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

updating readme

pull/12/head
Kevin Kwok 8 years ago
parent
ab13d7b891
commit
036f8debfc
9 changed files with 601 additions and 504 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. 3
      .gitignore
  2. 62
      README.md
  3. 462
      dist/tesseract.js
  4. 552
      dist/worker.js
  5. 2
      examples/file-input/demo.html
  6. 2
      package.json
  7. 15
      src/browser/index.js
  8. 6
      src/index.js
  9. 1
      src/node/index.js

3
.gitignore vendored
Unescape View File

@ -1,2 +1,3 @@
.DS_Store .DS_Store
node_modules/* node_modules/*
yarn.lock

62
README.md
Unescape View File

@ -27,15 +27,19 @@ Tesseract.js works with a `<script>` tag via local copy or cdn, with webpack and
You can either include Tesseract.js on you page with a cdn like this: You can either include Tesseract.js on you page with a cdn like this:
```html ```html
<script src='https://cdn.rawgit.com/naptha/tesseract.js/a01d2a2/dist/tesseract.js'></script> <script src='https://cdn.rawgit.com/naptha/tesseract.js/0.2.0/dist/tesseract.js'></script>
``` ```
Or you can grab copies of `tesseract.js` and `tesseract.worker.js` from the [dist folder](https://github.com/naptha/tesseract.js/tree/master/dist) and include your local copies like this: Or you can grab copies of `tesseract.js` and `worker.js` from the [dist folder](https://github.com/naptha/tesseract.js/tree/master/dist) and include your local copies like this:
```html ```html
<script src='/path/to/tesseract.js'></script> <script src='/path/to/tesseract.js'></script>
<script> <script>
Tesseract.workerUrl = 'http://www.absolute-path-to/tesseract.worker.js' var LocalTesseract = Tesseract.create({
workerPath: '/path/to/worker.js',
langPath: 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/',
tesseractPath: 'https://cdn.rawgit.com/naptha/tesseract.js-core/0.1.0/index.js',
})
// from now on use LocalTesseract instead of Tesseract
</script> </script>
``` ```
@ -69,7 +73,7 @@ You can [head to the docs](#docs) for a full treatment of the API.
* [TesseractJob](#tesseractjob) * [TesseractJob](#tesseractjob)
+ [TesseractJob.progress(callback: function) -> TesseractJob](#tesseractjobprogresscallback-function---tesseractjob) + [TesseractJob.progress(callback: function) -> TesseractJob](#tesseractjobprogresscallback-function---tesseractjob)
+ [TesseractJob.then(callback: function) -> TesseractJob](#tesseractjobthencallback-function---tesseractjob) + [TesseractJob.then(callback: function) -> TesseractJob](#tesseractjobthencallback-function---tesseractjob)
+ [TesseractJob.error(callback: function) -> TesseractJob](#tesseractjoberrorcallback-function---tesseractjob) + [TesseractJob.catch(callback: function) -> TesseractJob](#tesseractjoberrorcallback-function---tesseractjob)
* [Tesseract Remote File Options](#tesseract-remote-file-options) * [Tesseract Remote File Options](#tesseract-remote-file-options)
+ [Tesseract.coreUrl](#tesseractcoreurl) + [Tesseract.coreUrl](#tesseractcoreurl)
+ [Tesseract.workerUrl](#tesseractworkerurl) + [Tesseract.workerUrl](#tesseractworkerurl)
@ -87,11 +91,11 @@ Figures out what words are in `image`, where the words are in `image`, etc.
+ include properties that override some subset of the [default tesseract parameters](./tesseract_parameters.md) + include properties that override some subset of the [default tesseract parameters](./tesseract_parameters.md)
+ include a `lang` property with a value from the [list of lang parameters](./tesseract_lang_list.md) + include a `lang` property with a value from the [list of lang parameters](./tesseract_lang_list.md)
Returns a [TesseractJob](#tesseractjob) whose `then`, `progress`, and `error` methods can be used to act on the result. Returns a [TesseractJob](#tesseractjob) whose `then`, `progress`, and `catch` methods can be used to act on the result.
### Simple Example: ### Simple Example:
```javascript ```javascript
Tesseract.recognize('#my-image') Tesseract.recognize(document.querySelector('#my-image'))
.then(function(result){ .then(function(result){
console.log(result) console.log(result)
}) })
@ -100,7 +104,7 @@ Tesseract.recognize('#my-image')
### More Complicated Example: ### More Complicated Example:
```javascript ```javascript
// if we know our image is of spanish words without the letter 'e': // if we know our image is of spanish words without the letter 'e':
Tesseract.recognize('#my-image', { Tesseract.recognize(document.querySelector('#my-image'), {
lang: 'spa', lang: 'spa',
tessedit_char_blacklist: 'e' tessedit_char_blacklist: 'e'
}) })
@ -113,14 +117,16 @@ Tesseract.recognize('#my-image', {
## Tesseract.detect(image: [ImageLike](#imagelike)) -> [TesseractJob](#tesseractjob) ## Tesseract.detect(image: [ImageLike](#imagelike)) -> [TesseractJob](#tesseractjob)
Figures out what script (e.g. 'Latin', 'Chinese') the words in image are written in. Figures out what script (e.g. 'Latin', 'Chinese') the words in image are written in.
- `image` is any [ImageLike](#imagelike) object. - `image` is any [ImageLike](#imagelike) object.
Returns a [TesseractJob](#tesseractjob) whose `then`, `progress`, and `error` methods can be used to act on the result of the script. Returns a [TesseractJob](#tesseractjob) whose `then`, `progress`, and `error` methods can be used to act on the result of the script.
```javascript ```javascript
Tesseract.detect('#my-image') Tesseract.detect(document.querySelector('#my-image'))
.then(function(result){ .then(function(result){
console.log(result) console.log(result)
}) })
@ -128,13 +134,25 @@ Tesseract.detect('#my-image')
## ImageLike ## ImageLike
The main Tesseract.js functions take an `image` parameter, which should be something that is 'image-like'.
That means `image` should be The main Tesseract.js functions take an `image` parameter, which should be something that is like an image. What's considered "image-like" differs depending on whether it is being run from the browser or through NodeJS.
- an `img` element or querySelector that matches an `img` element
- a `video` element or querySelector that matches a `video` element
- a `canvas` element or querySelector that matches a `canvas` element On a browser, an image can be:
- an `img`, `video`, or `canvas` element
- a CanvasRenderingContext2D (returned by `canvas.getContext('2d')`) - a CanvasRenderingContext2D (returned by `canvas.getContext('2d')`)
- the absolute `url` of an image from the same website that is running your script. Browser security policies don't allow access to the content of images from other websites :( - a `File` object (from a file `<input>` or drag-drop event)
- a `Blob` object
- a `ImageData` instance (an object containing `width`, `height` and `data` properties)
- a path or URL to an accessible image (the image must either be hosted locally or accessible by CORS)
In NodeJS, an image can be
- a path to a local image
- a `Buffer` instance containing a `PNG` or `JPEG` image
- a `ImageData` instance (an object containing `width`, `height` and `data` properties)
## TesseractJob ## TesseractJob
@ -143,19 +161,19 @@ All methods of a given TesseractJob return that TesseractJob to enable chaining.
Typical use is: Typical use is:
```javascript ```javascript
Tesseract.recognize('#my-image') Tesseract.recognize(document.querySelector('#my-image'))
.progress(function(message){console.log(message)}) .progress(function(message){console.log(message)})
.error(function(err){console.error(err)}) .catch(function(err){console.error(err)})
.then(function(result){console.log(result)}) .then(function(result){console.log(result)})
``` ```
Which is equivalent to: Which is equivalent to:
```javascript ```javascript
var job1 = Tesseract.recognize('#my-image'); var job1 = Tesseract.recognize(document.querySelector('#my-image'));
job1.progress(function(message){console.log(message)}); job1.progress(function(message){console.log(message)});
job1.error(function(err){console.error(err)}); job1.catch(function(err){console.error(err)});
job1.then(function(result){console.log(result)}) job1.then(function(result){console.log(result)})
``` ```
@ -168,7 +186,7 @@ Sets `callback` as the function that will be called every time the job progresse
For example: For example:
```javascript ```javascript
Tesseract.recognize('#my-image') Tesseract.recognize(document.querySelector('#my-image'))
.progress(function(message){console.log('progress is: 'message)}) .progress(function(message){console.log('progress is: 'message)})
``` ```
@ -193,7 +211,7 @@ Sets `callback` as the function that will be called if and when the job successf
For example: For example:
```javascript ```javascript
Tesseract.recognize('#my-image') Tesseract.recognize(document.querySelector('#my-image'))
.then(function(result){console.log('result is: 'result)}) .then(function(result){console.log('result is: 'result)})
``` ```
@ -214,7 +232,7 @@ progress is: {
} }
``` ```
### TesseractJob.error(callback: function) -> TesseractJob ### TesseractJob.catch(callback: function) -> TesseractJob
Sets `callback` as the function that will be called if the job fails. Sets `callback` as the function that will be called if the job fails.
- `callback` is a function with the signature `callback(erros)` where `error` is a json object. - `callback` is a function with the signature `callback(erros)` where `error` is a json object.

