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

update contrib

pull/12/head
Guillermo 8 years ago
parent
c7f6364710
commit
c61ee098d7
3 changed files with 378 additions and 438 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. 6
      README.md
  2. 450
      dist/tesseract.js
  3. 360
      dist/worker.js

6
README.md
Unescape View File

@ -265,8 +265,10 @@ Then, cd in to the folder, `npm install`, and `npm start` @@ -265,8 +265,10 @@ Then, cd in to the folder, `npm install`, and `npm start`
... a bunch of npm stuff ...
tesseract.js@1.0.0 start /Users/guillermo/Desktop/code_static/tesseract.js
node devServer.js
Starting up http-server, serving ./
Available on:
http://127.0.0.1:7355
http://18.111.40.187:7355
Listening at http://localhost:7355
```

450
dist/tesseract.js vendored
Unescape View File

@ -1,85 +1,78 @@ @@ -1,85 +1,78 @@
(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',
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',
}
exports.spawnWorker = function spawnWorker(instance, workerOptions) {
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));
} else {
var worker = new Worker(workerOptions.workerPath);
}else{
var worker = new Worker(workerOptions.workerPath)
}
worker.onmessage = function (e) {
worker.onmessage = function(e){
var packet = e.data;
instance._recv(packet);
};
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)) {
instance._recv(packet)
}
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 {
return loadImage(document.querySelector(image), cb)
}else{
// url or path
var im = new Image();
var im = new Image
im.src = image;
im.onload = function (e) {
return loadImage(im, cb);
};
return;
im.onload = e => loadImage(im, cb);
return
}
} else if (image instanceof File) {
}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) {
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") {
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.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) {
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);
return loadImage(data, cb)
}
cb(image);
cb(image)
}
},{}],2:[function(require,module,exports){
"use strict";
// The result of dump.js is a big JSON tree
// which can be easily serialized (for instance
// to be sent from a webworker to the main app
@ -87,251 +80,212 @@ function loadImage(image, cb) { @@ -87,251 +80,212 @@ function loadImage(image, cb) {
// a (circular) DOM-like interface for walking
// through the data.
module.exports = function circularize(page) {
page.paragraphs = [];
page.lines = [];
page.words = [];
page.symbols = [];
module.exports = function circularize(page){
page.paragraphs = []
page.lines = []
page.words = []
page.symbols = []
page.blocks.forEach(function (block) {
page.blocks.forEach(function(block){
block.page = page;
block.lines = [];
block.words = [];
block.symbols = [];
block.lines = []
block.words = []
block.symbols = []
block.paragraphs.forEach(function (para) {
block.paragraphs.forEach(function(para){
para.block = block;
para.page = page;
para.words = [];
para.symbols = [];
para.lines.forEach(function (line) {
para.words = []
para.symbols = []
para.lines.forEach(function(line){
line.paragraph = para;
line.block = block;
line.page = page;
line.symbols = [];
line.symbols = []
line.words.forEach(function (word) {
line.words.forEach(function(word){
word.line = line;
word.paragraph = para;
word.block = block;
word.page = page;
word.symbols.forEach(function (sym) {
word.symbols.forEach(function(sym){
sym.word = word;
sym.line = line;
sym.paragraph = para;
sym.block = block;
sym.page = page;
sym.line.symbols.push(sym);
sym.paragraph.symbols.push(sym);
sym.block.symbols.push(sym);
sym.page.symbols.push(sym);
});
word.paragraph.words.push(word);
word.block.words.push(word);
word.page.words.push(word);
});
line.block.lines.push(line);
line.page.lines.push(line);
});
para.page.paragraphs.push(para);
});
});
return page;
};
sym.line.symbols.push(sym)
sym.paragraph.symbols.push(sym)
sym.block.symbols.push(sym)
sym.page.symbols.push(sym)
})
word.paragraph.words.push(word)
word.block.words.push(word)
word.page.words.push(word)
})
line.block.lines.push(line)
line.page.lines.push(line)
})
para.page.paragraphs.push(para)
})
})
return page
}
},{}],3:[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')
var circularize = require('./common/circularize.js')
var adapter = require('./node/index.js');
var circularize = require('./common/circularize.js');
