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

rewriting with node support

pull/12/head
Kevin Kwok 8 years ago
parent
b5267918fd
commit
c5a1726e1d
31 changed files with 13339 additions and 711 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 26
      devServer.js
  2. 252
      dist/tesseract.js
  3. 12293
      dist/worker.js
  4. 0
      docs/tesseract_lang_list.md
  5. 0
      docs/tesseract_parameters.md
  6. 5
      examples/file-input/demo.html
  7. 23
      examples/node/test.js
  8. 30
      index.html
  9. 47
      package.json
  10. 138
      src/browser/index.js
  11. 61
      src/browser/lang.js
  12. 21
      src/browser/worker.js
  13. 2
      src/common/circularize.js
  14. 23
      src/common/desaturate.js
  15. 161
      src/common/dump.js
  16. 1
      src/common/langdata.json
  17. 143
      src/common/worker.js
  18. 137
      src/index.js
  19. 83
      src/node/index.js
  20. 36
      src/node/lang.js
  21. 22
      src/node/worker.js
  22. 26
      src/shared/desaturate.js
  23. 166
      src/shared/dump.js
  24. 2
      src/shared/fileSizes.js
  25. 3
      src/worker/db.js
  26. 53
      src/worker/detect.js
  27. 37
      src/worker/index.js
  28. 99
      src/worker/loadLanguage.js
  29. 56
      src/worker/recognize.js
  30. 42
      webpack.config.dev.js
  31. 46
      webpack.config.prod.js

26
devServer.js
Unescape View File

@ -1,26 +0,0 @@ @@ -1,26 +0,0 @@
var path = require('path');
var express = require('express');
var webpack = require('webpack');
var config = require('./webpack.config.dev');
var app = express();
var compiler = webpack(config);
app.use(require('webpack-dev-middleware')(compiler, {
noInfo: true,
publicPath: config[0].output.publicPath
}));
// app.use(require('webpack-hot-middleware')(compiler));
app.use('/', express.static('./'));
var port = 7355
app.listen(port, 'localhost', function(err) {
if (err) {
console.log(err);
return;
}
console.log('Listening at http://localhost:' + port);
});

252
dist/tesseract.js vendored
Unescape View File

@ -1 +1,251 @@ @@ -1 +1,251 @@
!function(e,t){"object"==typeof exports&&"object"==typeof module?module.exports=t():"function"==typeof define&&define.amd?define([],t):"object"==typeof exports?exports.Tesseract=t():e.Tesseract=t()}(this,function(){return function(e){function t(n){if(r[n])return r[n].exports;var o=r[n]={exports:{},id:n,loaded:!1};return e[n].call(o.exports,o,o.exports,t),o.loaded=!0,o.exports}var r={};return t.m=e,t.c=r,t.p="",t(0)}([function(e,t){"use strict";function r(e,t){return a||(a=o(g.coreUrl,g.workerUrl,g.langUrl)),a.recognize(e,t)}function n(e){return a||(a=o(g.coreUrl,g.workerUrl,g.langUrl)),a.detect(e)}function o(){function e(e,t){var r=s++;u[r]={};var n=0;return Object.getOwnPropertyNames(t).filter(function(e){return"function"==typeof t[e]}).forEach(function(o){n++,t[o](function(a){t[o]=a,0==--n&&i.postMessage({jobId:r,action:e,args:t})})}),0==n&&i.postMessage({jobId:r,action:e,args:t}),{then:function(e){return u[r].result=e,this},error:function(e){return u[r].error=e,this},progress:function(e){return u[r].progress=e,this}}}function t(e){if(e.match&&e.match(/^https?:\/\//))return function(r){var n=new Image;n.src=e,n.onload=function(){return r(t(n))}};if("string"==typeof e&&(e=document.querySelector(e)),e.getContext)e=e.getContext("2d");else if("IMG"==e.tagName||"VIDEO"==e.tagName){var r=document.createElement("canvas");r.width=e.naturalWidth||e.videoWidth,r.height=e.naturalHeight||e.videoHeight;var n=r.getContext("2d");n.drawImage(e,0,0),e=n}return e.getImageData&&(e=e.getImageData(0,0,e.canvas.width,e.canvas.height)),e}var r=arguments.length>0&&void 0!==arguments[0]?arguments[0]:g.coreUrl,n=arguments.length>1&&void 0!==arguments[1]?arguments[1]:g.workerUrl,o=arguments.length>2&&void 0!==arguments[2]?arguments[2]:g.langUrl,a=new Blob(["importScripts('"+r+"');\n\t\t importScripts('"+n+"');"]),i=new Worker(window.URL.createObjectURL(a)),c=!1,s=0,u={};return i.onmessage=function(e){var t=e.data,r=t.jobId,n=t.progress,o=t.error,a=t.result,i=u[r];n&&i.progress&&i.progress(n),o&&i.error&&i.error(o),a&&i.result&&i.result(a)},e("init",{mem:100663296,langUrl:o}),{detect:function(r){return e("detect",{image:t(r)})},recognize:function(r){var n=arguments.length>1&&void 0!==arguments[1]?arguments[1]:"eng";return"string"==typeof n?n={lang:n}:n.lang=n.lang||"eng",c||["chi_sim","chi_tra","jpn"].indexOf(n.lang)==-1||(e("init",{mem:167772160,langUrl:o}),c=!0),e("recognize",{options:n,image:t(r)})}}}var a,i="https://cdn.rawgit.com/naptha/tesseract.js-core/master/index.js",c="https://cdn.rawgit.com/naptha/tesseract.js/8b915dc/dist/tesseract.worker.js",s="https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/",g={coreUrl:i,workerUrl:c,langUrl:s,recognize:r,detect:n,createWorker:o};e.exports=g}])});
(function(f){if(typeof exports==="object"&&typeof module!=="undefined"){module.exports=f()}else if(typeof define==="function"&&define.amd){define([],f)}else{var g;if(typeof window!=="undefined"){g=window}else if(typeof global!=="undefined"){g=global}else if(typeof self!=="undefined"){g=self}else{g=this}g.Tesseract = f()}})(function(){var define,module,exports;return (function e(t,n,r){function s(o,u){if(!n[o]){if(!t[o]){var a=typeof require=="function"&&require;if(!u&&a)return a(o,!0);if(i)return i(o,!0);var f=new Error("Cannot find module '"+o+"'");throw f.code="MODULE_NOT_FOUND",f}var l=n[o]={exports:{}};t[o][0].call(l.exports,function(e){var n=t[o][1][e];return s(n?n:e)},l,l.exports,e,t,n,r)}return n[o].exports}var i=typeof require=="function"&&require;for(var o=0;o<r.length;o++)s(r[o]);return s})({1:[function(require,module,exports){
'use strict';
exports.defaultOptions = {
langPath: 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/',
workerPath: 'dist/worker.js',
tesseractPath: 'https://cdn.rawgit.com/naptha/tesseract.js-core/0.1.0/index.js'
};
exports.spawnWorker = function spawnWorker(instance, workerOptions) {
var worker = new Worker(workerOptions.workerPath);
worker.onmessage = function (e) {
instance._recv(e.data);
};
return worker;
};
exports.terminateWorker = function (instance) {
instance.worker.terminate();
};
exports.sendPacket = function sendPacket(instance, packet) {
loadImage(packet.payload.image, function (img) {
packet.payload.image = img;
instance.worker.postMessage(packet);
});
};
function loadImage(image, cb) {
if (typeof image === 'string') {
if (/^\#/.test(image)) {
// element css selector
return loadImage(document.querySelector(image), cb);
} else {
// url or path
var im = new Image();
im.src = image;
im.onload = function (e) {
return loadImage(im, cb);
};
return;
}
} else if (image instanceof File) {
// files
var fr = new FileReader();
fr.onload = function (e) {
return loadImage(fr.result, cb);
};
fr.readAsDataURL(image);
return;
} else if (image instanceof Blob) {
return loadImage(URL.createObjectURL(image), cb);
} else if (image.getContext) {
// canvas element
return loadImage(image.getContext('2d'), cb);
} else if (image.tagName == "IMG" || image.tagName == "VIDEO") {
// image element or video element
var c = document.createElement('canvas');
c.width = image.naturalWidth || image.videoWidth;
c.height = image.naturalHeight || image.videoHeight;
var ctx = c.getContext('2d');
ctx.drawImage(image, 0, 0);
return loadImage(ctx, cb);
} else if (image.getImageData) {
// canvas context
var data = image.getImageData(0, 0, image.canvas.width, image.canvas.height);
return loadImage(data, cb);
}
cb(image);
}
},{}],2:[function(require,module,exports){
"use strict";
var _createClass = function () { function defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } return function (Constructor, protoProps, staticProps) { if (protoProps) defineProperties(Constructor.prototype, protoProps); if (staticProps) defineProperties(Constructor, staticProps); return Constructor; }; }();
function _classCallCheck(instance, Constructor) { if (!(instance instanceof Constructor)) { throw new TypeError("Cannot call a class as a function"); } }
var adapter = require('./node/index.js');
function createWorker(workerOptions) {
return new TesseractWorker(workerOptions);
}
var TesseractWorker = function () {
function TesseractWorker(workerOptions) {
_classCallCheck(this, TesseractWorker);
this.worker = null;
this.workerOptions = workerOptions;
this._currentJob = null;
this._queue = [];
}
_createClass(TesseractWorker, [{
key: 'recognize',
value: function recognize(image, options) {
var _this = this;
return this._delay(function (job) {
options = options || {};
options.lang = options.lang || 'eng';
job._send('recognize', { image: image, options: options, workerOptions: _this.workerOptions });
});
}
}, {
key: 'detect',
value: function detect(image, options) {
var _this2 = this;
options = options || {};
return this._delay(function (job) {
job._send('detect', { image: image, options: options, workerOptions: _this2.workerOptions });
});
}
}, {
key: 'terminate',
value: function terminate() {
if (this.worker) adapter.terminateWorker(this);
this.worker = null;
}
}, {
key: '_delay',
value: function _delay(fn) {
var _this3 = this;
if (!this.worker) this.worker = adapter.spawnWorker(this, this.workerOptions);
var job = new TesseractJob(this);
this._queue.push(function (e) {
_this3._queue.shift();
_this3._currentJob = job;
fn(job);
});
if (!this._currentJob) this._dequeue();
return job;
}
}, {
key: '_dequeue',
value: function _dequeue() {
this._currentJob = null;
if (this._queue.length > 0) {
this._queue[0]();
}
}
}, {
key: '_recv',
value: function _recv(packet) {
if (this._currentJob.id === packet.jobId) {
this._currentJob._handle(packet);
} else {
console.warn('Job ID ' + packet.jobId + ' not known.');
}
}
}]);
return TesseractWorker;
}();
var jobCounter = 0;
var TesseractJob = function () {
function TesseractJob(instance) {
_classCallCheck(this, TesseractJob);
this.id = 'Job-' + ++jobCounter + '-' + Math.random().toString(16).slice(3, 8);
this._instance = instance;
this._resolve = [];
this._reject = [];
this._progress = [];
}
_createClass(TesseractJob, [{
key: 'then',
value: function then(resolve, reject) {
if (this._resolve.push) {
this._resolve.push(resolve);
} else {
resolve(this._resolve);
}
if (reject) this.catch(reject);
return this;
}
}, {
key: 'catch',
value: function _catch(reject) {
if (this._reject.push) {
this._reject.push(reject);
} else {
reject(this._reject);
}
return this;
}
}, {
key: 'progress',
value: function progress(fn) {
this._progress.push(fn);
return this;
}
}, {
key: '_send',
value: function _send(action, payload) {
adapter.sendPacket(this._instance, {
jobId: this.id,
action: action,
payload: payload
});
}
}, {
key: '_handle',
value: function _handle(packet) {
var data = packet.data;
if (packet.status === 'resolve') {
if (this._resolve.length === 0) console.debug(data);
this._resolve.forEach(function (fn) {
var ret = fn(data);
if (ret && typeof ret.then == 'function') {
console.warn('TesseractJob instances do not chain like ES6 Promises. To convert it into a real promise, use Promise.resolve.');
}
});
this._resolve = data;
this._instance._dequeue();
} else if (packet.status === 'reject') {
if (this._reject.length === 0) console.error(data);
this._reject.forEach(function (fn) {
return fn(data);
});
this._reject = data;
this._instance._dequeue();
} else if (packet.status === 'progress') {
this._progress.forEach(function (fn) {
return fn(data);
});
} else {
console.warn('Message type unknown', packet.status);
}
}
}]);
return TesseractJob;
}();
var DefaultTesseract = createWorker(adapter.defaultOptions);
DefaultTesseract.createWorker = createWorker;
module.exports = DefaultTesseract;
},{"./node/index.js":1}]},{},[2])(2)
});