462
dist/tesseract.js vendored
Unescape View File

@ -1,78 +1,91 @@
(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){ (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){
exports.defaultOptions = { 'use strict';
langPath: 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/',
// workerPath: 'dist/worker.js', var defaultOptions = {
workerPath: 'https://cdn.rawgit.com/naptha/tesseract.js/0.1.3/dist/worker.js', workerPath: 'https://cdn.rawgit.com/naptha/tesseract.js/0.1.3/dist/worker.js',
tesseractPath: 'https://cdn.rawgit.com/naptha/tesseract.js-core/0.1.0/index.js', tesseractPath: 'https://cdn.rawgit.com/naptha/tesseract.js-core/0.1.0/index.js',
langPath: 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/'
};
if (location.hostname === '127.0.0.1' && location.port == '7355') {
console.debug('Using Development Configuration');
defaultOptions.workerPath = location.protocol + '//' + location.host + '/dist/worker.js';
} }
exports.spawnWorker = function spawnWorker(instance, workerOptions){ exports.defaultOptions = defaultOptions;
if(window.Blob && window.URL){
var blob = new Blob(['importScripts("' + workerOptions.workerPath + '");']) exports.spawnWorker = function spawnWorker(instance, workerOptions) {
if (window.Blob && window.URL) {
var blob = new Blob(['importScripts("' + workerOptions.workerPath + '");']);
var worker = new Worker(window.URL.createObjectURL(blob)); var worker = new Worker(window.URL.createObjectURL(blob));
}else{ } else {
var worker = new Worker(workerOptions.workerPath) var worker = new Worker(workerOptions.workerPath);
} }
worker.onmessage = function(e){ worker.onmessage = function (e) {
var packet = e.data; var packet = e.data;
instance._recv(packet) instance._recv(packet);
} };
return worker return worker;
} };
exports.terminateWorker = function(instance){ exports.terminateWorker = function (instance) {
instance.worker.terminate() instance.worker.terminate();
} };
exports.sendPacket = function sendPacket(instance, packet){ exports.sendPacket = function sendPacket(instance, packet) {
loadImage(packet.payload.image, function(img){ loadImage(packet.payload.image, function (img) {
packet.payload.image = img packet.payload.image = img;
instance.worker.postMessage(packet) instance.worker.postMessage(packet);
}) });
} };
function loadImage(image, cb) {
function loadImage(image, cb){ if (typeof image === 'string') {
if(typeof image === 'string'){ if (/^\#/.test(image)) {
if(/^\#/.test(image)){
// element css selector // element css selector
return loadImage(document.querySelector(image), cb) return loadImage(document.querySelector(image), cb);
}else{ } else {
// url or path // url or path
var im = new Image var im = new Image();
im.src = image; im.src = image;
im.onload = e => loadImage(im, cb); im.onload = function (e) {
return return loadImage(im, cb);
};
return;
} }
}else if(image instanceof File){ } else if (image instanceof File) {
// files // files
var fr = new FileReader() var fr = new FileReader();
fr.onload = e => loadImage(fr.result, cb); fr.onload = function (e) {
fr.readAsDataURL(image) return loadImage(fr.result, cb);
return };
}else if(image instanceof Blob){ fr.readAsDataURL(image);
return loadImage(URL.createObjectURL(image), cb) return;
}else if(image.getContext){ } else if (image instanceof Blob) {
return loadImage(URL.createObjectURL(image), cb);
} else if (image.getContext) {
// canvas element // canvas element
return loadImage(image.getContext('2d'), cb) return loadImage(image.getContext('2d'), cb);
}else if(image.tagName == "IMG" || image.tagName == "VIDEO"){ } else if (image.tagName == "IMG" || image.tagName == "VIDEO") {
// image element or video element // image element or video element
var c = document.createElement('canvas'); var c = document.createElement('canvas');
c.width = image.naturalWidth || image.videoWidth; c.width = image.naturalWidth || image.videoWidth;
c.height = image.naturalHeight || image.videoHeight; c.height = image.naturalHeight || image.videoHeight;
var ctx = c.getContext('2d'); var ctx = c.getContext('2d');
ctx.drawImage(image, 0, 0); ctx.drawImage(image, 0, 0);
return loadImage(ctx, cb) return loadImage(ctx, cb);
}else if(image.getImageData){ } else if (image.getImageData) {
// canvas context // canvas context
var data = image.getImageData(0, 0, image.canvas.width, image.canvas.height); var data = image.getImageData(0, 0, image.canvas.width, image.canvas.height);
return loadImage(data, cb) return loadImage(data, cb);
} }
cb(image) cb(image);
} }
},{}],2:[function(require,module,exports){ },{}],2:[function(require,module,exports){
"use strict";
// The result of dump.js is a big JSON tree // The result of dump.js is a big JSON tree
// which can be easily serialized (for instance // which can be easily serialized (for instance
// to be sent from a webworker to the main app // to be sent from a webworker to the main app
@ -80,212 +93,251 @@ function loadImage(image, cb){
// a (circular) DOM-like interface for walking // a (circular) DOM-like interface for walking
// through the data. // through the data.
module.exports = function circularize(page){ module.exports = function circularize(page) {
page.paragraphs = [] page.paragraphs = [];
page.lines = [] page.lines = [];
page.words = [] page.words = [];
page.symbols = [] page.symbols = [];
page.blocks.forEach(function(block){ page.blocks.forEach(function (block) {
block.page = page; block.page = page;
block.lines = [] block.lines = [];
block.words = [] block.words = [];
block.symbols = [] block.symbols = [];
block.paragraphs.forEach(function(para){ block.paragraphs.forEach(function (para) {
para.block = block; para.block = block;
para.page = page; para.page = page;
para.words = [] para.words = [];
para.symbols = [] para.symbols = [];
para.lines.forEach(function(line){ para.lines.forEach(function (line) {
line.paragraph = para; line.paragraph = para;
line.block = block; line.block = block;
line.page = page; line.page = page;
line.symbols = [] line.symbols = [];
line.words.forEach(function(word){ line.words.forEach(function (word) {
word.line = line; word.line = line;
word.paragraph = para; word.paragraph = para;
word.block = block; word.block = block;
word.page = page; word.page = page;
word.symbols.forEach(function(sym){ word.symbols.forEach(function (sym) {
sym.word = word; sym.word = word;
sym.line = line; sym.line = line;
sym.paragraph = para; sym.paragraph = para;
sym.block = block; sym.block = block;
sym.page = page; sym.page = page;
sym.line.symbols.push(sym) sym.line.symbols.push(sym);
sym.paragraph.symbols.push(sym) sym.paragraph.symbols.push(sym);
sym.block.symbols.push(sym) sym.block.symbols.push(sym);
sym.page.symbols.push(sym) sym.page.symbols.push(sym);
}) });
word.paragraph.words.push(word) word.paragraph.words.push(word);
word.block.words.push(word) word.block.words.push(word);
word.page.words.push(word) word.page.words.push(word);
}) });
line.block.lines.push(line) line.block.lines.push(line);
line.page.lines.push(line) line.page.lines.push(line);
}) });
para.page.paragraphs.push(para) para.page.paragraphs.push(para);
}) });
}) });
return page return page;
} };
},{}],3:[function(require,module,exports){ },{}],3:[function(require,module,exports){
"use strict"; "use strict";
var adapter = require('./node/index.js') 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; }; }();
var circularize = require('./common/circularize.js')
function _classCallCheck(instance, Constructor) { if (!(instance instanceof Constructor)) { throw new TypeError("Cannot call a class as a function"); } }
function createWorker(workerOptions){ var adapter = require('./node/index.js');
return new TesseractWorker(workerOptions) var circularize = require('./common/circularize.js');
function create(workerOptions) {
return new TesseractWorker(workerOptions);
} }
class TesseractWorker { var TesseractWorker = function () {
constructor(workerOptions){ function TesseractWorker(workerOptions) {
_classCallCheck(this, TesseractWorker);
this.worker = null; this.worker = null;
this.workerOptions = workerOptions; this.workerOptions = workerOptions;
this._currentJob = null; this._currentJob = null;
this._queue = [] this._queue = [];
}
recognize(image, options){
return this._delay(job => {
if(typeof options === 'string'){
options = { lang: options };
}else{
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(){ _createClass(TesseractWorker, [{
if(this.worker) adapter.terminateWorker(this); key: 'recognize',
this.worker = null; value: function recognize(image, options) {
} var _this = this;
_delay(fn){ return this._delay(function (job) {
if(!this.worker) this.worker = adapter.spawnWorker(this, this.workerOptions); if (typeof options === 'string') {
options = { lang: options };
var job = new TesseractJob(this); } else {
this._queue.push(e => { options = options || {};
this._queue.shift() options.lang = options.lang || 'eng';
this._currentJob = job; }
fn(job)
})
if(!this._currentJob) this._dequeue();
return job
}
_dequeue(){ job._send('recognize', { image: image, options: options, workerOptions: _this.workerOptions });
this._currentJob = null; });
if(this._queue.length > 0){
this._queue[0]()
} }
} }, {
key: 'detect',
_recv(packet){ 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(packet.status === 'resolve' && packet.action === 'recognize'){ if (packet.status === 'resolve' && packet.action === 'recognize') {
packet.data = circularize(packet.data); packet.data = circularize(packet.data);
} }
if(this._currentJob.id === packet.jobId){ if (this._currentJob.id === packet.jobId) {
this._currentJob._handle(packet) this._currentJob._handle(packet);
}else{ } else {
console.warn('Job ID ' + packet.jobId + ' not known.') console.warn('Job ID ' + packet.jobId + ' not known.');
}
} }
} }]);
}
return TesseractWorker;
}();
var jobCounter = 0; var jobCounter = 0;
class TesseractJob { var TesseractJob = function () {
constructor(instance){ function TesseractJob(instance) {
this.id = 'Job-' + (++jobCounter) + '-' + Math.random().toString(16).slice(3, 8) _classCallCheck(this, TesseractJob);
this.id = 'Job-' + ++jobCounter + '-' + Math.random().toString(16).slice(3, 8);
this._instance = instance; this._instance = instance;
this._resolve = [] this._resolve = [];
this._reject = [] this._reject = [];
this._progress = [] this._progress = [];
} }
then(resolve, reject){ _createClass(TesseractJob, [{
if(this._resolve.push){ key: 'then',
this._resolve.push(resolve) value: function then(resolve, reject) {
}else{ if (this._resolve.push) {
resolve(this._resolve) this._resolve.push(resolve);
} } else {
resolve(this._resolve);
}
if(reject) this.catch(reject); if (reject) this.catch(reject);
return this; return this;
}
catch(reject){
if(this._reject.push){
this._reject.push(reject)
}else{
reject(this._reject)
} }
return this; }, {
} key: 'catch',
progress(fn){ value: function _catch(reject) {
this._progress.push(fn) if (this._reject.push) {
return this; this._reject.push(reject);
} } else {
_send(action, payload){ reject(this._reject);
adapter.sendPacket(this._instance, { }
jobId: this.id, return this;
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)
} }
} }, {
} 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);
}
}
}]);
var DefaultTesseract = createWorker(adapter.defaultOptions) return TesseractJob;
DefaultTesseract.createWorker = createWorker; }();
module.exports = DefaultTesseract var DefaultTesseract = create(adapter.defaultOptions);
DefaultTesseract.create = create;
module.exports = DefaultTesseract;
},{"./common/circularize.js":2,"./node/index.js":1}]},{},[3])(3) },{"./common/circularize.js":2,"./node/index.js":1}]},{},[3])(3)
}); });

