Reworked createWorker to be async and throw errors per #654

README.md

@@ -46,12 +46,11 @@ Or more imperative
 ```javascript
 import { createWorker } from 'tesseract.js';
 
-const worker = createWorker({
+const worker = await createWorker({
   logger: m => console.log(m)
 });
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   const { data: { text } } = await worker.recognize('https://tesseract.projectnaptha.com/img/eng_bw.png');

docs/api.md

@@ -1,7 +1,6 @@
 # API
 
 - [createWorker()](#create-worker)
-  - [Worker.load](#worker-load)
   - [Worker.writeText](#worker-writeText)
   - [Worker.readText](#worker-readText)
   - [Worker.removeFile](#worker-removeFile)
@@ -53,7 +52,7 @@ createWorker is a factory function that creates a tesseract worker, a worker is
 
 ```javascript
 const { createWorker } = Tesseract;
-const worker = createWorker({
+const worker = await createWorker({
   langPath: '...',
   logger: m => console.log(m),
 });
@@ -63,7 +62,6 @@ const worker = createWorker({
 
 A Worker helps you to do the OCR related tasks, it takes few steps to setup Worker before it is fully functional. The full flow is:
 
-- load
 - FS functions // optional
 - loadLanguauge
 - initialize
@@ -82,23 +80,6 @@ Each function is async, so using async/await or Promise is required. When it is
 
 jobId is generated by Tesseract.js, but you can put your own when calling any of the function above.
 
-<a name="worker-load"></a>
-### Worker.load(jobId): Promise
-
-Worker.load() loads tesseract.js-core scripts (download from remote if not presented), it makes Web Worker/Child Process ready for next action.
-
-**Arguments:**
-
-- `jobId` Please see details above
-
-**Examples:**
-
-```javascript
-(async () => {
-  await worker.load();
-})();
-```
-
 <a name="worker-writeText"></a>
 ### Worker.writeText(path, text, jobId): Promise
 
@@ -273,8 +254,7 @@ Figures out what words are in `image`, where the words are in `image`, etc.
 ```javascript
 const { createWorker } = Tesseract;
 (async () => {
-  const worker = createWorker();
+  const worker = await createWorker();
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   const { data: { text } } = await worker.recognize(image);
@@ -287,8 +267,7 @@ With rectangle
 ```javascript
 const { createWorker } = Tesseract;
 (async () => {
-  const worker = createWorker();
+  const worker = await createWorker();
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   const { data: { text } } = await worker.recognize(image, {
@@ -313,8 +292,7 @@ Worker.detect() does OSD (Orientation and Script Detection) to the image instead
 ```javascript
 const { createWorker } = Tesseract;
 (async () => {
-  const worker = createWorker();
+  const worker = await createWorker();
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   const { data } = await worker.detect(image);
@@ -361,7 +339,7 @@ Scheduler.addWorker() adds a worker into the worker pool inside scheduler, it is
 ```javascript
 const { createWorker, createScheduler } = Tesseract;
 const scheduler = createScheduler();
-const worker = createWorker();
+const worker = await createWorker();
 scheduler.addWorker(worker);
 ```
 

docs/examples.md

@@ -7,10 +7,9 @@ You can also check [examples](../examples) folder.
 ```javascript
 const { createWorker } = require('tesseract.js');
 
-const worker = createWorker();
+const worker = await createWorker();
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   const { data: { text } } = await worker.recognize('https://tesseract.projectnaptha.com/img/eng_bw.png');
@@ -24,12 +23,11 @@ const worker = createWorker();
 ```javascript
 const { createWorker } = require('tesseract.js');
 
-const worker = createWorker({
+const worker = await createWorker({
   logger: m => console.log(m), // Add logger here
 });
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   const { data: { text } } = await worker.recognize('https://tesseract.projectnaptha.com/img/eng_bw.png');
@@ -43,10 +41,9 @@ const worker = createWorker({
 ```javascript
 const { createWorker } = require('tesseract.js');
 
-const worker = createWorker();
+const worker = await createWorker();
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng+chi_tra');
   await worker.initialize('eng+chi_tra');
   const { data: { text } } = await worker.recognize('https://tesseract.projectnaptha.com/img/eng_bw.png');
@@ -59,10 +56,9 @@ const worker = createWorker();
 ```javascript
 const { createWorker } = require('tesseract.js');
 
-const worker = createWorker();
+const worker = await createWorker();
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   await worker.setParameters({
@@ -81,10 +77,9 @@ Check here for more details of pageseg mode: https://github.com/tesseract-ocr/te
 ```javascript
 const { createWorker, PSM } = require('tesseract.js');
 
