@iyulab/unpdf 0.14.0 → 0.17.0

package/package.json

@@ -2,7 +2,7 @@
   "name": "@iyulab/unpdf",
   "type": "module",
   "description": "High-performance PDF extraction to Markdown, text, and JSON (WebAssembly)",
-  "version": "0.14.0",
+  "version": "0.17.0",
   "license": "MIT",
   "repository": {
     "type": "git",

package/unpdf_wasm.d.ts

@@ -7,8 +7,21 @@ export class ParseOptions {
     lenient(): ParseOptions;
     constructor();
     textOnly(): ParseOptions;
+    /**
+     * Minimum pixel width/height for an extracted image to be kept — images
+     * below this on either axis are dropped as decorative (logos, rule lines,
+     * tracking pixels). Default 64; `0` keeps every image regardless of size.
+     */
+    withMinImageDimension(min_px: number): ParseOptions;
     withPages(from: number, to: number): ParseOptions;
     withPassword(password: string): ParseOptions;
+    /**
+     * Enable or disable resource (embedded image) extraction. Off by default —
+     * large PDFs loading every embedded image into memory is the largest peak-
+     * memory vector in this library. See `withMinImageDimension` to also tune
+     * which images count.
+     */
+    withResources(extract: boolean): ParseOptions;
 }
 
 export class PdfDocument {

package/unpdf_wasm_bg.js

@@ -38,6 +38,18 @@ export class ParseOptions {
         const ret = wasm.parseoptions_textOnly(ptr);
         return ParseOptions.__wrap(ret);
     }
+    /**
+     * Minimum pixel width/height for an extracted image to be kept — images
+     * below this on either axis are dropped as decorative (logos, rule lines,
+     * tracking pixels). Default 64; `0` keeps every image regardless of size.
+     * @param {number} min_px
+     * @returns {ParseOptions}
+     */
+    withMinImageDimension(min_px) {
+        const ptr = this.__destroy_into_raw();
+        const ret = wasm.parseoptions_withMinImageDimension(ptr, min_px);
+        return ParseOptions.__wrap(ret);
+    }
     /**
      * @param {number} from
      * @param {number} to
@@ -59,6 +71,19 @@ export class ParseOptions {
         const ret = wasm.parseoptions_withPassword(ptr, ptr0, len0);
         return ParseOptions.__wrap(ret);
     }
+    /**
+     * Enable or disable resource (embedded image) extraction. Off by default —
+     * large PDFs loading every embedded image into memory is the largest peak-
+     * memory vector in this library. See `withMinImageDimension` to also tune
+     * which images count.
+     * @param {boolean} extract
+     * @returns {ParseOptions}
+     */
+    withResources(extract) {
+        const ptr = this.__destroy_into_raw();
+        const ret = wasm.parseoptions_withResources(ptr, extract);
+        return ParseOptions.__wrap(ret);
+    }
 }
 if (Symbol.dispose) ParseOptions.prototype[Symbol.dispose] = ParseOptions.prototype.free;