552
dist/worker.js vendored
Unescape View File

@ -82,7 +82,7 @@ AbstractChainedBatch.prototype.write = function (options, callback) {
module.exports = AbstractChainedBatch module.exports = AbstractChainedBatch
}).call(this,require('_process')) }).call(this,require('_process'))
},{"_process":36}],2:[function(require,module,exports){ },{"_process":37}],2:[function(require,module,exports){
(function (process){ (function (process){
/* Copyright (c) 2013 Rod Vagg, MIT License */ /* Copyright (c) 2013 Rod Vagg, MIT License */
@ -135,7 +135,7 @@ AbstractIterator.prototype.end = function (callback) {
module.exports = AbstractIterator module.exports = AbstractIterator
}).call(this,require('_process')) }).call(this,require('_process'))
},{"_process":36}],3:[function(require,module,exports){ },{"_process":37}],3:[function(require,module,exports){
(function (Buffer,process){ (function (Buffer,process){
/* Copyright (c) 2013 Rod Vagg, MIT License */ /* Copyright (c) 2013 Rod Vagg, MIT License */
@ -395,7 +395,7 @@ module.exports.AbstractIterator = AbstractIterator
module.exports.AbstractChainedBatch = AbstractChainedBatch module.exports.AbstractChainedBatch = AbstractChainedBatch
}).call(this,{"isBuffer":require("../is-buffer/index.js")},require('_process')) }).call(this,{"isBuffer":require("../is-buffer/index.js")},require('_process'))
},{"../is-buffer/index.js":10,"./abstract-chained-batch":1,"./abstract-iterator":2,"_process":36,"xtend":4}],4:[function(require,module,exports){ },{"../is-buffer/index.js":9,"./abstract-chained-batch":1,"./abstract-iterator":2,"_process":37,"xtend":4}],4:[function(require,module,exports){
module.exports = extend module.exports = extend
function extend() { function extend() {
@ -2323,7 +2323,7 @@ function isnan (val) {
} }
}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) }).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {})
},{"base64-js":5,"ieee754":8,"isarray":11}],7:[function(require,module,exports){ },{"base64-js":5,"ieee754":8,"isarray":10}],7:[function(require,module,exports){
/*global window:false, self:false, define:false, module:false */ /*global window:false, self:false, define:false, module:false */
/** /**
@ -3817,31 +3817,6 @@ exports.write = function (buffer, value, offset, isLE, mLen, nBytes) {
} }
},{}],9:[function(require,module,exports){ },{}],9:[function(require,module,exports){
if (typeof Object.create === 'function') {
// implementation from standard node.js 'util' module
module.exports = function inherits(ctor, superCtor) {
ctor.super_ = superCtor
ctor.prototype = Object.create(superCtor.prototype, {
constructor: {
value: ctor,
enumerable: false,
writable: true,
configurable: true
}
});
};
} else {
// old school shim for old browsers
module.exports = function inherits(ctor, superCtor) {
ctor.super_ = superCtor
var TempCtor = function () {}
TempCtor.prototype = superCtor.prototype
ctor.prototype = new TempCtor()
ctor.prototype.constructor = ctor
}
}
},{}],10:[function(require,module,exports){
/*! /*!
* Determine if an object is a Buffer * Determine if an object is a Buffer
* *
@ -3864,14 +3839,14 @@ function isSlowBuffer (obj) {
return typeof obj.readFloatLE === 'function' && typeof obj.slice === 'function' && isBuffer(obj.slice(0, 0)) return typeof obj.readFloatLE === 'function' && typeof obj.slice === 'function' && isBuffer(obj.slice(0, 0))
} }
},{}],11:[function(require,module,exports){ },{}],10:[function(require,module,exports){
var toString = {}.toString; var toString = {}.toString;
module.exports = Array.isArray || function (arr) { module.exports = Array.isArray || function (arr) {
return toString.call(arr) == '[object Array]'; return toString.call(arr) == '[object Array]';
}; };
},{}],12:[function(require,module,exports){ },{}],11:[function(require,module,exports){
var Buffer = require('buffer').Buffer; var Buffer = require('buffer').Buffer;
module.exports = isBuffer; module.exports = isBuffer;
@ -3881,7 +3856,7 @@ function isBuffer (o) {
|| /\[object (.+Array|Array.+)\]/.test(Object.prototype.toString.call(o)); || /\[object (.+Array|Array.+)\]/.test(Object.prototype.toString.call(o));
} }
},{"buffer":6}],13:[function(require,module,exports){ },{"buffer":6}],12:[function(require,module,exports){
(function (Buffer){ (function (Buffer){
module.exports = Level module.exports = Level
@ -4059,7 +4034,7 @@ var checkKeyValue = Level.prototype._checkKeyValue = function (obj, type) {
} }
}).call(this,require("buffer").Buffer) }).call(this,require("buffer").Buffer)
},{"./iterator":14,"abstract-leveldown":3,"buffer":6,"idb-wrapper":7,"isbuffer":12,"typedarray-to-buffer":37,"util":39,"xtend":41}],14:[function(require,module,exports){ },{"./iterator":13,"abstract-leveldown":3,"buffer":6,"idb-wrapper":7,"isbuffer":11,"typedarray-to-buffer":38,"util":41,"xtend":15}],13:[function(require,module,exports){
var util = require('util') var util = require('util')
var AbstractIterator = require('abstract-leveldown').AbstractIterator var AbstractIterator = require('abstract-leveldown').AbstractIterator
var ltgt = require('ltgt') var ltgt = require('ltgt')
@ -4133,7 +4108,43 @@ Iterator.prototype._next = function (callback) {
this.callback = callback this.callback = callback
} }
},{"abstract-leveldown":3,"ltgt":15,"util":39}],15:[function(require,module,exports){ },{"abstract-leveldown":3,"ltgt":16,"util":41}],14:[function(require,module,exports){
module.exports = hasKeys
function hasKeys(source) {
return source !== null &&
(typeof source === "object" ||
typeof source === "function")
}
},{}],15:[function(require,module,exports){
var Keys = require("object-keys")
var hasKeys = require("./has-keys")
module.exports = extend
function extend() {
var target = {}
for (var i = 0; i < arguments.length; i++) {
var source = arguments[i]
if (!hasKeys(source)) {
continue
}
var keys = Keys(source)
for (var j = 0; j < keys.length; j++) {
var name = keys[j]
target[name] = source[name]
}
}
return target
}
},{"./has-keys":14,"object-keys":18}],16:[function(require,module,exports){
(function (Buffer){ (function (Buffer){
exports.compare = function (a, b) { exports.compare = function (a, b) {
@ -4283,7 +4294,7 @@ exports.filter = function (range, compare) {
} }
}).call(this,{"isBuffer":require("../is-buffer/index.js")}) }).call(this,{"isBuffer":require("../is-buffer/index.js")})
},{"../is-buffer/index.js":10}],16:[function(require,module,exports){ },{"../is-buffer/index.js":9}],17:[function(require,module,exports){
var hasOwn = Object.prototype.hasOwnProperty; var hasOwn = Object.prototype.hasOwnProperty;
var toString = Object.prototype.toString; var toString = Object.prototype.toString;
@ -4325,11 +4336,11 @@ module.exports = function forEach(obj, fn) {
}; };
},{}],17:[function(require,module,exports){ },{}],18:[function(require,module,exports){
module.exports = Object.keys || require('./shim'); module.exports = Object.keys || require('./shim');
},{"./shim":19}],18:[function(require,module,exports){ },{"./shim":20}],19:[function(require,module,exports){
var toString = Object.prototype.toString; var toString = Object.prototype.toString;
module.exports = function isArguments(value) { module.exports = function isArguments(value) {