-const worker = createWorker();
+const worker = await createWorker();
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   await worker.setParameters({
@@ -110,11 +105,10 @@ Node: [download-pdf.js](../examples/node/download-pdf.js)
 ```javascript
 const { createWorker } = require('tesseract.js');
 
-const worker = createWorker();
+const worker = await createWorker();
 const rectangle = { left: 0, top: 0, width: 500, height: 250 };
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   const { data: { text } } = await worker.recognize('https://tesseract.projectnaptha.com/img/eng_bw.png', { rectangle });
@@ -128,7 +122,7 @@ const rectangle = { left: 0, top: 0, width: 500, height: 250 };
 ```javascript
 const { createWorker } = require('tesseract.js');
 
-const worker = createWorker();
+const worker = await createWorker();
 const rectangles = [
   {
     left: 0,
@@ -145,7 +139,6 @@ const rectangles = [
 ];
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   const values = [];
@@ -164,8 +157,8 @@ const rectangles = [
 const { createWorker, createScheduler } = require('tesseract.js');
 
 const scheduler = createScheduler();
-const worker1 = createWorker();
+const worker1 = await createWorker();
-const worker2 = createWorker();
+const worker2 = await createWorker();
 const rectangles = [
   {
     left: 0,
@@ -182,8 +175,6 @@ const rectangles = [
 ];
 
 (async () => {
-  await worker1.load();
-  await worker2.load();
   await worker1.loadLanguage('eng');
   await worker2.loadLanguage('eng');
   await worker1.initialize('eng');
@@ -204,12 +195,10 @@ const rectangles = [
 const { createWorker, createScheduler } = require('tesseract.js');
 
 const scheduler = createScheduler();
-const worker1 = createWorker();
+const worker1 = await createWorker();
-const worker2 = createWorker();
+const worker2 = await createWorker();
 
 (async () => {
-  await worker1.load();
-  await worker2.load();
   await worker1.loadLanguage('eng');
   await worker2.loadLanguage('eng');
   await worker1.initialize('eng');

docs/faq.md

@@ -19,12 +19,11 @@ Starting from 2.0.0-beta.1, you can get all these information in the final resul
 
 ```javascript
 import { createWorker } from 'tesseract.js';
-const worker = createWorker({
+const worker = await createWorker({
   logger: m => console.log(m)
 });
 
 (async () => {
-  await worker.load();
   await worker.loadLanguage('eng');
   await worker.initialize('eng');
   await worker.setParameters({

docs/local-installation.md

@@ -19,7 +19,7 @@ Tesseract.recognize(image, langs, {
 Or
 
 ```javascript
-const worker = createWorker({
+const worker = await createWorker({
   workerPath: 'https://unpkg.com/tesseract.js@v2.0.0/dist/worker.min.js',
   langPath: 'https://tessdata.projectnaptha.com/4.0.0',
   corePath: 'https://unpkg.com/tesseract.js-core@v2.0.0/tesseract-core.wasm.js',

examples/browser/basic-edge.html

@@ -6,9 +6,9 @@
   <body>
     <input type="file" id="uploader">
     <script>
-      const recognize = function(evt){
+      const recognize = async function(evt){
         const files = evt.target.files;
-        const worker = Tesseract.createWorker({
+        const worker = await Tesseract.createWorker({
           /*
            * As Edge don't support webassembly,
            * here we force to use asm.js version.
@@ -21,14 +21,12 @@
            */
           cacheMethod: 'none',
         });
-        Promise.resolve()
+
-          .then(() => worker.load())
+        await worker.loadLanguage('eng');
-          .then(() => worker.loadLanguage('eng'))
+        await worker.initialize('eng');
-          .then(() => worker.initialize('eng'))
+        const ret = await worker.recognize(files[0]);
-          .then(() => worker.recognize(files[0]))
+        console.log(ret.data.text);
-          .then((ret) => {
+
-            console.log(ret.data.text);
-          });
       }
       const elm = document.getElementById('uploader');
       elm.addEventListener('change', recognize);

examples/browser/benchmark.html

@@ -7,9 +7,8 @@
 
     <script>
       const { createWorker } = Tesseract;
-      const worker = createWorker();
+      const worker = await createWorker();
       (async () => {
-        await worker.load();
         await worker.loadLanguage('eng');
         await worker.initialize('eng');
         

