akira
/
tesseract.js
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
Browse Source

Update test & examples

develop
Jerome Wu 5 years ago
parent
d1c4a33c9e
commit
0e968d1829
13 changed files with 117 additions and 45 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Unified View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 2
      examples/browser/basic.html
  2. 52
      examples/browser/download-pdf.html
  3. 2
      examples/node/detect.js
  4. 22
      examples/node/download-pdf.js
  5. 7
      examples/node/recognize.js
  6. 17
      src/Tesseract.js
  7. 2
      src/createScheduler.js
  8. 9
      src/createWorker.js
  9. 2
      src/index.js
  10. 11
      src/utils/log.js
  11. 5
      src/worker-script/index.js
  12. 6
      tests/detect.test.js
  13. 25
      tests/recognize.test.js

2
examples/browser/basic.html
Unescape View File

@ -6,7 +6,7 @@
<input type="file" id="uploader"> <input type="file" id="uploader">
<script> <script>
const recognize = async ({ target: { files } }) => { const recognize = async ({ target: { files } }) => {
const { text } = await Tesseract.recognize(files[0], 'eng', { const { data: { text } } = await Tesseract.recognize(files[0], 'eng', {
corePath: '../../node_modules/tesseract.js-core/tesseract-core.wasm.js', corePath: '../../node_modules/tesseract.js-core/tesseract-core.wasm.js',
logger: m => console.log(m), logger: m => console.log(m),
}); });

52
examples/browser/download-pdf.html
Unescape View File

@ -0,0 +1,52 @@
<html>
<head>
<script src="/dist/tesseract.dev.js"></script>
</head>
<body>
<div>
<input type="file" id="uploader">
<button id="download-pdf" disabled="true">Download PDF</button>
</div>
<textarea id="board" readonly rows="8" cols="80">Upload an image file</textarea>
<script>
const { createWorker } = Tesseract;
const worker = 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]);
const board = document.getElementById('board');
board.value = text;
dlBtn.disabled = false;
};
const downloadPDF = async () => {
const filename = 'tesseract-ocr-result.pdf';
const { data } = await worker.getPDF('Tesseract OCR Result');
const blob = new Blob([new Uint8Array(data)], { type: 'application/pdf' });
if (navigator.msSaveBlob) {
// IE 10+
navigator.msSaveBlob(blob, filename);
} else {
const link = document.createElement('a');
if (link.download !== undefined) {
const url = URL.createObjectURL(blob);
link.setAttribute('href', url);
link.setAttribute('download', filename);
link.style.visibility = 'hidden';
document.body.appendChild(link);
link.click();
document.body.removeChild(link);
}
}
};
uploader.addEventListener('change', recognize);
dlBtn.addEventListener('click', downloadPDF);
</script>
</body>
</html>

2
examples/node/detect.js
Unescape View File

@ -8,6 +8,6 @@ const image = path.resolve(__dirname, (imagePath || '../../tests/assets/images/c
console.log(`Recognizing ${image}`); console.log(`Recognizing ${image}`);
Tesseract.detect(image, { logger: m => console.log(m) }) Tesseract.detect(image, { logger: m => console.log(m) })
.then((data) => { .then(({ data }) => {
console.log(data); console.log(data);
}); });

22
examples/node/download-pdf.js
Unescape View File

@ -0,0 +1,22 @@
#!/usr/bin/env node
const path = require('path');
const fs = require('fs');
const { createWorker } = require('../../');
const [,, imagePath] = process.argv;
const image = path.resolve(__dirname, (imagePath || '../../tests/assets/images/cosmic.png'));
console.log(`Recognizing ${image}`);
(async () => {
const worker = createWorker();
await worker.load();
await worker.loadLanguage('eng');
await worker.initialize('eng');
const { data: { text } } = await worker.recognize(image);
console.log(text);
const { data } = await worker.getPDF('Tesseract OCR Result');
fs.writeFileSync('tesseract-ocr-result.pdf', Buffer.from(data));
console.log('Generate PDF: tesseract-ocr-result.pdf');
await worker.terminate();
})();