@ -4347,7 +4358,7 @@ module.exports = function isArguments(value) {
}; };
},{}],19:[function(require,module,exports){ },{}],20:[function(require,module,exports){
(function () { (function () {
"use strict"; "use strict";
@ -4411,7 +4422,7 @@ module.exports = function isArguments(value) {
}()); }());
},{"./foreach":16,"./isArguments":18}],20:[function(require,module,exports){ },{"./foreach":17,"./isArguments":19}],21:[function(require,module,exports){
// Top level file is just a mixin of submodules & constants // Top level file is just a mixin of submodules & constants
'use strict'; 'use strict';
@ -4427,7 +4438,7 @@ assign(pako, deflate, inflate, constants);
module.exports = pako; module.exports = pako;
},{"./lib/deflate":21,"./lib/inflate":22,"./lib/utils/common":23,"./lib/zlib/constants":26}],21:[function(require,module,exports){ },{"./lib/deflate":22,"./lib/inflate":23,"./lib/utils/common":24,"./lib/zlib/constants":27}],22:[function(require,module,exports){
'use strict'; 'use strict';
@ -4829,7 +4840,7 @@ exports.deflate = deflate;
exports.deflateRaw = deflateRaw; exports.deflateRaw = deflateRaw;
exports.gzip = gzip; exports.gzip = gzip;
},{"./utils/common":23,"./utils/strings":24,"./zlib/deflate":28,"./zlib/messages":33,"./zlib/zstream":35}],22:[function(require,module,exports){ },{"./utils/common":24,"./utils/strings":25,"./zlib/deflate":29,"./zlib/messages":34,"./zlib/zstream":36}],23:[function(require,module,exports){
'use strict'; 'use strict';
@ -5249,7 +5260,7 @@ exports.inflate = inflate;
exports.inflateRaw = inflateRaw; exports.inflateRaw = inflateRaw;
exports.ungzip = inflate; exports.ungzip = inflate;
},{"./utils/common":23,"./utils/strings":24,"./zlib/constants":26,"./zlib/gzheader":29,"./zlib/inflate":31,"./zlib/messages":33,"./zlib/zstream":35}],23:[function(require,module,exports){ },{"./utils/common":24,"./utils/strings":25,"./zlib/constants":27,"./zlib/gzheader":30,"./zlib/inflate":32,"./zlib/messages":34,"./zlib/zstream":36}],24:[function(require,module,exports){
'use strict'; 'use strict';
@ -5353,7 +5364,7 @@ exports.setTyped = function (on) {
exports.setTyped(TYPED_OK); exports.setTyped(TYPED_OK);
},{}],24:[function(require,module,exports){ },{}],25:[function(require,module,exports){
// String encode/decode helpers // String encode/decode helpers
'use strict'; 'use strict';
@ -5540,7 +5551,7 @@ exports.utf8border = function (buf, max) {
return (pos + _utf8len[buf[pos]] > max) ? pos : max; return (pos + _utf8len[buf[pos]] > max) ? pos : max;
}; };
},{"./common":23}],25:[function(require,module,exports){ },{"./common":24}],26:[function(require,module,exports){
'use strict'; 'use strict';
// Note: adler32 takes 12% for level 0 and 2% for level 6. // Note: adler32 takes 12% for level 0 and 2% for level 6.
@ -5574,7 +5585,7 @@ function adler32(adler, buf, len, pos) {
module.exports = adler32; module.exports = adler32;
},{}],26:[function(require,module,exports){ },{}],27:[function(require,module,exports){
'use strict'; 'use strict';
@ -5626,7 +5637,7 @@ module.exports = {
//Z_NULL: null // Use -1 or null inline, depending on var type //Z_NULL: null // Use -1 or null inline, depending on var type
}; };
},{}],27:[function(require,module,exports){ },{}],28:[function(require,module,exports){
'use strict'; 'use strict';
// Note: we can't get significant speed boost here. // Note: we can't get significant speed boost here.
@ -5669,7 +5680,7 @@ function crc32(crc, buf, len, pos) {
module.exports = crc32; module.exports = crc32;
},{}],28:[function(require,module,exports){ },{}],29:[function(require,module,exports){
'use strict'; 'use strict';
var utils = require('../utils/common'); var utils = require('../utils/common');
@ -7526,7 +7537,7 @@ exports.deflatePrime = deflatePrime;
exports.deflateTune = deflateTune; exports.deflateTune = deflateTune;
*/ */
},{"../utils/common":23,"./adler32":25,"./crc32":27,"./messages":33,"./trees":34}],29:[function(require,module,exports){ },{"../utils/common":24,"./adler32":26,"./crc32":28,"./messages":34,"./trees":35}],30:[function(require,module,exports){
'use strict'; 'use strict';
@ -7568,7 +7579,7 @@ function GZheader() {
module.exports = GZheader; module.exports = GZheader;
},{}],30:[function(require,module,exports){ },{}],31:[function(require,module,exports){
'use strict'; 'use strict';
// See state defs from inflate.js // See state defs from inflate.js
@ -7896,7 +7907,7 @@ module.exports = function inflate_fast(strm, start) {
return; return;
}; };
},{}],31:[function(require,module,exports){ },{}],32:[function(require,module,exports){
'use strict'; 'use strict';
@ -9436,7 +9447,7 @@ exports.inflateSyncPoint = inflateSyncPoint;
exports.inflateUndermine = inflateUndermine; exports.inflateUndermine = inflateUndermine;
*/ */
},{"../utils/common":23,"./adler32":25,"./crc32":27,"./inffast":30,"./inftrees":32}],32:[function(require,module,exports){ },{"../utils/common":24,"./adler32":26,"./crc32":28,"./inffast":31,"./inftrees":33}],33:[function(require,module,exports){
'use strict'; 'use strict';
@ -9765,7 +9776,7 @@ module.exports = function inflate_table(type, lens, lens_index, codes, table, ta
return 0; return 0;
}; };
},{"../utils/common":23}],33:[function(require,module,exports){ },{"../utils/common":24}],34:[function(require,module,exports){
'use strict'; 'use strict';
module.exports = { module.exports = {
@ -9780,7 +9791,7 @@ module.exports = {
'-6': 'incompatible version' /* Z_VERSION_ERROR (-6) */ '-6': 'incompatible version' /* Z_VERSION_ERROR (-6) */
}; };
},{}],34:[function(require,module,exports){ },{}],35:[function(require,module,exports){
'use strict'; 'use strict';
@ -10984,7 +10995,7 @@ exports._tr_flush_block = _tr_flush_block;
exports._tr_tally = _tr_tally; exports._tr_tally = _tr_tally;
exports._tr_align = _tr_align; exports._tr_align = _tr_align;
},{"../utils/common":23}],35:[function(require,module,exports){ },{"../utils/common":24}],36:[function(require,module,exports){
'use strict'; 'use strict';
@ -11015,7 +11026,7 @@ function ZStream() {
module.exports = ZStream; module.exports = ZStream;
},{}],36:[function(require,module,exports){ },{}],37:[function(require,module,exports){
// shim for using process in browser // shim for using process in browser
var process = module.exports = {}; var process = module.exports = {};
@ -11197,7 +11208,7 @@ process.chdir = function (dir) {
}; };
process.umask = function() { return 0; }; process.umask = function() { return 0; };
},{}],37:[function(require,module,exports){ },{}],38:[function(require,module,exports){
(function (Buffer){ (function (Buffer){
/** /**
* Convert a typed array to a Buffer without a copy * Convert a typed array to a Buffer without a copy
@ -11220,14 +11231,39 @@ module.exports = function (arr) {
} }
}).call(this,require("buffer").Buffer) }).call(this,require("buffer").Buffer)
},{"buffer":6}],38:[function(require,module,exports){ },{"buffer":6}],39:[function(require,module,exports){
if (typeof Object.create === 'function') {
// implementation from standard node.js 'util' module
module.exports = function inherits(ctor, superCtor) {
ctor.super_ = superCtor
ctor.prototype = Object.create(superCtor.prototype, {
constructor: {
value: ctor,
enumerable: false,