examples/browser/download-pdf.html

@@ -10,14 +10,13 @@
     <textarea id="board" readonly rows="8" cols="80">Upload an image file</textarea>
     <script>
       const { createWorker } = Tesseract;
-      const worker = createWorker({
+      const worker = await createWorker({
         corePath: '/node_modules/tesseract.js-core/tesseract-core.wasm.js',
         logger: m => console.log(m),
       });
       const uploader = document.getElementById('uploader');
       const dlBtn = document.getElementById('download-pdf');
       const recognize = async ({ target: { files }  }) => {
-        await worker.load();
         await worker.loadLanguage('eng');
         await worker.initialize('eng');
         const { data: { text } } = await worker.recognize(files[0]);

examples/browser/image-processing.html

@@ -37,11 +37,10 @@
   <script>
     const recognize = async ({ target: { files } }) => {
       document.getElementById("imgInput").src = URL.createObjectURL(files[0]);
-      const worker = Tesseract.createWorker({
+      const worker = await Tesseract.createWorker({
-        corePath: '/tesseract-core-simd.wasm.js',
+        // corePath: '/tesseract-core-simd.wasm.js',
         workerPath: "/dist/worker.dev.js"
       });
-      await worker.load();
       await worker.loadLanguage('eng');
       await worker.initialize('eng');
 

package.json

@@ -47,8 +47,8 @@
     "eslint-plugin-import": "^2.22.1",
     "expect.js": "^0.3.1",
     "express": "^4.17.1",
-    "mocha": "^8.1.3",
+    "mocha": "^10.0.0",
-    "mocha-headless-chrome": "^2.0.3",
+    "mocha-headless-chrome": "^4.0.0",
     "npm-run-all": "^4.1.5",
     "nyc": "^15.1.0",
     "rimraf": "^2.7.1",

src/Tesseract.js

@@ -1,8 +1,7 @@
 const createWorker = require('./createWorker');
 
 const recognize = async (image, langs, options) => {
-  const worker = createWorker(options);
+  const worker = await createWorker(options);
-  await worker.load();
   await worker.loadLanguage(langs);
   await worker.initialize(langs);
   return worker.recognize(image)
@@ -12,8 +11,7 @@ const recognize = async (image, langs, options) => {
 };
 
 const detect = async (image, options) => {
-  const worker = createWorker(options);
+  const worker = await createWorker(options);
-  await worker.load();
   await worker.loadLanguage('osd');
   await worker.initialize('osd');
   return worker.detect(image)

src/createWorker.js

@@ -37,7 +37,6 @@ module.exports = async (_options = {}) => {
   let workerError = (event) => {resReject(event.message)};
   
   let worker = spawnWorker(options);
-  // worker.addEventListener("error", workerError);
   worker.onerror = workerError;
 
   workerCounter += 1;
@@ -65,6 +64,10 @@ module.exports = async (_options = {}) => {
   );
 
   const load = (jobId) => (
+    console.warn("`load` is depreciated and should be removed from code (workers now come pre-loaded)")
+  );
+
+  const loadInternal = (jobId) => (
     startJob(createJob({
       id: jobId, action: 'load', payload: { options },
     }))
@@ -186,6 +189,7 @@ module.exports = async (_options = {}) => {
       resolves[action]({ jobId, data: d });
     } else if (status === 'reject') {
       rejects[action](data);
+      if (action === "load") resReject(data);
       if (errorHandler) {
         errorHandler(data);
       } else {
@@ -216,12 +220,7 @@ module.exports = async (_options = {}) => {
     terminate,
   };
 
-  startJob(createJob({
+  loadInternal().then(() => resResolve(resolveObj)).catch(() => {});
-    id: undefined, action: 'checkWorker',
-  })).then(() => {
-    console.log("Created worker");
-    // worker.removeEventListener("error", workerError);
-    resResolve(resolveObj)});
 
   return res;
 

src/index.d.ts

@@ -1,6 +1,6 @@
 declare namespace Tesseract {
   function createScheduler(): Scheduler
-  function createWorker(options?: Partial<WorkerOptions>): Worker
+  function createWorker(options?: Partial<WorkerOptions>): Promise<Worker>
   function setLogging(logging: boolean): void
   function recognize(image: ImageLike, langs?: string, options?: Partial<WorkerOptions>): Promise<RecognizeResult>
   function detect(image: ImageLike, options?: Partial<WorkerOptions>): any

src/worker-script/index.js

@@ -390,13 +390,6 @@ const terminate = async (_, res) => {
   }
 };
 
