added progress callback

example.htm

@@ -13,5 +13,10 @@
 	// ctx.fillText('2小時可換乘2次2小時可換乘2次', 100, 80)
 	ctx.font = '30px sans-serif'
 	ctx.fillText('the Cosmic Void', 100, 120)
-	Tesseract.recognize(canvas,{tessedit_char_blacklist:'e'}).then( function(d){ console.log(d) } )
+	Tesseract.recognize(canvas,{
+		tessedit_char_blacklist:'e',
+		progress: function(e){
+			console.log(e)
+		}
+	}).then( function(d){ console.log(d) } )
 </script>

lib/Tesseract.js

@@ -34,11 +34,26 @@ Tesseract.recognize = function(image, options, callback){
 	}
 	if(image.getImageData) image = image.getImageData(0, 0, image.canvas.width, image.canvas.height);		
 
-	var worker = new Worker('https://cdn.rawgit.com/naptha/tesseract.js/master/lib/worker.js')
+	var worker = new Worker('/lib/worker.js')
+
+	var progress = (function(){
+		if(typeof options.progress === 'function'){
+			var p = options.progress
+			delete options.progress
+			return p
+		}
+		return function(){}
+	})()
+
 
 	if(typeof callback === "function"){
 		worker.onmessage = function(e){
-			callback(e.data.err, e.data.result)
+			if(e.data.progress){
+				progress(e.data.progress)
+			}
+			else{
+				callback(e.data.err, e.data.result)
+			}
 		}
 		worker.postMessage({image: image, lang: lang})		
 		console.log('callback')
@@ -46,7 +61,10 @@ Tesseract.recognize = function(image, options, callback){
 	else {
 		return new Promise(function(resolve, reject){
 			worker.onmessage = function(e){
-				if(e.data.err){
+				if(e.data.progress){
+					progress(e.data.progress)
+				}
+				else if(e.data.err){
 					reject(e.data.err)
 				}
 				else {

lib/worker.js

@@ -6489,7 +6489,11 @@ var recognize = (function createTesseractInstance(){
 	var Module = Tesseract304({
 		TOTAL_MEMORY: 90e6,
 		TesseractProgress: function(percent){
-			console.log('recognized',percent+'%')
+			postMessage({
+				'progress': {
+					'recognized': percent/100
+				}
+			})
 		}
 	})
 
@@ -6505,12 +6509,22 @@ var recognize = (function createTesseractInstance(){
 			xhr.open('GET', 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/'+lang+'.traineddata.gz', true);
 			xhr.responseType = 'arraybuffer';
 			xhr.onerror = function(){ cb(xhr, null) }
-			xhr.onprogress = function(e){console.log('loading',lang,'language model:',Math.round(e.loaded/filesizes[lang]*100)+'%')}
+			xhr.onprogress = function(e){
+				postMessage({
+					'progress': {
+						'loaded_lang_model': e.loaded/filesizes[lang]
+					}
+				})
+			}
 			xhr.onload = function(){
 				if (xhr.status == 200 || (xhr.status == 0 && xhr.response)) {
-					console.log('unzipping language model...')
-					var data = new Uint8Array(pako.deflate(new Uint8Array(xhr.response)))
-					console.log(lang +".traineddata", 'sucessfully unzipped')
+					postMessage({
+						'progress': 'unzipping_lang_model'
+					})
+					var data = pako.inflate(new Uint8Array(xhr.response))
+					postMessage({
+						'progress': 'unzipped_lang_model'
+					})
 					Module.FS_createDataFile('tessdata', lang +".traineddata", data, true, false);
 					loaded_langs.push(lang)
 					cb(null, lang)
@@ -6698,7 +6712,7 @@ var recognize = (function createTesseractInstance(){
 		var ptr = Module.allocate(image, 'i8', Module.ALLOC_NORMAL);
 		
 		loadLanguage(lang, function(err, result){
-			
+
 			if(err){
 				console.error("error loading", lang);
 				cb(err, null)
@@ -6707,7 +6721,14 @@ var recognize = (function createTesseractInstance(){
 			for (var option in options) {
 			    if (options.hasOwnProperty(option)) {
 			        base.SetVariable(option, options[option]);
-			        console.log('setting', option, '=', options[option]);
+			        postMessage({
+						progress: {
+							set_variable: {
+								variable: option,
+								value: options[option]
+							}
+						}
+					})
 			    }
 			}
 

worker_src/worker.js

@@ -79,7 +79,11 @@ var recognize = (function createTesseractInstance(){
 	var Module = Tesseract304({
 		TOTAL_MEMORY: 90e6,
 		TesseractProgress: function(percent){
-			console.log('recognized',percent+'%')
+			postMessage({
+				'progress': {
+					'recognized': percent/100
+				}
+			})
 		}
 	})
 
@@ -95,12 +99,22 @@ var recognize = (function createTesseractInstance(){
 			xhr.open('GET', 'https://cdn.rawgit.com/naptha/tessdata/gh-pages/3.02/'+lang+'.traineddata.gz', true);
 			xhr.responseType = 'arraybuffer';
 			xhr.onerror = function(){ cb(xhr, null) }
-			xhr.onprogress = function(e){console.log('loading',lang,'language model:',Math.round(e.loaded/filesizes[lang]*100)+'%')}
+			xhr.onprogress = function(e){
+				postMessage({
+					'progress': {
+						'loaded_lang_model': e.loaded/filesizes[lang]
+					}
+				})
+			}
 			xhr.onload = function(){
 				if (xhr.status == 200 || (xhr.status == 0 && xhr.response)) {
-					console.log('unzipping language model...')
-					var data = new Uint8Array(pako.deflate(new Uint8Array(xhr.response)))
-					console.log(lang +".traineddata", 'sucessfully unzipped')
+					postMessage({
+						'progress': 'unzipping_lang_model'
+					})
+					var data = pako.inflate(new Uint8Array(xhr.response))
+					postMessage({
+						'progress': 'unzipped_lang_model'
+					})
 					Module.FS_createDataFile('tessdata', lang +".traineddata", data, true, false);
 					loaded_langs.push(lang)
 					cb(null, lang)
@@ -288,7 +302,7 @@ var recognize = (function createTesseractInstance(){
 		var ptr = Module.allocate(image, 'i8', Module.ALLOC_NORMAL);
 		
 		loadLanguage(lang, function(err, result){
-			
+
 			if(err){
 				console.error("error loading", lang);
 				cb(err, null)
@@ -297,7 +311,14 @@ var recognize = (function createTesseractInstance(){
 			for (var option in options) {
 			    if (options.hasOwnProperty(option)) {
 			        base.SetVariable(option, options[option]);
-			        console.log('setting', option, '=', options[option]);
+			        postMessage({
+						progress: {
+							set_variable: {
+								variable: option,
+								value: options[option]
+							}
+						}
+					})
 			    }
 			}
 
@@ -312,7 +333,6 @@ var recognize = (function createTesseractInstance(){
 		})
 	}
 
-	// base._simple = _simple
 	return recognize
 })()