writable: true,
configurable: true
}
});
};
} else {
// old school shim for old browsers
module.exports = function inherits(ctor, superCtor) {
ctor.super_ = superCtor
var TempCtor = function () {}
TempCtor.prototype = superCtor.prototype
ctor.prototype = new TempCtor()
ctor.prototype.constructor = ctor
}
}
},{}],40:[function(require,module,exports){
module.exports = function isBuffer(arg) { module.exports = function isBuffer(arg) {
return arg && typeof arg === 'object' return arg && typeof arg === 'object'
&& typeof arg.copy === 'function' && typeof arg.copy === 'function'
&& typeof arg.fill === 'function' && typeof arg.fill === 'function'
&& typeof arg.readUInt8 === 'function'; && typeof arg.readUInt8 === 'function';
} }
},{}],39:[function(require,module,exports){ },{}],41:[function(require,module,exports){
(function (process,global){ (function (process,global){
// Copyright Joyent, Inc. and other Node contributors. // Copyright Joyent, Inc. and other Node contributors.
// //
@ -11817,185 +11853,170 @@ function hasOwnProperty(obj, prop) {
} }
}).call(this,require('_process'),typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) }).call(this,require('_process'),typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {})
},{"./support/isBuffer":38,"_process":36,"inherits":9}],40:[function(require,module,exports){ },{"./support/isBuffer":40,"_process":37,"inherits":39}],42:[function(require,module,exports){
module.exports = hasKeys 'use strict';
function hasKeys(source) {
return source !== null &&
(typeof source === "object" ||
typeof source === "function")
}
},{}],41:[function(require,module,exports){
var Keys = require("object-keys")
var hasKeys = require("./has-keys")
module.exports = extend
function extend() {
var target = {}
for (var i = 0; i < arguments.length; i++) {
var source = arguments[i]
if (!hasKeys(source)) {
continue
}
var keys = Keys(source)
for (var j = 0; j < keys.length; j++) {
var name = keys[j]
target[name] = source[name]
}
}
return target
}
},{"./has-keys":40,"object-keys":17}],42:[function(require,module,exports){ var leveljs = require('level-js');
var leveljs = require('level-js') var db = typeof indexedDB === 'undefined' ? { open: function open(_, cb) {
var db = typeof indexedDB === 'undefined' ? { open: (_, cb) => cb(true) } : leveljs('./tessdata2') return cb(true);
} } : leveljs('./tessdata2');
var langdata = require('../common/langdata.json') var langdata = require('../common/langdata.json');
module.exports = function getLanguageData(req, res, cb){ module.exports = function getLanguageData(req, res, cb) {
var lang = req.options.lang; var lang = req.options.lang;
function saveDataFile(data){ function saveDataFile(data) {
db.put(lang, data, err => console.log('cached', lang, err)) db.put(lang, data, function (err) {
cb(data) return console.log('cached', lang, err);
});
cb(data);
} }
db.open({ compression: false }, err => { db.open({ compression: false }, function (err) {
if (err) return fetchLanguageData(req, res, cb); if (err) return fetchLanguageData(req, res, cb);
db.get(lang, (err, data) => { db.get(lang, function (err, data) {
if (err) return fetchLanguageData(req, res, saveDataFile); if (err) return fetchLanguageData(req, res, saveDataFile);
res.progress({ status: 'found in cache ' + lang + '.traineddata' }) res.progress({ status: 'found in cache ' + lang + '.traineddata' });
cb(data) cb(data);
}) });
}) });
} };
var ungzip = require('pako').ungzip; var ungzip = require('pako').ungzip;
function fetchLanguageData(req, res, cb){ function fetchLanguageData(req, res, cb) {
var lang = req.options.lang; var lang = req.options.lang;
var langfile = lang + '.traineddata.gz'; var langfile = lang + '.traineddata.gz';
var url = req.workerOptions.langPath + langfile; var url = req.workerOptions.langPath + langfile;
var xhr = new XMLHttpRequest(); var xhr = new XMLHttpRequest();
xhr.responseType = 'arraybuffer';
xhr.open('GET', url, true); xhr.open('GET', url, true);
xhr.onerror = e => { xhr.responseType = 'arraybuffer';
xhr.onprogress = xhr.onload = null xhr.onerror = function (e) {
cb(xhr, null) xhr.onprogress = xhr.onload = null;
} cb(xhr, null);
xhr.onprogress = e => };
res.progress({ xhr.onprogress = function (e) {
return res.progress({
status: 'downloading ' + langfile, status: 'downloading ' + langfile,
loaded: e.loaded, loaded: e.loaded,
progress: Math.min(1, e.loaded / langdata[lang]) progress: Math.min(1, e.loaded / langdata[lang])
}); });
};
xhr.onload = e => { xhr.onload = function (e) {
if (!(xhr.status == 200 || (xhr.status == 0 && xhr.response))) return res.reject('Error downloading language ' + url); if (!(xhr.status == 200 || xhr.status == 0 && xhr.response)) return res.reject('Error downloading language ' + url);
res.progress({ status: 'unzipping ' + langfile }) res.progress({ status: 'unzipping ' + langfile });
// in case the gzips are already ungzipped or extra gzipped // in case the gzips are already ungzipped or extra gzipped
var response = new Uint8Array(xhr.response) var response = new Uint8Array(xhr.response);
try { try {
while(response[0] == 0x1f && response[1] == 0x8b) response = ungzip(response); while (response[0] == 0x1f && response[1] == 0x8b) {
response = ungzip(response);
}
} catch (err) { } catch (err) {
return res.reject('Error unzipping language file ' + langfile + '\n' + err.message) return res.reject('Error unzipping language file ' + langfile + '\n' + err.message);
} }
cb(response) cb(response);
} };
xhr.send() xhr.send();
} }
},{"../common/langdata.json":46,"level-js":13,"pako":20}],43:[function(require,module,exports){ },{"../common/langdata.json":46,"level-js":12,"pako":21}],43:[function(require,module,exports){
(function (global){ (function (global){
"use strict"; "use strict";
var workerUtils = require('../common/worker.js') var workerUtils = require('../common/worker.js');
global.addEventListener('message', function(e){ global.addEventListener('message', function (e) {
var packet = e.data; var packet = e.data;
workerUtils.dispatchHandlers(packet, obj => postMessage(obj)) workerUtils.dispatchHandlers(packet, function (obj) {
}) return postMessage(obj);
});
});
exports.getCore = function(req, res){ exports.getCore = function (req, res) {
if(!global.TesseractCore){ if (!global.TesseractCore) {
res.progress({ status: 'loading tesseract core' }) res.progress({ status: 'loading tesseract core' });
importScripts(req.workerOptions.tesseractPath) importScripts(req.workerOptions.tesseractPath);
res.progress({ status: 'loaded tesseract core' }) res.progress({ status: 'loaded tesseract core' });
} }
return TesseractCore return TesseractCore;
} };
exports.getLanguageData = require('./lang.js') exports.getLanguageData = require('./lang.js');
workerUtils.setAdapter(module.exports); workerUtils.setAdapter(module.exports);
}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) }).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {})
},{"../common/worker.js":47,"./lang.js":42}],44:[function(require,module,exports){ },{"../common/worker.js":47,"./lang.js":42}],44:[function(require,module,exports){
'use strict';
// This converts an image to grayscale // This converts an image to grayscale