12293
dist/worker.js vendored
View File

File diff suppressed because it is too large Load Diff

0
tesseract_lang_list.md → docs/tesseract_lang_list.md
Unescape View File

0
tesseract_parameters.md → docs/tesseract_parameters.md
Unescape View File

5
examples/file-input/demo.html
Unescape View File

@ -0,0 +1,5 @@ @@ -0,0 +1,5 @@
<script src="dist/tesseract.js"></script>
<input type="file" onchange="Tesseract.recognize(this.files[0]).progress(console.log)">
<input type="file" onchange="Tesseract.recognize(this.files[0], {lang: 'chi_sim'}).progress(console.log)">

23
examples/node/test.js
Unescape View File

@ -0,0 +1,23 @@ @@ -0,0 +1,23 @@
var Tesseract = require('./src/index.js')
global.Tesseract = Tesseract;
// Tesseract.recognize('yolop.png', {
// lang: 'eng'
// }).progress(function(info){
// console.log('--', info)
// })
// .then(function(data){
// console.log('--', data)
// })
// Tesseract.recognize('cosmic.jpg', {
// lang: 'eng'
// })
Tesseract.detect('cosmic.jpg')
.progress(function(info){
console.log(info)
})
.then(function(data){
console.log('done', data)
})

30
index.html
Unescape View File

@ -1,30 +0,0 @@ @@ -1,30 +0,0 @@
<canvas id="c"></canvas>
<script type="text/javascript" src="./tesseract/tesseract.js"></script>
<script type="text/javascript">
var canvas = document.getElementById('c')
canvas.width = 400
canvas.height = 400
var ctx = canvas.getContext('2d');
ctx.font = '30px "Arial Black"'
ctx.fillText('Hell0 World', 100, 40)
// ctx.fillText("囚犯離奇掙脫囚犯離奇掙脫", 100, 40)
ctx.font = '30px "Times New Roman"'
ctx.fillText('from beyond', 100, 80)
// ctx.fillText('2小時可換乘2次2小時可換乘2次', 100, 80)
ctx.font = '30px sans-serif'
ctx.fillText('the Cosmic Void', 100, 120)
Tesseract.workerUrl = location.protocol+'//'+location.host+'/tesseract/tesseract.worker.js'
// Tesseract.detect(canvas)
// Tesseract.recognize('http://localhost:7355/westmorland.jpg')
// Tesseract.recognize(canvas, { tessedit_char_blacklist: 'e' })
Tesseract.recognize(canvas)
.progress(function(e){
console.log('progress', e)
})
.then(function(e){
console.log('result', e)
})
</script>