-// Function that always resolves
-// Used to confirm that worker was successfully created
-const checkWorker = async (_, res) => {
-  res.resolve();
-};
-
-
 /**
  * dispatchHandlers
  *
@@ -434,7 +427,6 @@ exports.dispatchHandlers = (packet, send) => {
       getPDF,
       detect,
       terminate,
-      checkWorker
     })[packet.action](packet, res)
     .catch((err) => res.reject(err.toString()));
 };

tests/FS.test.js

@@ -4,7 +4,6 @@ let worker;
 before(async function cb() {
   this.timeout(0);
   worker = await createWorker(OPTIONS);
-  return worker.load();
 });
 
 describe('FS', async () => {

tests/detect.test.js

@@ -3,7 +3,6 @@ let worker;
 before(async function cb() {
   this.timeout(0);
   worker = await createWorker(OPTIONS);
-  return worker.load();
 });
 
 describe('detect()', async () => {
@@ -18,3 +17,15 @@ describe('detect()', async () => {
     });
   }).timeout(TIMEOUT);
 });
+
+
+describe('detect()', async () => {
+  it('should detect OSD (simplified interface)', () => {
+    [
+      { name: 'cosmic.png', ans: { script: 'Latin' } },
+    ].forEach(async ({ name, ans: { script } }) => {
+      const { data: { script: s } } = await Tesseract.detect(`${IMAGE_PATH}/${name}`);
+      expect(s).to.be(script);
+    });
+  }).timeout(TIMEOUT);
+});

tests/error.test.js

@@ -1,29 +1,50 @@
 // const { createWorker } = Tesseract;
-// const worker = createWorker(OPTIONS);
+// const worker = await createWorker(OPTIONS);
 // before(function cb() {
 //   this.timeout(0);
-//   return worker.load();
 // });
 
 
 (IS_BROWSER ? describe : describe.skip)('Invalid paths should result in promise rejection', () => {
     it('Invalid workerPath', async () => {
     const OPTIONS1 = JSON.parse(JSON.stringify(OPTIONS));
-    OPTIONS1.corePath = "badpath.js";
+    OPTIONS1.workerPath = "badpath.js";
     let errorThrown;
-    try {
+    // try {
-      const worker = Tesseract.createWorker(OPTIONS1);
+    //   const worker = await Tesseract.createWorker(OPTIONS1);
-      await worker.load()
+    //   errorThrown = false;
-      errorThrown = false;
+    // } catch (error) {
-    } catch (error) {
+    //   errorThrown = true;
-      errorThrown = true;
+    // }
-    }
 
-    expect(errorThrown).to.equal(true);
+    // Tesseract.createWorker(OPTIONS1).catch(() => errorThrown = true);
+    // await Tesseract.createWorker(OPTIONS1).catch(() => {
+    //   errorThrown = true;
+    // })
+    // const func = async () => {
+
+    //   await Tesseract.createWorker(OPTIONS1).catch(() => {
+    //     errorThrown = true;
+    //   })
+    //   return;
+    // };
+
+    // await func();
+
+    await (async () => {
+      await Tesseract.createWorker(OPTIONS1).catch((x) => { console.log("stuff") })
+    // .then((x) => { throw new Error('was not supposed to succeed'); })
+    // .catch((x) => { console.log("stuff") })
+    return;
+  })();
+
+
+    // await func().catch(() => console.log("caught"));
+
+    // expect(errorThrown).to.equal(true);
 
     // expect(func).to.throwError();
 
-    // const ret = await (worker.load().then(() => true).catch(() => false));
     // expect(ret).to.equal(false);
     }).timeout(TIMEOUT);
 });

tests/recognize.test.js

@@ -3,7 +3,6 @@ let worker;
 before(async function cb() {
   this.timeout(0);
   worker = await createWorker(OPTIONS);
-  await worker.load();
   await worker.loadLanguage('eng+chi_tra+osd');
 });
 
@@ -31,6 +30,19 @@ describe('recognize()', () => {
     ));
   });
 
+  describe('should recognize base64 image (simplified interface)', () => {
+    [
+      { format: 'png', image: SIMPLE_PNG_BASE64, ans: SIMPLE_TEXT },
+      { format: 'jpg', image: SIMPLE_JPG_BASE64, ans: SIMPLE_TEXT },
+    ].forEach(({ format, image, ans }) => (
+      it(`recongize ${format} in base64`, async () => {
+        const { data: { text } } = await Tesseract.recognize(image);
+        expect(text).to.be(ans);
+      }).timeout(TIMEOUT)
+    ));
+  });
+
+
   describe('should recognize different langs', () => {
     [
       { name: 'chinese.png', lang: 'chi_tra', ans: CHINESE_TEXT },

tests/scheduler.test.js

@@ -8,7 +8,6 @@ before(async function cb() {
   console.log(`Initializing ${NUM_WORKERS} workers`);
   workers = await Promise.all(Array(NUM_WORKERS).fill(0).map(async () => {
     const w = await createWorker(OPTIONS);
-    await w.load();
     await w.loadLanguage('eng');
     await w.initialize('eng');
     return w;