module.exports = function desaturate(image){ module.exports = function desaturate(image) {
var width, height; var width, height;
if(image.data){ if (image.data) {
var src = image.data; var src = image.data;
width = image.width, width = image.width, height = image.height;
height = image.height; var dst = new Uint8Array(width * height);
var dst = new Uint8Array(width * height); var srcLength = src.length | 0,
var srcLength = src.length | 0, srcLength_16 = (srcLength - 16) | 0; srcLength_16 = srcLength - 16 | 0;
for (var i = 0, j = 0; i <= srcLength_16; i += 16, j += 4) { 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 // 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] = (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 + 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 + 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 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 for (; i < srcLength; i += 4, ++j) {
dst[j] = (((src[i] * 77 + src[i+1] * 151 + src[i+2] * 28) * src[i+3]) + ((255-src[i+3]) << 15) + 32768) >> 16 //finish up
image = dst; dst[j] = (src[i] * 77 + src[i + 1] * 151 + src[i + 2] * 28) * src[i + 3] + (255 - src[i + 3] << 15) + 32768 >> 16;
} else { throw 'Invalid ImageData' } }image = dst;
return image } else {
} throw 'Invalid ImageData';
}
return image;
};
},{}],45:[function(require,module,exports){ },{}],45:[function(require,module,exports){
module.exports = function DumpLiterallyEverything(Module, base){ 'use strict';
module.exports = function DumpLiterallyEverything(Module, base) {
var ri = base.GetIterator(); var ri = base.GetIterator();
var blocks = []; var blocks = [];
var block, para, textline, word, symbol; var block, para, textline, word, symbol;
function enumToString(value, prefix){ function enumToString(value, prefix) {
return (Object.keys(Module) return Object.keys(Module).filter(function (e) {
.filter(function(e){ return e.substr(0, prefix.length + 1) == prefix + '_' }) return e.substr(0, prefix.length + 1) == prefix + '_';
.filter(function(e){ return Module[e] === value }) }).filter(function (e) {
.map(function(e){ return e.slice(prefix.length + 1) })[0]) return Module[e] === value;
}).map(function (e) {
return e.slice(prefix.length + 1);
})[0];
} }
ri.Begin() ri.Begin();
do { do {
if(ri.IsAtBeginningOf(Module.RIL_BLOCK)){ if (ri.IsAtBeginningOf(Module.RIL_BLOCK)) {
var poly = ri.BlockPolygon(); var poly = ri.BlockPolygon();
var polygon = null; var polygon = null;
// BlockPolygon() returns null when automatic page segmentation is off // BlockPolygon() returns null when automatic page segmentation is off
if(Module.getPointer(poly) > 0){ if (Module.getPointer(poly) > 0) {
var n = poly.get_n(), var n = poly.get_n(),
px = poly.get_x(), px = poly.get_x(),
py = poly.get_y(), py = poly.get_y(),
polygon = []; polygon = [];
for(var i = 0; i < n; i++){ for (var i = 0; i < n; i++) {
polygon.push([px.getValue(i), py.getValue(i)]); polygon.push([px.getValue(i), py.getValue(i)]);
} }
Module._ptaDestroy(Module.getPointer(poly)); Module._ptaDestroy(Module.getPointer(poly));
} }
block = { block = {
paragraphs: [], paragraphs: [],
@ -12006,10 +12027,10 @@ module.exports = function DumpLiterallyEverything(Module, base){
blocktype: enumToString(ri.BlockType(), 'PT'), blocktype: enumToString(ri.BlockType(), 'PT'),
polygon: polygon polygon: polygon
} };
blocks.push(block) blocks.push(block);
} }
if(ri.IsAtBeginningOf(Module.RIL_PARA)){ if (ri.IsAtBeginningOf(Module.RIL_PARA)) {
para = { para = {
lines: [], lines: [],
@ -12019,10 +12040,10 @@ module.exports = function DumpLiterallyEverything(Module, base){
bbox: ri.getBoundingBox(Module.RIL_PARA), bbox: ri.getBoundingBox(Module.RIL_PARA),
is_ltr: !!ri.ParagraphIsLtr() is_ltr: !!ri.ParagraphIsLtr()
} };
block.paragraphs.push(para) block.paragraphs.push(para);
} }
if(ri.IsAtBeginningOf(Module.RIL_TEXTLINE)){ if (ri.IsAtBeginningOf(Module.RIL_TEXTLINE)) {
textline = { textline = {
words: [], words: [],
@ -12030,10 +12051,10 @@ module.exports = function DumpLiterallyEverything(Module, base){
confidence: ri.Confidence(Module.RIL_TEXTLINE), confidence: ri.Confidence(Module.RIL_TEXTLINE),
baseline: ri.getBaseline(Module.RIL_TEXTLINE), baseline: ri.getBaseline(Module.RIL_TEXTLINE),
bbox: ri.getBoundingBox(Module.RIL_TEXTLINE) bbox: ri.getBoundingBox(Module.RIL_TEXTLINE)
} };
para.lines.push(textline) para.lines.push(textline);
} }
if(ri.IsAtBeginningOf(Module.RIL_WORD)){ if (ri.IsAtBeginningOf(Module.RIL_WORD)) {
var fontInfo = ri.getWordFontAttributes(), var fontInfo = ri.getWordFontAttributes(),
wordDir = ri.WordDirection(); wordDir = ri.WordDirection();
word = { word = {
@ -12058,25 +12079,25 @@ module.exports = function DumpLiterallyEverything(Module, base){
is_smallcaps: fontInfo.is_smallcaps, is_smallcaps: fontInfo.is_smallcaps,
font_size: fontInfo.pointsize, font_size: fontInfo.pointsize,
font_id: fontInfo.font_id, font_id: fontInfo.font_id,
font_name: fontInfo.font_name, font_name: fontInfo.font_name
} };
var wc = new Module.WordChoiceIterator(ri); var wc = new Module.WordChoiceIterator(ri);
do { do {
word.choices.push({ word.choices.push({
text: wc.GetUTF8Text(), text: wc.GetUTF8Text(),
confidence: wc.Confidence() confidence: wc.Confidence()
}) });
} while (wc.Next()); } while (wc.Next());
Module.destroy(wc) Module.destroy(wc);
textline.words.push(word) textline.words.push(word);
} }
var image = null; var image = null;
// var pix = ri.GetBinaryImage(Module.RIL_SYMBOL) // var pix = ri.GetBinaryImage(Module.RIL_SYMBOL)
// var image = pix2array(pix); // var image = pix2array(pix);
// // for some reason it seems that things stop working if you destroy pics // // for some reason it seems that things stop working if you destroy pics
// Module._pixDestroy(Module.getPointer(pix)); // Module._pixDestroy(Module.getPointer(pix));
if(ri.IsAtBeginningOf(Module.RIL_SYMBOL)){ if (ri.IsAtBeginningOf(Module.RIL_SYMBOL)) {
symbol = { symbol = {
choices: [], choices: [],
image: image, image: image,
@ -12088,20 +12109,20 @@ module.exports = function DumpLiterallyEverything(Module, base){
is_superscript: !!ri.SymbolIsSuperscript(), is_superscript: !!ri.SymbolIsSuperscript(),
is_subscript: !!ri.SymbolIsSubscript(), is_subscript: !!ri.SymbolIsSubscript(),
is_dropcap: !!ri.SymbolIsDropcap(), is_dropcap: !!ri.SymbolIsDropcap()
} };
word.symbols.push(symbol) word.symbols.push(symbol);
var ci = new Module.ChoiceIterator(ri); var ci = new Module.ChoiceIterator(ri);
do { do {
symbol.choices.push({ symbol.choices.push({
text: ci.GetUTF8Text(), text: ci.GetUTF8Text(),
confidence: ci.Confidence() confidence: ci.Confidence()
}) });
} while (ci.Next()); } while (ci.Next());
Module.destroy(ci) Module.destroy(ci);
} }
} while (ri.Next(Module.RIL_SYMBOL)); } while (ri.Next(Module.RIL_SYMBOL));
Module.destroy(ri) Module.destroy(ri);
return { return {
text: base.GetUTF8Text(), text: base.GetUTF8Text(),
@ -12113,120 +12134,116 @@ module.exports = function DumpLiterallyEverything(Module, base){
psm: enumToString(base.GetPageSegMode(), 'PSM'), psm: enumToString(base.GetPageSegMode(), 'PSM'),
oem: enumToString(base.oem(), 'OEM'), oem: enumToString(base.oem(), 'OEM'),
version: base.Version(), version: base.Version()
} };
} };
