Updated Tesseract.js to fix memory leaks

scribeocr · Dec 28, 2024 · 267bef0 · 267bef0
1 parent 6f28688
commit 267bef0
Show file tree

Hide file tree

Showing 18 changed files with 1,431 additions and 1,405 deletions.
diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -50,6 +50,6 @@
   "dependencies": {
     "canvaskit-wasm": "^0.39.1",
     "commander": "^11.1.0",
-    "@scribe.js/tesseract.js": "^5.1.6"
+    "@scribe.js/tesseract.js": "^6.0.2"
   }
 }
diff --git a/tess/core/tesseract-core-lstm.wasm.js b/tess/core/tesseract-core-lstm.wasm.js
diff --git a/tess/core/tesseract-core-simd-lstm.wasm.js b/tess/core/tesseract-core-simd-lstm.wasm.js
diff --git a/tess/core/tesseract-core-simd.wasm.js b/tess/core/tesseract-core-simd.wasm.js
diff --git a/tess/core/tesseract-core.wasm.js b/tess/core/tesseract-core.wasm.js
diff --git a/tess/core_vanilla/tesseract-core-lstm.wasm.js b/tess/core_vanilla/tesseract-core-lstm.wasm.js
diff --git a/tess/core_vanilla/tesseract-core-simd-lstm.wasm.js b/tess/core_vanilla/tesseract-core-simd-lstm.wasm.js
diff --git a/tess/core_vanilla/tesseract-core-simd.wasm.js b/tess/core_vanilla/tesseract-core-simd.wasm.js
diff --git a/tess/core_vanilla/tesseract-core.wasm.js b/tess/core_vanilla/tesseract-core.wasm.js
diff --git a/tess/tesseract.esm.min.js b/tess/tesseract.esm.min.js
diff --git a/tess/tesseract.esm.min.js.map b/tess/tesseract.esm.min.js.map
diff --git a/tess/tesseract.min.js b/tess/tesseract.min.js
diff --git a/tess/tesseract.min.js.map b/tess/tesseract.min.js.map
diff --git a/tess/worker.min.js b/tess/worker.min.js
diff --git a/tess/worker.min.js.map b/tess/worker.min.js.map
diff --git a/tests/assets/italics_1.png b/tests/assets/italics_1.png
diff --git a/tests/module/recognize.spec.js b/tests/module/recognize.spec.js
@@ -26,6 +26,28 @@ describe('Check basic recognition features.', function () {
   }).timeout(10000);
 });
 
+describe('Check style detection.', function () {
+  this.timeout(20000);
+  before(async () => {
+    // This article page contains mostly italic text.
+    await scribe.importFiles([`${ASSETS_PATH_KARMA}/italics_1.png`]);
+    await scribe.recognize({
+      modeAdv: 'legacy',
+    });
+  });
+
+  it('Italic words are identified correctly', async () => {
+    assert.strictEqual(scribe.data.ocr.active[0].lines[0].words[3].style, 'normal');
+    assert.strictEqual(scribe.data.ocr.active[0].lines[0].words[4].style, 'italic');
+    assert.strictEqual(scribe.data.ocr.active[0].lines[0].words[5].style, 'italic');
+    assert.strictEqual(scribe.data.ocr.active[0].lines[0].words[6].style, 'normal');
+  }).timeout(10000);
+
+  after(async () => {
+    await scribe.terminate();
+  });
+});
+
 describe('Check font optimization features.', function () {
   this.timeout(20000);
   before(async () => {