function createWorker(workerOptions) {
return new TesseractWorker(workerOptions);
function createWorker(workerOptions){
return new TesseractWorker(workerOptions)
}
var TesseractWorker = function () {
function TesseractWorker(workerOptions) {
_classCallCheck(this, TesseractWorker);
class TesseractWorker {
constructor(workerOptions){
this.worker = null;
this.workerOptions = workerOptions;
this._currentJob = null;
this._queue = [];
this._queue = []
}
_createClass(TesseractWorker, [{
key: 'recognize',
value: function recognize(image, options) {
var _this = this;
return this._delay(function (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 });
});
}
}, {
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]();
recognize(image, options){
return this._delay(job => {
if(typeof options === 'string'){
options = { lang: options };
}else{
options = options || {}
options.lang = options.lang || 'eng';
}
}
}, {
key: '_recv',
value: function _recv(packet) {
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 })
})
}
if (packet.status === 'resolve' && packet.action === 'recognize') {
packet.data = circularize(packet.data);
}
terminate(){
if(this.worker) adapter.terminateWorker(this);
this.worker = null;
}
if (this._currentJob.id === packet.jobId) {
this._currentJob._handle(packet);
} else {
console.warn('Job ID ' + packet.jobId + ' not known.');
}
_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]()
}
}]);
}
return TesseractWorker;
}();
_recv(packet){
var jobCounter = 0;
if(packet.status === 'resolve' && packet.action === 'recognize'){
packet.data = circularize(packet.data);
}
var TesseractJob = function () {
function TesseractJob(instance) {
_classCallCheck(this, TesseractJob);
if(this._currentJob.id === packet.jobId){
this._currentJob._handle(packet)
}else{
console.warn('Job ID ' + packet.jobId + ' not known.')
}
}
}
this.id = 'Job-' + ++jobCounter + '-' + Math.random().toString(16).slice(3, 8);
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 = [];
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;
then(resolve, reject){
if(this._resolve.push){
this._resolve.push(resolve)
}else{
resolve(this._resolve)
}
}, {
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
});
if(reject) this.catch(reject);
return this;
}
catch(reject){
if(this._reject.push){
this._reject.push(reject)
}else{
reject(this._reject)
}
}, {
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 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)
}
}]);
}
}
return TesseractJob;
}();
var DefaultTesseract = createWorker(adapter.defaultOptions);
var DefaultTesseract = createWorker(adapter.defaultOptions)
DefaultTesseract.createWorker = createWorker;
module.exports = DefaultTesseract;
module.exports = DefaultTesseract
},{"./common/circularize.js":2,"./node/index.js":1}]},{},[3])(3)
});

360
dist/worker.js vendored
Unescape View File

@ -11854,38 +11854,33 @@ function extend() { @@ -11854,38 +11854,33 @@ function extend() {
}
},{"./has-keys":40,"object-keys":17}],42:[function(require,module,exports){
'use strict';
var leveljs = require('level-js');
var db = typeof indexedDB === 'undefined' ? { open: function open(_, cb) {
return cb(true);
} } : leveljs('./tessdata2');
var leveljs = require('level-js')
var db = typeof indexedDB === 'undefined' ? { open: (_, cb) => 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;
function saveDataFile(data) {
db.put(lang, data, function (err) {
return console.log('cached', lang, err);
});
cb(data);
function saveDataFile(data){
db.put(lang, data, err => console.log('cached', lang, err))
cb(data)
}
db.open({ compression: false }, function (err) {
db.open({ compression: false }, err => {
if (err) return fetchLanguageData(req, res, cb);
db.get(lang, function (err, data) {
db.get(lang, (err, data) => {
if (err) return fetchLanguageData(req, res, saveDataFile);
res.progress({ status: 'found in cache ' + lang + '.traineddata' });
cb(data);
});
});
};
res.progress({ status: 'found in cache ' + lang + '.traineddata' })
cb(data)
})
})
}
var ungzip = require('pako').ungzip;
function fetchLanguageData(req, res, cb) {
function fetchLanguageData(req, res, cb){
var lang = req.options.lang;
var langfile = lang + '.traineddata.gz';
var url = req.workerOptions.langPath + langfile;
@ -11893,130 +11888,114 @@ function fetchLanguageData(req, res, cb) { @@ -11893,130 +11888,114 @@ function fetchLanguageData(req, res, cb) {
var xhr = new XMLHttpRequest();