// the generated HOCR is excessively indented, so // the generated HOCR is excessively indented, so
// we get rid of that indentation // we get rid of that indentation
function deindent(html){ function deindent(html) {
var lines = html.split('\n') var lines = html.split('\n');
if(lines[0].substring(0, 2) === " "){ if (lines[0].substring(0, 2) === " ") {
for (var i = 0; i < lines.length; i++) { for (var i = 0; i < lines.length; i++) {
if (lines[i].substring(0,2) === " ") { if (lines[i].substring(0, 2) === " ") {
lines[i] = lines[i].slice(2) lines[i] = lines[i].slice(2);
} }
}; };
} }
return lines.join('\n') return lines.join('\n');
} }
},{}],46:[function(require,module,exports){ },{}],46:[function(require,module,exports){
module.exports={"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} module.exports={"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}
},{}],47:[function(require,module,exports){ },{}],47:[function(require,module,exports){
'use strict';
var latestJob; var latestJob;
var Module; var Module;
var base; var base;
var adapter = {}; var adapter = {};
function dispatchHandlers(packet, send){ function dispatchHandlers(packet, send) {
function respond(status, data){ function respond(status, data) {
send({ send({
jobId: packet.jobId, jobId: packet.jobId,
status: status, status: status,
action: packet.action, action: packet.action,
data: data data: data
}) });
} }
respond.resolve = respond.bind(this, 'resolve') respond.resolve = respond.bind(this, 'resolve');
respond.reject = respond.bind(this, 'reject') respond.reject = respond.bind(this, 'reject');
respond.progress = respond.bind(this, 'progress') respond.progress = respond.bind(this, 'progress');
latestJob = respond; latestJob = respond;
if(packet.action === 'recognize'){ if (packet.action === 'recognize') {
handleRecognize(packet.payload, respond) handleRecognize(packet.payload, respond);
}else if(packet.action === 'detect'){ } else if (packet.action === 'detect') {
handleDetect(packet.payload, respond) handleDetect(packet.payload, respond);
} }
} }
exports.dispatchHandlers = dispatchHandlers; exports.dispatchHandlers = dispatchHandlers;
exports.setAdapter = function setAdapter(impl){ exports.setAdapter = function setAdapter(impl) {
adapter = impl; adapter = impl;
} };
function handleInit(req, res){ function handleInit(req, res) {
if(!Module){ if (!Module) {
var Core = adapter.getCore(req, res); var Core = adapter.getCore(req, res);
res.progress({ status: 'initializing tesseract api' }) res.progress({ status: 'initializing tesseract api' });
Module = Core({ Module = Core({
TOTAL_MEMORY: req.memory, TOTAL_MEMORY: req.memory,
TesseractProgress(percent){ TesseractProgress: function TesseractProgress(percent) {
latestJob.progress({ status: 'recognizing text', progress: Math.max(0, (percent-30)/70) }) latestJob.progress({ status: 'recognizing text', progress: Math.max(0, (percent - 30) / 70) });
}, },
onRuntimeInitialized() {} onRuntimeInitialized: function onRuntimeInitialized() {}
}) });
Module.FS_createPath("/", "tessdata", true, true) Module.FS_createPath("/", "tessdata", true, true);
base = new Module.TessBaseAPI() base = new Module.TessBaseAPI();
res.progress({ status: 'initialized tesseract api' }) res.progress({ status: 'initialized tesseract api' });
} }
} }
var dump = require('./dump.js');
var desaturate = require('./desaturate.js');
function setImage(Module, base, image) {
var dump = require('./dump.js')
var desaturate = require('./desaturate.js')
function setImage(Module, base, image){
var imgbin = desaturate(image), var imgbin = desaturate(image),
width = image.width, width = image.width,
height = image.height; height = image.height;
var ptr = Module.allocate(imgbin, 'i8', Module.ALLOC_NORMAL); var ptr = Module.allocate(imgbin, 'i8', Module.ALLOC_NORMAL);
base.SetImage(Module.wrapPointer(ptr), width, height, 1, width); base.SetImage(Module.wrapPointer(ptr), width, height, 1, width);
base.SetRectangle(0, 0, width, height) base.SetRectangle(0, 0, width, height);
return ptr; return ptr;
} }
function loadLanguage(req, res, cb){ function loadLanguage(req, res, cb) {
var lang = req.options.lang; var lang = req.options.lang;
if(!Module._loadedLanguages) Module._loadedLanguages = {};
if(lang in Module._loadedLanguages) return cb();
adapter.getLanguageData(req, res, function(data){ 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); Module.FS_createDataFile('tessdata', lang + ".traineddata", data, true, false);
res.progress({ status: 'loaded ' + lang + '.traineddata' }) res.progress({ status: 'loaded ' + lang + '.traineddata' });
Module._loadedLanguages[lang] = true; Module._loadedLanguages[lang] = true;
cb() cb();
}) });
} }
function handleRecognize(req, res) {
handleInit(req, res);
loadLanguage(req, res, function () {
function handleRecognize(req, res){
handleInit(req, res)
loadLanguage(req, res, function(){
var lang = req.options.lang; var lang = req.options.lang;
base.Init(null, lang) base.Init(null, lang);
res.progress({ status: 'initialized with language' }) res.progress({ status: 'initialized with language' });
var options = req.options; var options = req.options;
for (var option in options) { for (var option in options) {
@ -12236,38 +12253,37 @@ function handleRecognize(req, res){
} }
var ptr = setImage(Module, base, req.image); var ptr = setImage(Module, base, req.image);
base.Recognize(null) base.Recognize(null);
var result = dump(Module, base) var result = dump(Module, base);
base.End(); base.End();
Module._free(ptr); Module._free(ptr);
res.resolve(result); res.resolve(result);
}) });
} }
function handleDetect(req, res) {
function handleDetect(req, res){ handleInit(req, res);
handleInit(req, res)
req.options.lang = 'osd'; req.options.lang = 'osd';
loadLanguage(req, res, function(){ loadLanguage(req, res, function () {
base.Init(null, 'osd');
base.SetPageSegMode(Module.PSM_OSD_ONLY);
base.Init(null, 'osd')
base.SetPageSegMode(Module.PSM_OSD_ONLY)
var ptr = setImage(Module, base, req.image); var ptr = setImage(Module, base, req.image);
var results = new Module.OSResults(); var results = new Module.OSResults();
var success = base.DetectOS(results); var success = base.DetectOS(results);
if(!success){ if (!success) {
base.End(); base.End();
Module._free(ptr); Module._free(ptr);
res.reject("failed to detect os") res.reject("failed to detect os");
} else { } else {
var charset = results.get_unicharset() var charset = results.get_unicharset();
var best = results.get_best_result() var best = results.get_best_result();
var oid = best.get_orientation_id(), var oid = best.get_orientation_id(),
sid = best.get_script_id(); sid = best.get_script_id();
@ -12277,14 +12293,14 @@ function handleDetect(req, res){
script_confidence: best.get_sconfidence(), script_confidence: best.get_sconfidence(),
orientation_degrees: [0, 270, 180, 90][oid], orientation_degrees: [0, 270, 180, 90][oid],
orientation_confidence: best.get_oconfidence() orientation_confidence: best.get_oconfidence()
} };
base.End(); base.End();
Module._free(ptr); Module._free(ptr);
res.resolve(result) res.resolve(result);
} }
}) });
} }
},{"./desaturate.js":44,"./dump.js":45}]},{},[43]); },{"./desaturate.js":44,"./dump.js":45}]},{},[43]);