47
package.json
Unescape View File

@ -2,34 +2,31 @@ @@ -2,34 +2,31 @@
"name": "tesseract.js",
"version": "1.0.0",
"description": "",
"main": "Tesseract.js",
"main": "index.js",
"scripts": {
"start": "node devServer.js",
"build": "webpack --config webpack.config.prod.js"
"test": "echo \"Error: no test specified\" & exit 1",
"start": "watchify src/index.js -o dist/tesseract.js --standalone Tesseract & watchify src/browser/worker.js -o dist/worker.js & http-server -p 7355",
"build": "browserify src/index.js -t [ babelify --presets [ es2015 ] ] -o dist/tesseract.js --standalone Tesseract && browserify src/browser/worker.js -t [ babelify --presets [ es2015 ] ] -o dist/worker.js"
},
"dependencies": {
"level-js": "^2.1.6",
"pako": "^0.2.7",
"tesseract.js-core": "^1.0.0"
},
"devDependencies": {
"babel": "^6.5.2",
"babel-core": "^6.7.0",
"babel-loader": "^6.2.4",
"express": "^4.13.4",
"webpack": "^1.13.0",
"webpack-dev-middleware": "^1.5.1",
"babel-preset-stage-1": "^6.5.0",
"babel-preset-es2015": "^6.6.0"
},
"repository": {
"type": "git",
"url": "https://github.com/naptha/tesseract.js.git"
"browser": {
"./src/node/index.js": "./src/browser/index.js"
},
"author": "",
"license": "MIT",
"bugs": {
"url": "https://github.com/naptha/tesseract.js/issues"
"license": "ISC",
"devDependencies": {
"babel-preset-es2015": "^6.16.0",
"babel-preset-react": "^6.16.0",
"babelify": "^7.3.0",
"browserify": "^13.1.0",
"http-server": "^0.9.0",
"watchify": "^3.7.0"
},
"homepage": "https://github.com/naptha/tesseract.js"
"dependencies": {
"file-type": "^3.8.0",
"jpeg-js": "^0.2.0",
"level-js": "^2.2.4",
"pako": "^1.0.3",
"png.js": "^0.2.1",
"tesseract.js-core": "^1.0.2"
}
}

138
src/browser/index.js
Unescape View File

@ -1,106 +1,64 @@ @@ -1,106 +1,64 @@
var coreUrl = 'https://cdn.rawgit.com/naptha/tesseract.js-core/master/index.js',
workerUrl = 'https://cdn.rawgit.com/naptha/tesseract.js/8b915dc/dist/tesseract.worker.js',
langUrl = 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/',
worker;
function recognize(image, options){
if(!worker) worker = createWorker( Tesseract.coreUrl, Tesseract.workerUrl, Tesseract.langUrl )
return worker.recognize(image, options)
exports.defaultOptions = {
langPath: 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/',
workerPath: 'dist/worker.js',
tesseractPath: 'https://cdn.rawgit.com/naptha/tesseract.js-core/0.1.0/index.js',
}
function detect(image){
if(!worker) worker = createWorker( Tesseract.coreUrl, Tesseract.workerUrl, Tesseract.langUrl )
return worker.detect(image)
exports.spawnWorker = function spawnWorker(instance, workerOptions){
var worker = new Worker(workerOptions.workerPath)
worker.onmessage = function(e){
instance._recv(e.data)
}
return worker
}
function createWorker(coreUrl=Tesseract.coreUrl, workerUrl=Tesseract.workerUrl, langUrl=Tesseract.langUrl){
var blob = new Blob([`importScripts('${coreUrl}');
importScripts('${workerUrl}');`])
var worker = new Worker(window.URL.createObjectURL(blob));
var bigworker = false
var jobCounter = 0
var handlers = {}
function runAsync(action, args){
var jobId = jobCounter++
handlers[jobId] = {}
exports.terminateWorker = function(instance){
instance.worker.terminate()
}
var waitingCount = 0
Object.getOwnPropertyNames(args)
.filter(name => typeof args[name] === 'function')
.forEach(name => {
waitingCount++
args[name](value => {
args[name] = value
if(--waitingCount == 0) worker.postMessage({jobId, action, args})
})
exports.sendPacket = function sendPacket(instance, packet){
loadImage(packet.payload.image, function(img){
packet.payload.image = img
instance.worker.postMessage(packet)
})
}
if(waitingCount == 0) worker.postMessage({jobId, action, args})
return {
then (f){ handlers[jobId].result = f; return this},
error (f){ handlers[jobId].error = f; return this},
progress(f){ handlers[jobId].progress = f; return this}
}
}
worker.onmessage = function(e){
var {jobId, progress, error, result} = e.data
var handler = handlers[jobId]
if(progress && handler.progress) handler.progress(progress);
if(error && handler.error) handler.error(error);
if(result && handler.result) handler.result(result);
}
function convertToImageData(image){
if(image.match && image.match(/^https?:\/\//)) {
return function thunk(cb){
var img = new Image()
img.src = image
img.onload = () => cb(convertToImageData(img))
}
function loadImage(image, cb){
if(typeof image === 'string'){
if(/^\#/.test(image)){
// element css selector
return loadImage(document.querySelector(image), cb)
}else{
// url or path
var im = new Image
im.src = image;
im.onload = e => loadImage(im, cb);
return
}
if(typeof image === 'string') image = document.querySelector(image)
if(image.getContext) image = image.getContext('2d');
else if(image.tagName == "IMG" || image.tagName == "VIDEO"){
}else if(image instanceof File){
// files
var fr = new FileReader()
fr.onload = e => loadImage(fr.result, cb);
fr.readAsDataURL(image)
return
}else if(image instanceof Blob){
return loadImage(URL.createObjectURL(image), cb)
}else if(image.getContext){
// canvas element
return loadImage(image.getContext('2d'), cb)
}else if(image.tagName == "IMG" || image.tagName == "VIDEO"){
// image element or video element
var c = document.createElement('canvas');
c.width = image.naturalWidth || image.videoWidth;
c.height = image.naturalHeight || image.videoHeight;
var ctx = c.getContext('2d');
ctx.drawImage(image, 0, 0);
image = ctx;
}
if(image.getImageData) image = image.getImageData(0, 0, image.canvas.width, image.canvas.height);
return image
}
runAsync('init', {mem: (1<<24) * 6, langUrl})
return {
detect(image){
return runAsync('detect', {image: convertToImageData(image)})
},
recognize(image, options='eng'){
if (typeof options === 'string') options = {lang: options};
else options.lang = options.lang || 'eng';
if (!bigworker && ['chi_sim', 'chi_tra', 'jpn'].indexOf(options.lang) != -1){
runAsync('init', {mem: (1<<24) * 10, langUrl})
bigworker = true
}
return runAsync('recognize', {options, image: convertToImageData(image)})
}
return loadImage(ctx, cb)
}else if(image.getImageData){
// canvas context
var data = image.getImageData(0, 0, image.canvas.width, image.canvas.height);
return loadImage(data, cb)
}
cb(image)
}
var Tesseract = {coreUrl, workerUrl, langUrl, recognize, detect, createWorker}
module.exports = Tesseract