7
examples/node/recognize.js
Unescape View File

@ -8,11 +8,6 @@ const image = path.resolve(__dirname, (imagePath || '../../tests/assets/images/c
console.log(`Recognizing ${image}`); console.log(`Recognizing ${image}`);
Tesseract.recognize(image, 'eng', { logger: m => console.log(m) }) Tesseract.recognize(image, 'eng', { logger: m => console.log(m) })
.then(({ text }) => { .then(({ data: { text } }) => {
console.log(text); console.log(text);
}); });
//Tesseract.detect(image, { logger: m => console.log(m) })
// .then((data) => {
// console.log(data);
// });

17
src/Tesseract.js
Unescape View File

@ -1,29 +1,24 @@
const createScheduler = require('./createScheduler');
const createWorker = require('./createWorker'); const createWorker = require('./createWorker');
const recognize = async (image, langs, options) => { const recognize = async (image, langs, options) => {
const scheduler = createScheduler();
const worker = createWorker(options); const worker = createWorker(options);
await worker.load(); await worker.load();
await worker.loadLanguage(langs); await worker.loadLanguage(langs);
await worker.initialize(langs); await worker.initialize(langs);
scheduler.addWorker(worker); return worker.recognize(image)
return scheduler.addJob('recognize', image) .finally(async () => {
.finally(() => { await worker.terminate();
scheduler.terminate();
}); });
}; };
const detect = async (image, options) => { const detect = async (image, options) => {
const scheduler = createScheduler();
const worker = createWorker(options); const worker = createWorker(options);
await worker.load(); await worker.load();
await worker.loadLanguage('osd'); await worker.loadLanguage('osd');
await worker.initialize('osd'); await worker.initialize('osd');
scheduler.addWorker(worker); return worker.detect(image)
return scheduler.addJob('detect', image) .finally(async () => {
.finally(() => { await worker.terminate();
scheduler.terminate();
}); });
}; };

2
src/createScheduler.js
Unescape View File

@ -1,5 +1,5 @@
const createJob = require('./createJob'); const createJob = require('./createJob');
const log = require('./utils/log'); const { log } = require('./utils/log');
const getId = require('./utils/getId'); const getId = require('./utils/getId');
let schedulerCounter = 0; let schedulerCounter = 0;

9
src/createWorker.js
Unescape View File

@ -1,7 +1,7 @@
const resolvePaths = require('./utils/resolvePaths'); const resolvePaths = require('./utils/resolvePaths');
const circularize = require('./utils/circularize'); const circularize = require('./utils/circularize');
const createJob = require('./createJob'); const createJob = require('./createJob');
const log = require('./utils/log'); const { log } = require('./utils/log');
const getId = require('./utils/getId'); const getId = require('./utils/getId');
const { defaultOEM } = require('./constants/config'); const { defaultOEM } = require('./constants/config');
const { const {
@ -106,9 +106,12 @@ module.exports = (_options = {}) => {
})) }))
); );
const terminate = async () => { const terminate = async (jobId) => {
if (worker !== null) { if (worker !== null) {
await startJob('terminate'); await startJob(createJob({
id: jobId,
action: 'terminate',
}));
terminateWorker(worker); terminateWorker(worker);
worker = null; worker = null;
} }

2
src/index.js
Unescape View File

@ -14,6 +14,7 @@ const createJob = require('./createJob');
const Tesseract = require('./Tesseract'); const Tesseract = require('./Tesseract');
const OEM = require('./constants/OEM'); const OEM = require('./constants/OEM');
const PSM = require('./constants/PSM'); const PSM = require('./constants/PSM');
const { setLogging } = require('./utils/log');
module.exports = { module.exports = {
OEM, OEM,
@ -21,5 +22,6 @@ module.exports = {
createScheduler, createScheduler,
createWorker, createWorker,
createJob, createJob,
setLogging,
...Tesseract, ...Tesseract,
}; };

11
src/utils/log.js
Unescape View File

@ -1,2 +1,9 @@
module.exports = (typeof process.env !== 'undefined' && process.env.NODE_ENV === 'development') let logging = false;
? console.log : () => {};
exports.logging = logging;
exports.setLogging = (_logging) => {
logging = _logging;
};
exports.log = (...args) => (logging ? console.log.apply(this, args) : null);