2
examples/file-input/demo.html
Unescape View File

@ -1,2 +1,2 @@
<script src="dist/tesseract.js"></script> <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]).progress(console.log)">

2
package.json
Unescape View File

@ -5,7 +5,7 @@
"main": "src/index.js", "main": "src/index.js",
"scripts": { "scripts": {
"test": "echo \"Error: no test specified\" & exit 1", "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", "start": "watchify src/index.js -t [ babelify --presets [ es2015 ] ] -o dist/tesseract.js --standalone Tesseract & watchify src/browser/worker.js -t [ babelify --presets [ es2015 ] ] -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" "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"
}, },
"browser": { "browser": {

15
src/browser/index.js
Unescape View File

@ -1,10 +1,19 @@
exports.defaultOptions = { var defaultOptions = {
langPath: 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/',
// workerPath: 'dist/worker.js',
workerPath: 'https://cdn.rawgit.com/naptha/tesseract.js/0.1.3/dist/worker.js', workerPath: 'https://cdn.rawgit.com/naptha/tesseract.js/0.1.3/dist/worker.js',
tesseractPath: 'https://cdn.rawgit.com/naptha/tesseract.js-core/0.1.0/index.js', tesseractPath: 'https://cdn.rawgit.com/naptha/tesseract.js-core/0.1.0/index.js',
langPath: 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/',
}
if(location.hostname === '127.0.0.1' && location.port == '7355'){
console.debug('Using Development Configuration')
defaultOptions.workerPath = location.protocol + '//' + location.host + '/dist/worker.js'
} }
exports.defaultOptions = defaultOptions;
exports.spawnWorker = function spawnWorker(instance, workerOptions){ exports.spawnWorker = function spawnWorker(instance, workerOptions){
if(window.Blob && window.URL){ if(window.Blob && window.URL){
var blob = new Blob(['importScripts("' + workerOptions.workerPath + '");']) var blob = new Blob(['importScripts("' + workerOptions.workerPath + '");'])

6
src/index.js
Unescape View File

@ -4,7 +4,7 @@ var adapter = require('./node/index.js')
var circularize = require('./common/circularize.js') var circularize = require('./common/circularize.js')
function createWorker(workerOptions){ function create(workerOptions){
return new TesseractWorker(workerOptions) return new TesseractWorker(workerOptions)
} }
@ -142,8 +142,8 @@ class TesseractJob {
} }
var DefaultTesseract = createWorker(adapter.defaultOptions) var DefaultTesseract = create(adapter.defaultOptions)
DefaultTesseract.createWorker = createWorker; DefaultTesseract.create = create;
module.exports = DefaultTesseract module.exports = DefaultTesseract

1
src/node/index.js
Unescape View File

@ -29,6 +29,7 @@ exports.sendPacket = function sendPacket(instance, packet){
function loadImage(image, cb){ function loadImage(image, cb){
// TODO: support URLs
if(typeof image === 'string'){ if(typeof image === 'string'){
fs.readFile(image, function(err, buffer){ fs.readFile(image, function(err, buffer){
loadImage(buffer, cb) loadImage(buffer, cb)

Write Preview
Loading…
Cancel
Save