61
src/browser/lang.js
Unescape View File

@ -0,0 +1,61 @@ @@ -0,0 +1,61 @@
var leveljs = require('level-js')
var db = typeof indexedDB === 'undefined' ? { open: (_, cb) => cb(true) } : leveljs('./tessdata2')
var langdata = require('../common/langdata.json')
module.exports = function getLanguageData(req, res, cb){
var lang = req.options.lang;
function saveDataFile(data){
db.put(lang, data, err => console.log('cached', lang, err))
cb(data)
}
db.open({ compression: false }, err => {
if (err) return fetchLanguageData(req, res, cb);
db.get(lang, (err, data) => {
if (err) return fetchLanguageData(req, res, saveDataFile);
res.progress({ status: 'found in cache ' + lang + '.traineddata' })
cb(data)
})
})
}
var ungzip = require('pako').ungzip;
function fetchLanguageData(req, res, cb){
var lang = req.options.lang;
var langfile = lang + '.traineddata.gz';
var url = req.workerOptions.langPath + langfile;
var xhr = new XMLHttpRequest();
xhr.responseType = 'arraybuffer';
xhr.open('GET', url, true);
xhr.onerror = e => {
xhr.onprogress = xhr.onload = null
cb(xhr, null)
}
xhr.onprogress = e =>
res.progress({
status: 'downloading ' + langfile,
loaded: e.loaded,
progress: Math.min(1, e.loaded / langdata[lang])
});
xhr.onload = e => {
if (!(xhr.status == 200 || (xhr.status == 0 && xhr.response))) return res.reject('Error downloading language ' + url);
res.progress({ status: 'unzipping ' + langfile })
// in case the gzips are already ungzipped or extra gzipped
var response = new Uint8Array(xhr.response)
try {
while(response[0] == 0x1f && response[1] == 0x8b) response = ungzip(response);
} catch (err) {
return res.reject('Error unzipping language file ' + langfile + '\n' + err.message)
}
cb(response)
}
xhr.send()
}

21
src/browser/worker.js
Unescape View File

@ -0,0 +1,21 @@ @@ -0,0 +1,21 @@
"use strict";
var workerUtils = require('../common/worker.js')
global.addEventListener('message', function(e){
var packet = e.data;
workerUtils.dispatchHandlers(packet, obj => postMessage(obj))
})
exports.getLanguageData = require('./lang.js')
exports.getCore = function(req, res){
if(!global.TesseractCore){
res.progress({ status: 'loading tesseract core' })
importScripts(req.workerOptions.tesseractPath)
res.progress({ status: 'loaded tesseract core' })
}
return TesseractCore
}
workerUtils.setAdapter(module.exports);

2
src/shared/circularize.js → src/common/circularize.js
Unescape View File

@ -1,4 +1,4 @@ @@ -1,4 +1,4 @@
export default function circularize(page){
module.exports = function circularize(page){
page.paragraphs = []
page.lines = []
page.words = []

23
src/common/desaturate.js
Unescape View File

@ -0,0 +1,23 @@ @@ -0,0 +1,23 @@
module.exports = function desaturate(image){
var width, height;
if(image.data){
var src = image.data;
width = image.width, height = image.height;
var dst = new Uint8Array(width * height);
var srcLength = src.length | 0, srcLength_16 = (srcLength - 16) | 0;
for (var i = 0, j = 0; i <= srcLength_16; i += 16, j += 4) {
// convert to grayscale 4 pixels at a time; eveything with alpha gets put in front of 50% gray
dst[j] = (((src[i] * 77 + src[i+1] * 151 + src[i+2] * 28) * src[i+3]) + ((255-src[i+3]) << 15) + 32768) >> 16
dst[j+1] = (((src[i+4] * 77 + src[i+5] * 151 + src[i+6] * 28) * src[i+7]) + ((255-src[i+7]) << 15) + 32768) >> 16
dst[j+2] = (((src[i+8] * 77 + src[i+9] * 151 + src[i+10] * 28) * src[i+11]) + ((255-src[i+11]) << 15) + 32768) >> 16
dst[j+3] = (((src[i+12] * 77 + src[i+13] * 151 + src[i+14] * 28) * src[i+15]) + ((255-src[i+15]) << 15) + 32768) >> 16
}
for (; i < srcLength; i += 4, ++j) //finish up
dst[j] = (((src[i] * 77 + src[i+1] * 151 + src[i+2] * 28) * src[i+3]) + ((255-src[i+3]) << 15) + 32768) >> 16
image = dst;
} else {
throw 'Expected ImageData'
}
return image
}

161
src/common/dump.js
Unescape View File

@ -0,0 +1,161 @@ @@ -0,0 +1,161 @@
function deindent(html){
var lines = html.split('\n')
if(lines[0].substring(0,2) === " "){
for (var i = 0; i < lines.length; i++) {
if (lines[i].substring(0,2) === " ") {
lines[i] = lines[i].slice(2)
}
};
}
return lines.join('\n')
}
module.exports = function DumpLiterallyEverything(Module, base){
var ri = base.GetIterator();
var blocks = [];
var block, para, textline, word, symbol;
function enumToString(value, prefix){
return (Object.keys(Module)
.filter(function(e){ return e.substr(0, prefix.length + 1) == prefix + '_' })
.filter(function(e){ return Module[e] === value })
.map(function(e){ return e.slice(prefix.length + 1) })[0])
}
ri.Begin()
do {
if(ri.IsAtBeginningOf(Module.RIL_BLOCK)){
var poly = ri.BlockPolygon();
var polygon = null;
// BlockPolygon() returns null when automatic page segmentation is off
if(Module.getPointer(poly) > 0){
var n = poly.get_n(),
px = poly.get_x(),
py = poly.get_y(),
polygon = [];
for(var i = 0; i < n; i++){
polygon.push([px.getValue(i), py.getValue(i)]);
}
Module._ptaDestroy(Module.getPointer(poly));
}
block = {
paragraphs: [],
text: ri.GetUTF8Text(Module.RIL_BLOCK),
confidence: ri.Confidence(Module.RIL_BLOCK),
baseline: ri.getBaseline(Module.RIL_BLOCK),
bbox: ri.getBoundingBox(Module.RIL_BLOCK),
blocktype: enumToString(ri.BlockType(), 'PT'),
polygon: polygon
}
blocks.push(block)
}
if(ri.IsAtBeginningOf(Module.RIL_PARA)){
para = {
lines: [],
text: ri.GetUTF8Text(Module.RIL_PARA),
confidence: ri.Confidence(Module.RIL_PARA),
baseline: ri.getBaseline(Module.RIL_PARA),
bbox: ri.getBoundingBox(Module.RIL_PARA),
is_ltr: !!ri.ParagraphIsLtr()
}
block.paragraphs.push(para)
}
if(ri.IsAtBeginningOf(Module.RIL_TEXTLINE)){
textline = {
words: [],
text: ri.GetUTF8Text(Module.RIL_TEXTLINE),
confidence: ri.Confidence(Module.RIL_TEXTLINE),
baseline: ri.getBaseline(Module.RIL_TEXTLINE),
bbox: ri.getBoundingBox(Module.RIL_TEXTLINE)
}
para.lines.push(textline)
}
if(ri.IsAtBeginningOf(Module.RIL_WORD)){
var fontInfo = ri.getWordFontAttributes(),
wordDir = ri.WordDirection();
word = {
symbols: [],
choices: [],
text: ri.GetUTF8Text(Module.RIL_WORD),
confidence: ri.Confidence(Module.RIL_WORD),
baseline: ri.getBaseline(Module.RIL_WORD),
bbox: ri.getBoundingBox(Module.RIL_WORD),
is_numeric: !!ri.WordIsNumeric(),
in_dictionary: !!ri.WordIsFromDictionary(),
direction: enumToString(wordDir, 'DIR'),
language: ri.WordRecognitionLanguage(),
is_bold: fontInfo.is_bold,
is_italic: fontInfo.is_italic,
is_underlined: fontInfo.is_underlined,
is_monospace: fontInfo.is_monospace,
is_serif: fontInfo.is_serif,
is_smallcaps: fontInfo.is_smallcaps,
font_size: fontInfo.pointsize,
font_id: fontInfo.font_id,
font_name: fontInfo.font_name,
}
var wc = new Module.WordChoiceIterator(ri);
do {
word.choices.push({
text: wc.GetUTF8Text(),
confidence: wc.Confidence()
})
} while (wc.Next());
Module.destroy(wc)
textline.words.push(word)
}
var image = null;
// var pix = ri.GetBinaryImage(Module.RIL_SYMBOL)
// var image = pix2array(pix);
// // for some reason it seems that things stop working if you destroy pics
// Module._pixDestroy(Module.getPointer(pix));
if(ri.IsAtBeginningOf(Module.RIL_SYMBOL)){
symbol = {
choices: [],
image: image,
text: ri.GetUTF8Text(Module.RIL_SYMBOL),
confidence: ri.Confidence(Module.RIL_SYMBOL),
baseline: ri.getBaseline(Module.RIL_SYMBOL),
bbox: ri.getBoundingBox(Module.RIL_SYMBOL),
is_superscript: !!ri.SymbolIsSuperscript(),
is_subscript: !!ri.SymbolIsSubscript(),
is_dropcap: !!ri.SymbolIsDropcap(),
}
word.symbols.push(symbol)
var ci = new Module.ChoiceIterator(ri);
do {
symbol.choices.push({
text: ci.GetUTF8Text(),
confidence: ci.Confidence()
})
} while (ci.Next());
Module.destroy(ci)
}
} while (ri.Next(Module.RIL_SYMBOL));
Module.destroy(ri)
return {
text: base.GetUTF8Text(),
html: deindent(base.GetHOCRText()),
confidence: base.MeanTextConf(),
blocks: blocks,
psm: enumToString(base.GetPageSegMode(), 'PSM'),
oem: enumToString(base.oem(), 'OEM'),
version: base.Version(),
}
}