5
src/worker-script/index.js
Unescape View File

@ -15,7 +15,7 @@ const dump = require('./utils/dump');
const isBrowser = require('../utils/getEnvironment')('type') === 'browser'; const isBrowser = require('../utils/getEnvironment')('type') === 'browser';
const setImage = require('./utils/setImage'); const setImage = require('./utils/setImage');
const defaultParams = require('./constants/defaultParams'); const defaultParams = require('./constants/defaultParams');
const log = require('../utils/log'); const { log, setLogging } = require('../utils/log');
/* /*
* Tesseract Module returned by TesseractCore. * Tesseract Module returned by TesseractCore.
@ -29,7 +29,8 @@ let latestJob;
let adapter = {}; let adapter = {};
let params = defaultParams; let params = defaultParams;
const load = ({ workerId, jobId, payload: { options: { corePath } } }, res) => { const load = ({ workerId, jobId, payload: { options: { corePath, logging } } }, res) => {
setLogging(logging);
if (!TessModule) { if (!TessModule) {
const Core = adapter.getCore(corePath, res); const Core = adapter.getCore(corePath, res);

6
tests/detect.test.js
Unescape View File

@ -1,7 +1,5 @@
const { createScheduler, createWorker } = Tesseract; const { createWorker } = Tesseract;
const scheduler = createScheduler();
const worker = createWorker(OPTIONS); const worker = createWorker(OPTIONS);
scheduler.addWorker(worker);
before(function cb() { before(function cb() {
this.timeout(0); this.timeout(0);
return worker.load(); return worker.load();
@ -14,7 +12,7 @@ describe('detect()', async () => {
].forEach(async ({ name, ans: { script } }) => { ].forEach(async ({ name, ans: { script } }) => {
await worker.loadLanguage('osd'); await worker.loadLanguage('osd');
await worker.initialize('osd'); await worker.initialize('osd');
const { data: { script: s } } = await scheduler.addJob('detect', `${IMAGE_PATH}/${name}`); const { data: { script: s } } = await worker.detect(`${IMAGE_PATH}/${name}`);
expect(s).to.be(script); expect(s).to.be(script);
}); });
}).timeout(TIMEOUT); }).timeout(TIMEOUT);

