officeparser 7.2.1 → 7.2.2

package/README.md

@@ -751,6 +751,7 @@ Pass as the second argument to `parseOffice(file, config)`.
 | `ignoreInternalLinks` | `boolean` | `false` | Strip bookmarks and internal cross-references from AST |
 | `fileType` | `SupportedFileType \| null` | `null` | **Required for text-based binary data** (`'md'`, `'html'`, `'csv'`) as these lack magic bytes. |
 | `csvDelimiter` | `string` | `','` | Input delimiter when parsing CSV files |
+| `decompressionLimits` | `DecompressionLimits` | `{ maxUncompressedBytes: 512MB, maxZipEntries: 10000 }` | **New**: Limits applied during ZIP extraction to protect against excessive memory and resource usage |
 | `pdfWorkerSrc` | `string` | CDN (jsDelivr) | Path/URL to `pdf.worker.min.mjs` (required in browser) |
 | `onWarning` | `(issue: OfficeIssue) => void` | — | Callback for non-fatal parsing issues |
 | `abortSignal` | `AbortSignal \| null` | `null` | Optional signal to cancel parsing (rejects with AbortError) |

package/dist/defaults.js

@@ -59,6 +59,10 @@ exports.DEFAULT_OFFICE_PARSER_CONFIG = {
     ignoreInternalLinks: false,
     fileType: null,
     csvDelimiter: ',',
+    decompressionLimits: {
+        maxUncompressedBytes: 512 * 1024 * 1024,
+        maxZipEntries: 10000,
+    },
 };
 /**
  * Default configuration for HTML generation.

package/dist/generators/ChunkingGenerator.js

@@ -644,7 +644,7 @@ class ChunkingGenerator extends BaseGenerator_js_1.BaseGenerator {
                     let timerId;
                     const timeoutPromise = new Promise((_, reject) => {
                         timerId = setTimeout(() => {
-                            reject(new Error(`Embedding call timed out after ${timeoutMs}ms`));
+                            reject((0, errorUtils_js_1.getOfficeError)(types_js_1.OfficeErrorType.EMBEDDING_TIMEOUT, this.config, timeoutMs));
                         }, timeoutMs);
                     });
                     return Promise.race([call, timeoutPromise]).finally(() => {

package/dist/generators/HtmlGenerator.js

@@ -469,7 +469,7 @@ class HtmlGenerator extends BaseGenerator_js_1.BaseGenerator {
                         src = `data:${attachment.mimeType || 'image/png'};base64,${attachment.data}`;
                     }
                 }
-                const img = `<img src="${src}" alt="${this.escape(node.text || meta?.altText || '')}"${className}${mappedAttrs}${styleAttr}>`;
+                const img = `<img src="${this.escape(src)}" alt="${this.escape(node.text || meta?.altText || '')}"${className}${mappedAttrs}${styleAttr}>`;
                 const content = this.config.includeFormatting ? `<div class="image-container">${img}<div class="caption">${this.escape(attachmentName || '')}</div></div>` : img;
                 return `${extraAnchors}<div${idAttr}>${content}</div>`;
             }

package/dist/officeparser.browser.d.ts

@@ -34,7 +34,15 @@ export declare enum OfficeErrorType {
 	/** Semantic chunking strategy is selected but no embedding function is provided */
 	MISSING_EMBEDDING_FUNCTION = "MISSING_EMBEDDING_FUNCTION",
 	/** The operation was aborted */
-	OPERATION_ABORTED = "OPERATION_ABORTED"
+	OPERATION_ABORTED = "OPERATION_ABORTED",
+	/** ZIP entry count exceeds limit */
+	ZIP_ENTRY_COUNT_LIMIT_EXCEEDED = "ZIP_ENTRY_COUNT_LIMIT_EXCEEDED",
+	/** ZIP entry missing a valid declared size */
+	ZIP_ENTRY_INVALID_SIZE = "ZIP_ENTRY_INVALID_SIZE",
+	/** ZIP uncompressed size limit exceeded */
+	ZIP_SIZE_LIMIT_EXCEEDED = "ZIP_SIZE_LIMIT_EXCEEDED",
+	/** Embedding call timed out */
+	EMBEDDING_TIMEOUT = "EMBEDDING_TIMEOUT"
 }
 /**
  * Standard warning types for OfficeParser.
@@ -349,6 +357,28 @@ export interface OfficeParserConfig {
 	 * Defaults to ',' but can be overridden (e.g., ';', '\t').
 	 */
 	csvDelimiter?: string;
+	/**
+	 * Limits and checks applied during ZIP extraction to protect against excessive
+	 * memory and resource usage.
+	 */
+	decompressionLimits?: DecompressionLimits;
+}
+/**
+ * Limits applied to ZIP archive decompression.
+ */
+export interface DecompressionLimits {
+	/**
+	 * Maximum allowed total uncompressed size (in bytes) of files extracted from a ZIP archive.
+	 * Applies to OOXML (DOCX, XLSX, PPTX) and ODF (ODT, ODP, ODS) formats.
+	 * Default is 536870912 (512 MB).
+	 */
+	maxUncompressedBytes?: number;
+	/**
+	 * Maximum allowed number of entries (files and directories) in a ZIP archive.
+	 * Applies to OOXML (DOCX, XLSX, PPTX) and ODF (ODT, ODP, ODS) formats.
+	 * Default is 10000.
+	 */
+	maxZipEntries?: number;
 }
 /**
  * Represents a single issue (warning, error, or info) generated during document processing.