1
src/common/langdata.json
Unescape View File

@ -0,0 +1 @@ @@ -0,0 +1 @@
{"afr": 1079573, "ara": 1701536, "aze": 1420865, "bel": 1276820, "ben": 6772012, "bul": 1605615, "cat": 1652368, "ces": 1035441, "chi_sim": 17710414, "chi_tra": 24717749, "chr": 320649, "dan-frak": 677656, "dan": 1972936, "deu-frak": 822644, "deu": 991656, "ell": 859719, "eng": 9453554, "enm": 619254, "epo": 1241212, "equ": 821130, "est": 1905040, "eus": 1641190, "fin": 979418, "fra": 1376221, "frk": 5912963, "frm": 5147082, "glg": 1674938, "grc": 3012615, "heb": 1051501, "hin": 6590065, "hrv": 1926995, "hun": 3074473, "ind": 1874776, "isl": 1634041, "ita": 948593, "ita_old": 3436571, "jpn": 13507168, "kan": 4390317, "kor": 5353098, "lav": 1843944, "lit": 1779240, "mal": 5966263, "meme": 88453, "mkd": 1163087, "mlt": 1463001, "msa": 1665427, "nld": 1134708, "nor": 2191610, "osd": 4274649, "pol": 7024662, "por": 909359, "ron": 915680, "rus": 5969957, "slk-frak": 289885, "slk": 2217342, "slv": 1611338, "spa": 883170, "spa_old": 5647453, "sqi": 1667041, "srp": 1770244, "swa": 757916, "swe": 2451917, "tam": 3498763, "tel": 5795246, "tgl": 1496256, "tha": 3811136, "tur": 3563264, "ukr": 937566, "vie": 2195922}

143
src/common/worker.js
Unescape View File

@ -0,0 +1,143 @@ @@ -0,0 +1,143 @@
var latestJob;
var Module;
var base;
var adapter = {};
function dispatchHandlers(packet, send){
function respond(status, data){
send({
jobId: packet.jobId,
status: status,
data: data
})
}
respond.resolve = respond.bind(this, 'resolve')
respond.reject = respond.bind(this, 'reject')
respond.progress = respond.bind(this, 'progress')
latestJob = respond;
if(packet.action === 'recognize'){
handleRecognize(packet.payload, respond)
}else if(packet.action === 'detect'){
handleDetect(packet.payload, respond)
}
}
exports.dispatchHandlers = dispatchHandlers;
exports.setAdapter = function setAdapter(impl){
adapter = impl;
}
function handleInit(req, res){
if(!Module){
var Core = adapter.getCore(req, res);
res.progress({ status: 'initializing tesseract api' })
Module = Core({
TOTAL_MEMORY: req.memory,
TesseractProgress(percent){
latestJob.progress({ status: 'recognizing text', progress: Math.max(0, (percent-30)/70) })
},
onRuntimeInitialized() {}
})
Module.FS_createPath("/", "tessdata", true, true)
base = new Module.TessBaseAPI()
res.progress({ status: 'initialized tesseract api' })
}
}
var dump = require('./dump.js')
var desaturate = require('./desaturate.js')
function setImage(Module, base, image){
var imgbin = desaturate(image),
width = image.width,
height = image.height;
var ptr = Module.allocate(imgbin, 'i8', Module.ALLOC_NORMAL);
base.SetImage(Module.wrapPointer(ptr), width, height, 1, width);
base.SetRectangle(0, 0, width, height)
return ptr;
}
function loadLanguage(req, res, cb){
var lang = req.options.lang;
if(!Module._loadedLanguages) Module._loadedLanguages = {};
if(lang in Module._loadedLanguages) return cb();
adapter.getLanguageData(req, res, function(data){
Module.FS_createDataFile('tessdata', lang + ".traineddata", data, true, false);
res.progress({ status: 'loaded ' + lang + '.traineddata' })
Module._loadedLanguages[lang] = true;
cb()
})
}
function handleRecognize(req, res){
handleInit(req, res)
loadLanguage(req, res, function(){
var lang = req.options.lang;
base.Init(null, lang)
res.progress({ status: 'initialized with language' })
var ptr = setImage(Module, base, req.image);
base.Recognize(null)
var result = dump(Module, base)
base.End();
Module._free(ptr);
res.resolve(result);
})
}
function handleDetect(req, res){
handleInit(req, res)
req.options.lang = 'osd';
loadLanguage(req, res, function(){
base.Init(null, 'osd')
base.SetPageSegMode(Module.PSM_OSD_ONLY)
var ptr = setImage(Module, base, req.image);
var results = new Module.OSResults();
var success = base.DetectOS(results);
if(!success){
base.End();
Module._free(ptr);
res.reject("failed to detect os")
} else {
var charset = results.get_unicharset()
var best = results.get_best_result()
var oid = best.get_orientation_id(),
sid = best.get_script_id();
var result = {
tesseract_script_id: sid,
script: charset.get_script_from_script_id(sid),
script_confidence: best.get_sconfidence(),
orientation_degrees: [0, 270, 180, 90][oid],
orientation_confidence: best.get_oconfidence()
}
base.End();
Module._free(ptr);
res.resolve(result)
}
})
}