xhr.responseType = 'arraybuffer';
xhr.open('GET', url, true);
xhr.onerror = function (e) {
xhr.onprogress = xhr.onload = null;
cb(xhr, null);
};
xhr.onprogress = function (e) {
return res.progress({
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 = function (e) {
if (!(xhr.status == 200 || xhr.status == 0 && xhr.response)) return res.reject('Error downloading language ' + url);
res.progress({ status: 'unzipping ' + langfile });
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);
var response = new Uint8Array(xhr.response)
try {
while (response[0] == 0x1f && response[1] == 0x8b) {
response = ungzip(response);
}
while(response[0] == 0x1f && response[1] == 0x8b) response = ungzip(response);
} 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);
};
xhr.send();
cb(response)
}
xhr.send()
}
},{"../common/langdata.json":46,"level-js":13,"pako":20}],43:[function(require,module,exports){
(function (global){
"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;
workerUtils.dispatchHandlers(packet, function (obj) {
return postMessage(obj);
});
});
workerUtils.dispatchHandlers(packet, obj => postMessage(obj))
})
exports.getCore = function (req, res) {
if (!global.TesseractCore) {
res.progress({ status: 'loading tesseract core' });
importScripts(req.workerOptions.tesseractPath);
res.progress({ status: 'loaded tesseract core' });
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;
};
return TesseractCore
}
exports.getLanguageData = require('./lang.js');
exports.getLanguageData = require('./lang.js')
workerUtils.setAdapter(module.exports);
}).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){
'use strict';
// This converts an image to grayscale
module.exports = function desaturate(image) {
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;
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;
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 'Invalid ImageData';
}
return image;
};
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 'Invalid ImageData' }
return image
}
},{}],45:[function(require,module,exports){
'use strict';
module.exports = function DumpLiterallyEverything(Module, base) {
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];
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();
ri.Begin()
do {
if (ri.IsAtBeginningOf(Module.RIL_BLOCK)) {
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) {
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++) {
for(var i = 0; i < n; i++){
polygon.push([px.getValue(i), py.getValue(i)]);
}
Module._ptaDestroy(Module.getPointer(poly));
Module._ptaDestroy(Module.getPointer(poly));
}
block = {
paragraphs: [],
@ -12027,10 +12006,10 @@ module.exports = function DumpLiterallyEverything(Module, base) { @@ -12027,10 +12006,10 @@ module.exports = function DumpLiterallyEverything(Module, base) {
blocktype: enumToString(ri.BlockType(), 'PT'),
polygon: polygon
};
blocks.push(block);
}
blocks.push(block)
}
if (ri.IsAtBeginningOf(Module.RIL_PARA)) {
if(ri.IsAtBeginningOf(Module.RIL_PARA)){
para = {
lines: [],
@ -12040,10 +12019,10 @@ module.exports = function DumpLiterallyEverything(Module, base) { @@ -12040,10 +12019,10 @@ module.exports = function DumpLiterallyEverything(Module, base) {
bbox: ri.getBoundingBox(Module.RIL_PARA),
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 = {
words: [],
@ -12051,10 +12030,10 @@ module.exports = function DumpLiterallyEverything(Module, base) { @@ -12051,10 +12030,10 @@ module.exports = function DumpLiterallyEverything(Module, base) {
confidence: ri.Confidence(Module.RIL_TEXTLINE),
baseline: ri.getBaseline(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(),
wordDir = ri.WordDirection();
word = {
@ -12079,25 +12058,25 @@ module.exports = function DumpLiterallyEverything(Module, base) { @@ -12079,25 +12058,25 @@ module.exports = function DumpLiterallyEverything(Module, base) {
is_smallcaps: fontInfo.is_smallcaps,
font_size: fontInfo.pointsize,
font_id: fontInfo.font_id,
font_name: fontInfo.font_name
};
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);
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)) {
if(ri.IsAtBeginningOf(Module.RIL_SYMBOL)){
symbol = {
choices: [],
image: image,
@ -12109,20 +12088,20 @@ module.exports = function DumpLiterallyEverything(Module, base) { @@ -12109,20 +12088,20 @@ module.exports = function DumpLiterallyEverything(Module, base) {
is_superscript: !!ri.SymbolIsSuperscript(),
is_subscript: !!ri.SymbolIsSubscript(),
