /xl\/charts\/chart\d+.xml/g;
@@ -456,7 +460,7 @@ const parseExcel = async (buffer, config) => {
                     const typeMatch = cAttrs.match(/t="([a-zA-Z]+)"/);
                     const type = typeMatch ? typeMatch[1] : 'n'; // n = number (default)
                     const vMatch = cContent.match(/<v>([\s\S]*?)<\/v>/);
-                    const tMatch = cContent.match(/<t>([\s\S]*?)<\/t>/);
+                    const tMatch = cContent.match(/<t\b[^>]*>([\s\S]*?)<\/t>/);
                     let text = '';
                     let cellNodes = [];
                     if (type === 's' && vMatch) {
@@ -473,7 +477,7 @@ const parseExcel = async (buffer, config) => {
                         }
                     }
                     else if (type === 'inlineStr' && tMatch) {
-                        text = tMatch[1].trim();
+                        text = (0, xmlUtils_js_1.decodeXmlEntities)(tMatch[1].trim());
                     }
                     else if (vMatch) {
                         text = vMatch[1].trim();

package/dist/parsers/HtmlParser.js

@@ -2,6 +2,7 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parseHtml = void 0;
 const astUtils_js_1 = require("../utils/astUtils.js");
+const errorUtils_js_1 = require("../utils/errorUtils.js");
 const parseAttributes = (attrString) => {
     const attrs = {};
     const regex = /([a-zA-Z0-9\-:]+)(?:\s*=\s*(?:"([^"]*)"|'([^']*)'|([^\s>]+)))?/g;
@@ -95,6 +96,10 @@ const parseHtmlTree = (html) => {
     return root;
 };
 const parseHtml = async (buffer, config) => {
+    // Honour cancellation requests before the HTML tree is built and traversed.
+    // The custom recursive HTML parser can be expensive for large documents;
+    // rejecting early here prevents both the parsing and the subsequent AST construction.
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     const textStr = buffer.toString('utf-8');
     const root = parseHtmlTree(textStr);
     // Find head and body

package/dist/parsers/MarkdownParser.js

@@ -2,7 +2,12 @@
 Object.defineProperty(exports, "__esModule", { value: true });
 exports.parseMarkdown = void 0;
 const astUtils_js_1 = require("../utils/astUtils.js");
+const errorUtils_js_1 = require("../utils/errorUtils.js");
 const parseMarkdown = async (buffer, config) => {
+    // Honour cancellation requests before the line-by-line Markdown scanning loop begins.
+    // Markdown parsing is entirely synchronous and CPU-bound, so failing fast avoids
+    // processing content whose result will be discarded anyway.
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     let textStr = buffer.toString('utf-8');
     textStr = textStr.replace(/\r\n/g, '\n');
     const content = [];

package/dist/parsers/OpenOfficeParser.js

@@ -39,6 +39,10 @@ const zipUtils_js_1 = require("../utils/zipUtils.js");
  * @returns A promise resolving to the parsed AST
  */
 const parseOpenOffice = async (buffer, config) => {
+    // Honour cancellation requests immediately — before extracting the ZIP archive.
+    // ODF containers (ODT/ODS/ODP) bundle content.xml, styles.xml, and media files;
+    // aborting early avoids needlessly inflating and parsing all of those resources.
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     const contentFileRegex = /content\.xml/;
     const objectContentFileRegex = /Object \d+\/content\.xml/;
     const mediaFileRegex = /(Pictures|media)\/.*/;

package/dist/parsers/PdfParser.js

@@ -266,6 +266,7 @@ function convertToRgbaBuffer(data, width, height, kind) {
  * @returns Promise resolving to the parsed AST
  */
 const parsePdf = async (buffer, config) => {
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     const pdfjs = await (0, moduleLoader_js_1.loadPdfJs)();
     // Configure worker
     const workerSrc = config.pdfWorkerSrc;
@@ -398,6 +399,7 @@ const parsePdf = async (buffer, config) => {
     }
     // --- First Pass: Collect all items for font statistics ---
     for (let i = 1; i <= numPages; i++) {
+        (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
         let page;
         let textContent;
         const pageItems = [];
@@ -575,6 +577,7 @@ const parsePdf = async (buffer, config) => {
     const fontStats = calculateFontStats(allPageItems);
     // --- Second Pass: Process pages with font statistics ---
     for (let i = 0; i < allPageItems.length; i++) {
+        (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
         const pageNum = i + 1;
         let page;
         try {

package/dist/parsers/PowerPointParser.js

@@ -40,6 +40,10 @@ const zipUtils_js_1 = require("../utils/zipUtils.js");
  * @returns A promise resolving to the parsed AST
  */
 const parsePowerPoint = async (buffer, config) => {
+    // Honour cancellation requests immediately — before extracting the ZIP archive.
+    // PPTX presentations can have many slides with media/charts and optional OCR per image,
+    // so an early abort prevents decompressing and traversing data that will be discarded.
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     const allFilesRegex = /ppt\/(notesSlides|slides)\/(notesSlide|slide)\d+.xml/g;
     const slidesRegex = /ppt\/slides\/slide\d+.xml/g;
     const slideRelsRegex = /ppt\/slides\/_rels\/slide\d+\.xml\.rels/;

package/dist/parsers/RtfParser.js

@@ -359,6 +359,7 @@ exports.SimpleRtfParser = SimpleRtfParser;
  * @returns The parsed AST.
  */
 const parseRtf = async (buffer, config) => {
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     const parser = new SimpleRtfParser(buffer);
     const doc = parser.parse();
     // Extract font and color tables
@@ -1653,6 +1654,7 @@ const parseRtf = async (buffer, config) => {
     // Perform OCR if enabled
     if (config.ocr && config.extractAttachments) {
         for (const attachment of attachments) {
+            (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
             if (attachment.mimeType.startsWith('image/')) {
                 try {
                     // Convert base64 data back to Buffer for Tesseract.js

package/dist/parsers/WordParser.js

@@ -86,6 +86,10 @@ const zipUtils_js_1 = require("../utils/zipUtils.js");
  * @returns A promise resolving to the parsed AST
  */
 const parseWord = async (buffer, config) => {
+    // Honour cancellation requests immediately — before opening the ZIP archive, loading XML
+    // files, or kicking off any OCR work.  DOCX files can be large and the inflate + XML-parse
+    // steps are synchronous-heavy, so failing fast here avoids wasted CPU time.
+    (0, errorUtils_js_1.checkAbortSignal)(config.abortSignal);
     const documentFileRegex = /word\/document[\d+]?.xml/;
     const footnotesFileRegex = /word\/footnotes[\d+]?.xml/;
     const endnotesFileRegex = /word\/endnotes[\d+]?.xml/;