137
src/index.js
Unescape View File

@ -0,0 +1,137 @@ @@ -0,0 +1,137 @@
"use strict";
var adapter = require('./node/index.js')
function createWorker(workerOptions){
return new TesseractWorker(workerOptions)
}
class TesseractWorker {
constructor(workerOptions){
this.worker = null;
this.workerOptions = workerOptions;
this._currentJob = null;
this._queue = []
}
recognize(image, options){
return this._delay(job => {
options = options || {}
options.lang = options.lang || 'eng';
job._send('recognize', { image: image, options: options, workerOptions: this.workerOptions })
})
}
detect(image, options){
options = options || {}
return this._delay(job => {
job._send('detect', { image: image, options: options, workerOptions: this.workerOptions })
})
}
terminate(){
if(this.worker) adapter.terminateWorker(this);
this.worker = null;
}
_delay(fn){
if(!this.worker) this.worker = adapter.spawnWorker(this, this.workerOptions);
var job = new TesseractJob(this);
this._queue.push(e => {
this._queue.shift()
this._currentJob = job;
fn(job)
})
if(!this._currentJob) this._dequeue();
return job
}
_dequeue(){
this._currentJob = null;
if(this._queue.length > 0){
this._queue[0]()
}
}
_recv(packet){
if(this._currentJob.id === packet.jobId){
this._currentJob._handle(packet)
}else{
console.warn('Job ID ' + packet.jobId + ' not known.')
}
}
}
var jobCounter = 0;
class TesseractJob {
constructor(instance){
this.id = 'Job-' + (++jobCounter) + '-' + Math.random().toString(16).slice(3, 8)
this._instance = instance;
this._resolve = []
this._reject = []
this._progress = []
}
then(resolve, reject){
if(this._resolve.push){
this._resolve.push(resolve)
}else{
resolve(this._resolve)
}
if(reject) this.catch(reject);
return this;
}
catch(reject){
if(this._reject.push){
this._reject.push(reject)
}else{
reject(this._reject)
}
return this;
}
progress(fn){
this._progress.push(fn)
return this;
}
_send(action, payload){
adapter.sendPacket(this._instance, {
jobId: this.id,
action: action,
payload: payload
})
}
_handle(packet){
var data = packet.data;
if(packet.status === 'resolve'){
if(this._resolve.length === 0) console.debug(data);
this._resolve.forEach(fn => {
var ret = fn(data);
if(ret && typeof ret.then == 'function'){
console.warn('TesseractJob instances do not chain like ES6 Promises. To convert it into a real promise, use Promise.resolve.')
}
})
this._resolve = data;
this._instance._dequeue()
}else if(packet.status === 'reject'){
if(this._reject.length === 0) console.error(data);
this._reject.forEach(fn => fn(data))
this._reject = data;
this._instance._dequeue()
}else if(packet.status === 'progress'){
this._progress.forEach(fn => fn(data))
}else{
console.warn('Message type unknown', packet.status)
}
}
}
var DefaultTesseract = createWorker(adapter.defaultOptions)
DefaultTesseract.createWorker = createWorker;
module.exports = DefaultTesseract

83
src/node/index.js
Unescape View File

@ -0,0 +1,83 @@ @@ -0,0 +1,83 @@
var path = require('path')
exports.defaultOptions = {
workerPath: path.join(__dirname, 'worker.js'),
langPath: 'http://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/',
}
var fork = require('child_process').fork;
var fs = require('fs')
exports.spawnWorker = function spawnWorker(instance, workerOptions){
var cp = fork(workerOptions.workerPath);
cp.on('message', function(packet){
instance._recv(packet)
})
return cp;
}
exports.terminateWorker = function(instance){
instance.worker.kill()
}
exports.sendPacket = function sendPacket(instance, packet){
loadImage(packet.payload.image, function(img){
packet.payload.image = img
instance.worker.send(packet)
})
}
function loadImage(image, cb){
if(typeof image === 'string'){
fs.readFile(image, function(err, buffer){
loadImage(buffer, cb)
})
return
}else if(image instanceof Buffer){
var fileType = require('file-type');
var mime = fileType(image).mime
if(mime === 'image/png'){
var PNGReader = require('png.js');
var reader = new PNGReader(image);
reader.parse(function(err, png){
if (err) throw err;
var image = {
width: png.getWidth(),
height: png.getHeight()
}
image.data = new Uint8Array(image.width * image.height * 4)
for(var j = 0; j < image.height; j++){
for(var i = 0; i < image.width; i++){
var offset = 4 * (i + j * image.width),
pix = png.getPixel(i, j);
image.data[offset] = pix[0]
image.data[offset + 1] = pix[1]
image.data[offset + 2] = pix[2]
image.data[offset + 3] = pix[3];
}
}
// console.log(image)
loadImage(image, cb)
});
return
}else if(mime === 'image/jpeg'){
var jpeg = require('jpeg-js');
loadImage(jpeg.decode(image), cb)
return
}
// TODO: support for TIFF, NetPBM, BMP, etc.
}
// node uses json.stringify for ipc which means we need to turn
// fancy arrays into raw arrays
if(image && image.data && image.data.length && !Array.isArray(image.data)){
image.data = Array.from(image.data)
return loadImage(image, cb)
}
cb(image)
}

36
src/node/lang.js
Unescape View File

@ -0,0 +1,36 @@ @@ -0,0 +1,36 @@
var http = require("http"),
zlib = require("zlib"),
fs = require("fs"),
path = require("path");
var langdata = require('../common/langdata.json')
function getLanguageData(req, res, cb){
var lang = req.options.lang;
var langfile = lang + '.traineddata.gz';
var url = req.workerOptions.langPath + langfile;
fs.readFile(lang + '.traineddata', function (err, data) {
if(!err) return cb(new Uint8Array(data));
http.get(url, function(stream){
var received_bytes = 0;
stream.on('data', function(chunk) {
received_bytes += chunk.length;
res.progress({
status: 'downloading ' + langfile,
loaded: received_bytes,
progress: Math.min(1, received_bytes / langdata[lang])
});
});
var gunzip = zlib.createGunzip();
stream.pipe(gunzip).pipe(fs.createWriteStream(lang + '.traineddata'))
gunzip.on('end', function(){ getLanguageData(req, stream, cb) })
})
});
}
module.exports = getLanguageData;

22
src/node/worker.js
Unescape View File

@ -0,0 +1,22 @@ @@ -0,0 +1,22 @@
"use strict";
var workerUtils = require('../common/worker.js')
process.on('message', function(packet){
workerUtils.dispatchHandlers(packet, obj => process.send(obj))
})
exports.getLanguageData = require('./lang.js')
var TesseractCore;
exports.getCore = function(req, res){
if(!TesseractCore){
res.progress({ status: 'loading tesseract core' })
TesseractCore = require('tesseract.js-core')
res.progress({ status: 'loaded tesseract core' })
}
return TesseractCore
}
workerUtils.setAdapter(module.exports);

26
src/shared/desaturate.js
Unescape View File

@ -1,26 +0,0 @@ @@ -1,26 +0,0 @@
export default function desaturate(image){
var width, height;
if(image.data){
var src = image.data;
width = image.width, height = image.height;
var dst = new Uint8Array(width * height);
var srcLength = src.length | 0, srcLength_16 = (srcLength - 16) | 0;
for (var i = 0, j = 0; i <= srcLength_16; i += 16, j += 4) {
// convert to grayscale 4 pixels at a time; eveything with alpha gets put in front of 50% gray
dst[j] = (((src[i] * 77 + src[i+1] * 151 + src[i+2] * 28) * src[i+3]) + ((255-src[i+3]) << 15) + 32768) >> 16
dst[j+1] = (((src[i+4] * 77 + src[i+5] * 151 + src[i+6] * 28) * src[i+7]) + ((255-src[i+7]) << 15) + 32768) >> 16
dst[j+2] = (((src[i+8] * 77 + src[i+9] * 151 + src[i+10] * 28) * src[i+11]) + ((255-src[i+11]) << 15) + 32768) >> 16
dst[j+3] = (((src[i+12] * 77 + src[i+13] * 151 + src[i+14] * 28) * src[i+15]) + ((255-src[i+15]) << 15) + 32768) >> 16
}
for (; i < srcLength; i += 4, ++j) //finish up
dst[j] = (((src[i] * 77 + src[i+1] * 151 + src[i+2] * 28) * src[i+3]) + ((255-src[i+3]) << 15) + 32768) >> 16
image = dst;
}
else {
throw 'Expected ImageData'
}
return image
}