25
tests/recognize.test.js
Unescape View File

@ -1,7 +1,5 @@
const { createScheduler, createWorker, PSM } = Tesseract; const { createWorker, PSM } = Tesseract;
const scheduler = createScheduler();
const worker = createWorker(OPTIONS); const worker = createWorker(OPTIONS);
scheduler.addWorker(worker);
before(async function cb() { before(async function cb() {
this.timeout(0); this.timeout(0);
await worker.load(); await worker.load();
@ -13,7 +11,7 @@ describe('recognize()', () => {
FORMATS.forEach(format => ( FORMATS.forEach(format => (
it(`support ${format} format`, async () => { it(`support ${format} format`, async () => {
await worker.initialize('eng'); await worker.initialize('eng');
const { data: { text } } = await scheduler.addJob('recognize', `${IMAGE_PATH}/simple.${format}`); const { data: { text } } = await worker.recognize(`${IMAGE_PATH}/simple.${format}`);
expect(text).to.be(SIMPLE_TEXT); expect(text).to.be(SIMPLE_TEXT);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));
@ -26,7 +24,7 @@ describe('recognize()', () => {
].forEach(({ format, image, ans }) => ( ].forEach(({ format, image, ans }) => (
it(`recongize ${format} in base64`, async () => { it(`recongize ${format} in base64`, async () => {
await worker.initialize('eng'); await worker.initialize('eng');
const { data: { text } } = await scheduler.addJob('recognize', image); const { data: { text } } = await worker.recognize(image);
expect(text).to.be(ans); expect(text).to.be(ans);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));
@ -38,7 +36,7 @@ describe('recognize()', () => {
].forEach(({ name, lang, ans }) => ( ].forEach(({ name, lang, ans }) => (
it(`recongize ${lang}`, async () => { it(`recongize ${lang}`, async () => {
await worker.initialize(lang); await worker.initialize(lang);
const { data: { text } } = await scheduler.addJob('recognize', `${IMAGE_PATH}/${name}`); const { data: { text } } = await worker.recognize(`${IMAGE_PATH}/${name}`);
expect(text).to.be(ans); expect(text).to.be(ans);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));
@ -52,7 +50,7 @@ describe('recognize()', () => {
].forEach(({ name, desc, ans }) => ( ].forEach(({ name, desc, ans }) => (
it(`recongize ${desc} image`, async () => { it(`recongize ${desc} image`, async () => {
await worker.initialize('eng'); await worker.initialize('eng');
const { data: { text } } = await scheduler.addJob('recognize', `${IMAGE_PATH}/${name}`); const { data: { text } } = await worker.recognize(`${IMAGE_PATH}/${name}`);
expect(text).to.be(ans); expect(text).to.be(ans);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));
@ -68,8 +66,7 @@ describe('recognize()', () => {
}) => ( }) => (
it(`recongize half ${name}`, async () => { it(`recongize half ${name}`, async () => {
await worker.initialize('eng'); await worker.initialize('eng');
const { data: { text } } = await scheduler.addJob( const { data: { text } } = await worker.recognize(
'recognize',
`${IMAGE_PATH}/${name}`, `${IMAGE_PATH}/${name}`,
{ {
rectangles: [ rectangles: [
@ -94,7 +91,7 @@ describe('recognize()', () => {
await worker.setParameters({ await worker.setParameters({
tessedit_pageseg_mode: mode, tessedit_pageseg_mode: mode,
}); });
const { data } = await scheduler.addJob('recognize', `${IMAGE_PATH}/simple.png`); const { data } = await worker.recognize(`${IMAGE_PATH}/simple.png`);
expect(Object.keys(data).length).not.to.be(0); expect(Object.keys(data).length).not.to.be(0);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));
@ -105,7 +102,7 @@ describe('recognize()', () => {
it(`support ${format} format`, async () => { it(`support ${format} format`, async () => {
const buf = fs.readFileSync(path.join(__dirname, 'assets', 'images', `simple.${format}`)); const buf = fs.readFileSync(path.join(__dirname, 'assets', 'images', `simple.${format}`));
await worker.initialize('eng'); await worker.initialize('eng');
const { data: { text } } = await scheduler.addJob('recognize', buf); const { data: { text } } = await worker.recognize(buf);
expect(text).to.be(SIMPLE_TEXT); expect(text).to.be(SIMPLE_TEXT);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));
@ -117,7 +114,7 @@ describe('recognize()', () => {
const imageDOM = document.createElement('img'); const imageDOM = document.createElement('img');
imageDOM.setAttribute('src', `${IMAGE_PATH}/simple.${format}`); imageDOM.setAttribute('src', `${IMAGE_PATH}/simple.${format}`);
await worker.initialize('eng'); await worker.initialize('eng');
const { data: { text } } = await scheduler.addJob('recognize', imageDOM); const { data: { text } } = await worker.recognize(imageDOM);
expect(text).to.be(SIMPLE_TEXT); expect(text).to.be(SIMPLE_TEXT);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));
@ -129,7 +126,7 @@ describe('recognize()', () => {
const videoDOM = document.createElement('video'); const videoDOM = document.createElement('video');
videoDOM.setAttribute('poster', `${IMAGE_PATH}/simple.${format}`); videoDOM.setAttribute('poster', `${IMAGE_PATH}/simple.${format}`);
await worker.initialize('eng'); await worker.initialize('eng');
const { data: { text } } = await scheduler.addJob('recognize', videoDOM); const { data: { text } } = await worker.recognize(videoDOM);
expect(text).to.be(SIMPLE_TEXT); expect(text).to.be(SIMPLE_TEXT);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));
@ -161,7 +158,7 @@ describe('recognize()', () => {
formats.forEach(format => ( formats.forEach(format => (
it(`support ${format} format`, async () => { it(`support ${format} format`, async () => {
await worker.initialize('eng'); await worker.initialize('eng');
const { data: { text } } = await scheduler.addJob('recognize', canvasDOM); const { data: { text } } = await worker.recognize(canvasDOM);
expect(text).to.be(SIMPLE_TEXT); expect(text).to.be(SIMPLE_TEXT);
}).timeout(TIMEOUT) }).timeout(TIMEOUT)
)); ));

Write Preview
Loading…
Cancel
Save