is_dropcap: !!ri.SymbolIsDropcap()
};
word.symbols.push(symbol);
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);
Module.destroy(ci)
}
} while (ri.Next(Module.RIL_SYMBOL));
Module.destroy(ri);
Module.destroy(ri)
return {
text: base.GetUTF8Text(),
@ -12134,116 +12113,120 @@ module.exports = function DumpLiterallyEverything(Module, base) { @@ -12134,116 +12113,120 @@ module.exports = function DumpLiterallyEverything(Module, base) {
psm: enumToString(base.GetPageSegMode(), 'PSM'),
oem: enumToString(base.oem(), 'OEM'),
version: base.Version()
};
};
version: base.Version(),
}
}
// the generated HOCR is excessively indented, so
// we get rid of that indentation
function deindent(html) {
var lines = html.split('\n');
if (lines[0].substring(0, 2) === " ") {
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);
if (lines[i].substring(0,2) === " ") {
lines[i] = lines[i].slice(2)
}
};
}
return lines.join('\n');
return lines.join('\n')
}
},{}],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}
},{}],47:[function(require,module,exports){
'use strict';
var latestJob;
var Module;
var base;
var adapter = {};
function dispatchHandlers(packet, send) {
function respond(status, data) {
function dispatchHandlers(packet, send){
function respond(status, data){
send({
jobId: packet.jobId,
status: status,
action: packet.action,
data: data
});
})
}
respond.resolve = respond.bind(this, 'resolve');
respond.reject = respond.bind(this, 'reject');
respond.progress = respond.bind(this, 'progress');
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);
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) {
exports.setAdapter = function setAdapter(impl){
adapter = impl;
};
}
function handleInit(req, res) {
if (!Module) {
function handleInit(req, res){
if(!Module){
var Core = adapter.getCore(req, res);
res.progress({ status: 'initializing tesseract api' });
res.progress({ status: 'initializing tesseract api' })
Module = Core({
TOTAL_MEMORY: req.memory,
TesseractProgress: function TesseractProgress(percent) {
latestJob.progress({ status: 'recognizing text', progress: Math.max(0, (percent - 30) / 70) });
TesseractProgress(percent){
latestJob.progress({ status: 'recognizing text', progress: Math.max(0, (percent-30)/70) })
},
onRuntimeInitialized: function onRuntimeInitialized() {}
});
Module.FS_createPath("/", "tessdata", true, true);
base = new Module.TessBaseAPI();
res.progress({ status: 'initialized tesseract api' });
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 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);
base.SetRectangle(0, 0, width, height)
return ptr;
}
function loadLanguage(req, res, cb) {
function loadLanguage(req, res, cb){
var lang = req.options.lang;
if(!Module._loadedLanguages) Module._loadedLanguages = {};
if(lang in Module._loadedLanguages) return cb();
if (!Module._loadedLanguages) Module._loadedLanguages = {};
if (lang in Module._loadedLanguages) return cb();
adapter.getLanguageData(req, res, function (data) {
adapter.getLanguageData(req, res, function(data){
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;
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;
base.Init(null, lang);
res.progress({ status: 'initialized with language' });
base.Init(null, lang)
res.progress({ status: 'initialized with language' })
var options = req.options;
for (var option in options) {
@ -12253,37 +12236,38 @@ function handleRecognize(req, res) { @@ -12253,37 +12236,38 @@ function handleRecognize(req, res) {
}
var ptr = setImage(Module, base, req.image);
base.Recognize(null);
var result = dump(Module, base);
base.Recognize(null)
var result = dump(Module, base)
base.End();
Module._free(ptr);
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);
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) {
if(!success){
base.End();
Module._free(ptr);
res.reject("failed to detect os");
res.reject("failed to detect os")
} else {
var charset = results.get_unicharset();
var best = results.get_best_result();
var charset = results.get_unicharset()
var best = results.get_best_result()
var oid = best.get_orientation_id(),
sid = best.get_script_id();
@ -12293,14 +12277,14 @@ function handleDetect(req, res) { @@ -12293,14 +12277,14 @@ function handleDetect(req, res) {
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);
res.resolve(result)
}
});
})
}
},{"./desaturate.js":44,"./dump.js":45}]},{},[43]);

Write Preview
Loading…
Cancel
Save