166
src/shared/dump.js
Unescape View File

@ -1,166 +0,0 @@ @@ -1,166 +0,0 @@
function deindent(html){
var lines = html.split('\n')
if(lines[0].substring(0,2) === " "){
for (var i = 0; i < lines.length; i++) {
if (lines[i].substring(0,2) === " ") {
lines[i] = lines[i].slice(2)
}
};
}
return lines.join('\n')
}
export default function DumpLiterallyEverything(){
var {module, base} = self
var ri = base.GetIterator();
var blocks = [];
var block, para, textline, word, symbol;
function enumToString(value, prefix){
return (Object.keys(module)
.filter(function(e){ return e.substr(0, prefix.length + 1) == prefix + '_' })
.filter(function(e){ return module[e] === value })
.map(function(e){ return e.slice(prefix.length + 1) })[0])
}
const {RIL_BLOCK, RIL_PARA, RIL_TEXTLINE, RIL_WORD, RIL_SYMBOL} = module
ri.Begin()
do {
if(ri.IsAtBeginningOf(RIL_BLOCK)){
var poly = ri.BlockPolygon();
var polygon = null;
// BlockPolygon() returns null when automatic page segmentation is off
if(module.getPointer(poly) > 0){
var n = poly.get_n(),
px = poly.get_x(),
py = poly.get_y(),
polygon = [];
for(var i = 0; i < n; i++){
polygon.push([px.getValue(i), py.getValue(i)]);
}
module._ptaDestroy(module.getPointer(poly));
}
block = {
paragraphs: [],
text: ri.GetUTF8Text(RIL_BLOCK),
confidence: ri.Confidence(RIL_BLOCK),
baseline: ri.getBaseline(RIL_BLOCK),
bbox: ri.getBoundingBox(RIL_BLOCK),
blocktype: enumToString(ri.BlockType(), 'PT'),
polygon: polygon
}
blocks.push(block)
}
if(ri.IsAtBeginningOf(RIL_PARA)){
para = {
lines: [],
text: ri.GetUTF8Text(RIL_PARA),
confidence: ri.Confidence(RIL_PARA),
baseline: ri.getBaseline(RIL_PARA),
bbox: ri.getBoundingBox(RIL_PARA),
is_ltr: !!ri.ParagraphIsLtr()
}
block.paragraphs.push(para)
}
if(ri.IsAtBeginningOf(RIL_TEXTLINE)){
textline = {
words: [],
text: ri.GetUTF8Text(RIL_TEXTLINE),
confidence: ri.Confidence(RIL_TEXTLINE),
baseline: ri.getBaseline(RIL_TEXTLINE),
bbox: ri.getBoundingBox(RIL_TEXTLINE)
}
para.lines.push(textline)
}
if(ri.IsAtBeginningOf(RIL_WORD)){
var fontInfo = ri.getWordFontAttributes(),
wordDir = ri.WordDirection();
word = {
symbols: [],
choices: [],
text: ri.GetUTF8Text(RIL_WORD),
confidence: ri.Confidence(RIL_WORD),
baseline: ri.getBaseline(RIL_WORD),
bbox: ri.getBoundingBox(RIL_WORD),
is_numeric: !!ri.WordIsNumeric(),
in_dictionary: !!ri.WordIsFromDictionary(),
direction: enumToString(wordDir, 'DIR'),
language: ri.WordRecognitionLanguage(),
is_bold: fontInfo.is_bold,
is_italic: fontInfo.is_italic,
is_underlined: fontInfo.is_underlined,
is_monospace: fontInfo.is_monospace,
is_serif: fontInfo.is_serif,
is_smallcaps: fontInfo.is_smallcaps,
font_size: fontInfo.pointsize,
font_id: fontInfo.font_id,
font_name: fontInfo.font_name,
}
var wc = new module.WordChoiceIterator(ri);
do {
word.choices.push({
text: wc.GetUTF8Text(),
confidence: wc.Confidence()
})
} while (wc.Next());
module.destroy(wc)
textline.words.push(word)
}
var image = null;
// var pix = ri.GetBinaryImage(RIL_SYMBOL)
// var image = pix2array(pix);
// // for some reason it seems that things stop working if you destroy pics
// module._pixDestroy(module.getPointer(pix));
if(ri.IsAtBeginningOf(RIL_SYMBOL)){
symbol = {
choices: [],
image: image,
text: ri.GetUTF8Text(RIL_SYMBOL),
confidence: ri.Confidence(RIL_SYMBOL),
baseline: ri.getBaseline(RIL_SYMBOL),
bbox: ri.getBoundingBox(RIL_SYMBOL),
is_superscript: !!ri.SymbolIsSuperscript(),
is_subscript: !!ri.SymbolIsSubscript(),
is_dropcap: !!ri.SymbolIsDropcap(),
}
word.symbols.push(symbol)
var ci = new module.ChoiceIterator(ri);
do {
symbol.choices.push({
text: ci.GetUTF8Text(),
confidence: ci.Confidence()
})
} while (ci.Next());
module.destroy(ci)
}
} while (ri.Next(RIL_SYMBOL));
module.destroy(ri)
return {
text: base.GetUTF8Text(),
html: deindent(base.GetHOCRText()),
confidence: base.MeanTextConf(),
blocks: blocks,
psm: enumToString(base.GetPageSegMode(), 'PSM'),
oem: enumToString(base.oem(), 'OEM'),
version: base.Version(),
}
}

2
src/shared/fileSizes.js
Unescape View File

@ -1,2 +0,0 @@ @@ -1,2 +0,0 @@
const fileSizes = {"afr": 1079573, "ara": 1701536, "aze": 1420865, "bel": 1276820, "ben": 6772012, "bul": 1605615, "cat": 1652368, "ces": 1035441, "chi_sim": 17710414, "chi_tra": 24717749, "chr": 320649, "dan-frak": 677656, "dan": 1972936, "deu-frak": 822644, "deu": 991656, "ell": 859719, "eng": 9453554, "enm": 619254, "epo": 1241212, "equ": 821130, "est": 1905040, "eus": 1641190, "fin": 979418, "fra": 1376221, "frk": 5912963, "frm": 5147082, "glg": 1674938, "grc": 3012615, "heb": 1051501, "hin": 6590065, "hrv": 1926995, "hun": 3074473, "ind": 1874776, "isl": 1634041, "ita": 948593, "ita_old": 3436571, "jpn": 13507168, "kan": 4390317, "kor": 5353098, "lav": 1843944, "lit": 1779240, "mal": 5966263, "meme": 88453, "mkd": 1163087, "mlt": 1463001, "msa": 1665427, "nld": 1134708, "nor": 2191610, "osd": 4274649, "pol": 7024662, "por": 909359, "ron": 915680, "rus": 5969957, "slk-frak": 289885, "slk": 2217342, "slv": 1611338, "spa": 883170, "spa_old": 5647453, "sqi": 1667041, "srp": 1770244, "swa": 757916, "swe": 2451917, "tam": 3498763, "tel": 5795246, "tgl": 1496256, "tha": 3811136, "tur": 3563264, "ukr": 937566, "vie": 2195922}
export default fileSizes;

3
src/worker/db.js
Unescape View File

@ -1,3 +0,0 @@ @@ -1,3 +0,0 @@
import leveljs from 'level-js'
var db = typeof indexedDB === 'undefined' ? { open: (_, cb) => cb(true) } : leveljs('./tessdata')
export default db

53
src/worker/detect.js
Unescape View File

@ -1,53 +0,0 @@ @@ -1,53 +0,0 @@
import desaturate from '../shared/desaturate'
import loadLanguage from './loadLanguage'
export default function detect(jobId, image, cb){
var width = image.width, height = image.height;
image = desaturate(image)
var ptr = self.module.allocate(image, 'i8', self.module.ALLOC_NORMAL);
// console.log('allocated image')
loadLanguage(jobId, 'osd', err => {
self.module._free(ptr);
cb(err)
}, success => {
self.base.Init(null, 'osd')
self.base.SetPageSegMode(self.module.PSM_OSD_ONLY)
// console.log('loaded language')
self.base.SetImage(self.module.wrapPointer(ptr), width, height, 1, width)
self.base.SetRectangle(0, 0, width, height)
var results = new self.module.OSResults();
var success = self.base.DetectOS(results);
if(!success){
self.base.End();
self.module._free(ptr);
cb("failed to detect os")
}
else {
var charset = results.get_unicharset()
// console.log(charset)
// results.print_scores()
var best = results.get_best_result()
var oid = best.get_orientation_id(),
sid = best.get_script_id();
// console.log('orientation id', oid, [0, 270, 180, 90][oid], best.get_oconfidence())
// console.log('script id', sid, charset.get_script_from_script_id(sid), best.get_sconfidence())
// console.log(best)
cb(null, {
tesseract_script_id: sid,
script: charset.get_script_from_script_id(sid),
script_confidence: best.get_sconfidence(),
orientation_degrees: [0, 270, 180, 90][oid],
orientation_confidence: best.get_oconfidence()
})
self.base.End();
self.module._free(ptr);
}
})
}

37
src/worker/index.js
Unescape View File

@ -1,37 +0,0 @@ @@ -1,37 +0,0 @@
import recognize from './recognize'
import detect from './detect'
var module, base, jobId
onmessage = function(e) {
var {action, args} = e.data;
jobId = e.data.jobId
console.log('worker got action', action)
if(action == 'init'){
self.langUrl = args.langUrl
self.module = TesseractCore({
TOTAL_MEMORY: args.mem, //must be a multiple of 10 megabytes
TesseractProgress(percent){
postMessage({ jobId,
'progress': {
'recognized': Math.max(0,(percent-30)/70)
}
})
},
onRuntimeInitialized() {}
})
self.module.FS_createPath("/","tessdata",true,true)
self.base = new self.module.TessBaseAPI()
} else if(action === 'recognize'){
var {image, options} = args
recognize(jobId, image, options,
(error, result) => postMessage({jobId, error: error.message, result}))
} else if(action === 'detect'){
detect(jobId, args.image,
(error, result) => postMessage({jobId, error: error.message, result}))
}
}

99
src/worker/loadLanguage.js
Unescape View File

@ -1,99 +0,0 @@ @@ -1,99 +0,0 @@
import {ungzip} from 'pako'
import db from './db'
import fileSizes from '../shared/fileSizes'
function getLanguageData(lang, progress, cb){
var xhr = new XMLHttpRequest();
xhr.responseType = 'arraybuffer';
xhr.open('GET', self.langUrl + lang + '.traineddata.gz', true);
xhr.onerror = e => {
xhr.onprogress = xhr.onload = null
cb(xhr, null)
}
xhr.onprogress = e => progress({
'loaded_lang_model': e.loaded/fileSizes[lang], //this is kinda wrong on safari
cached: false
})
xhr.onload = e => {
if (!(xhr.status == 200 || (xhr.status == 0 && xhr.response))) return cb(xhr, null);
progress({'unzipping_lang_model': true})
var response = new Uint8Array(xhr.response)
while(response[0] == 0x1f && response[1] == 0x8b) response = ungzip(response);
progress({
'unzipped_lang_model': true,
'lang_model_size': response.length
})
cb(null, response)
}
progress({
'loaded_lang_model': 0,
cached: false,
requesting: true
})
xhr.send()
}
function load(lang, jobId, cb){
console.log('loadLanguage jobId', jobId)
function progressMessage(progress){
postMessage({ jobId, progress })
}
function finish(err, data) {
if(err) return cb(err);
// loaded_langs.push(lang)
cb(null, data)
}
function createDataFile(err, data){
progressMessage({ created_virtual_datafile: true})
finish(err, data)
}
function createDataFileCached(err, data) {
if(err) return createDataFile(err);
db.put(lang, data, err => console.log('cached', lang, err))
progressMessage({cached_lang: lang})
createDataFile(null, data)
}
db.open({compression: false}, err => {
if (err) return getLanguageData(lang, progressMessage, createDataFile);
db.get(lang, (err, data) => {
if (err) return getLanguageData(lang, progressMessage, createDataFileCached)
while(data[0] == 0x1f && data[1] == 0x8b) data = ungzip(data);
progressMessage({ loaded_lang_model: lang, from_cache: true })
cb(null, data)
})
})
}
var loaded_langs = []
export default function loadLanguage(jobId, lang, error, success){
if(loaded_langs.indexOf(lang) == -1) load(lang, jobId, function(err, result){
if(err) return error(err)
loaded_langs.push(lang)
self.module.FS_createDataFile('tessdata', lang +".traineddata", result, true, false);
success()
})
else success();
}

56
src/worker/recognize.js
Unescape View File

@ -1,56 +0,0 @@ @@ -1,56 +0,0 @@
import desaturate from '../shared/desaturate'
import loadLanguage from './loadLanguage'
import circularize from '../shared/circularize'
import dump from '../shared/dump'
var loaded_langs = []
export default function recognize(jobId, image, options, cb){
console.log('recognize id', jobId)
var {lang} = options
var width = image.width, height = image.height;
image = desaturate(image)
var ptr = self.module.allocate(image, 'i8', self.module.ALLOC_NORMAL);
loadLanguage(jobId, lang, err => {
self.module._free(ptr)
cb(err)
}, success => {
self.base.Init(null, lang)
postMessage({
jobId,
'progress': {
'initialized_with_lang': lang
}
})
for (var option in options) {
if (options.hasOwnProperty(option)) {
self.base.SetVariable(option, options[option]);
postMessage({
jobId: jobId,
'progress': {
'set_variable': {
variable: option,
value: options[option]
}
}
})
}
}
self.base.SetImage(self.module.wrapPointer(ptr), width, height, 1, width)
self.base.SetRectangle(0, 0, width, height)
// self.base.GetUTF8Text()
self.base.Recognize(null)
var everything = circularize(dump())
self.base.End();
self.module._free(ptr);
cb(null, everything)
})
}

42
webpack.config.dev.js
Unescape View File

@ -1,42 +0,0 @@ @@ -1,42 +0,0 @@
var path = require('path');
var webpack = require('webpack');
function config(opt) {
return {
devtool: 'cheap-module-eval-source-map',
entry: opt.entry,
output: Object.assign({}, opt.output, {
path: path.join(__dirname, 'build'),
publicPath: '/tesseract/',
}),
plugins: [
new webpack.NoErrorsPlugin()
],
module: {
loaders: [{
test: /\.js$/,
loaders: ['babel'],
include: opt.include
}]
},
node: {
fs: "empty"
}
}
}
module.exports = [{
entry: './src/browser/index.js',
output: {
filename: 'tesseract.js',
library: "Tesseract",
libraryTarget: "umd"
},
include: [path.join(__dirname, 'src/browser'), path.join(__dirname, 'src/shared')]
}, {
entry: './src/worker/index.js',
output: {
filename: 'tesseract.worker.js',
},
include: [path.join(__dirname, 'src/worker'), path.join(__dirname, 'src/shared')]
}].map(config);

46
webpack.config.prod.js
Unescape View File

@ -1,46 +0,0 @@ @@ -1,46 +0,0 @@
var path = require('path');
var webpack = require('webpack');
function config({entry, output, include}) {
return {
entry,
output: Object.assign({}, output, {
path: path.join(__dirname, 'dist')
}),
plugins: [
new webpack.optimize.OccurenceOrderPlugin(),
new webpack.optimize.DedupePlugin(),
new webpack.optimize.UglifyJsPlugin({
compressor: {
warnings: false
}
})
],
module: {
loaders: [{
test: /\.js$/,
loaders: ['babel'],
include
}]
},
node: {
fs: "empty"
}
}
}
module.exports = [{
entry: './src/browser/index.js',
output: {
filename: 'tesseract.js',
library: "Tesseract",
libraryTarget: "umd"
},
include: [path.join(__dirname, 'src/browser')]
}, {
entry: './src/worker/index.js',
output: {
filename: 'tesseract.worker.js',
},
include: [path.join(__dirname, 'src/worker')]
}].map(config);
Write Preview
Loading…
Cancel
Save