1
|
|
|
/* -*- Mode: Java; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */ |
2
|
|
|
/* Copyright 2012 Mozilla Foundation |
3
|
|
|
* |
4
|
|
|
* Licensed under the Apache License, Version 2.0 (the "License"); |
5
|
|
|
* you may not use this file except in compliance with the License. |
6
|
|
|
* You may obtain a copy of the License at |
7
|
|
|
* |
8
|
|
|
* http://www.apache.org/licenses/LICENSE-2.0 |
9
|
|
|
* |
10
|
|
|
* Unless required by applicable law or agreed to in writing, software |
11
|
|
|
* distributed under the License is distributed on an "AS IS" BASIS, |
12
|
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
13
|
|
|
* See the License for the specific language governing permissions and |
14
|
|
|
* limitations under the License. |
15
|
|
|
*/ |
16
|
|
|
/* globals CustomStyle, PDFJS */ |
17
|
|
|
|
18
|
|
|
'use strict'; |
19
|
|
|
|
20
|
|
|
var MAX_TEXT_DIVS_TO_RENDER = 100000; |
21
|
|
|
|
22
|
|
|
var NonWhitespaceRegexp = /\S/; |
23
|
|
|
|
24
|
|
|
function isAllWhitespace(str) { |
25
|
|
|
return !NonWhitespaceRegexp.test(str); |
26
|
|
|
} |
27
|
|
|
|
28
|
|
|
/** |
29
|
|
|
* @typedef {Object} TextLayerBuilderOptions |
30
|
|
|
* @property {HTMLDivElement} textLayerDiv - The text layer container. |
31
|
|
|
* @property {number} pageIndex - The page index. |
32
|
|
|
* @property {PageViewport} viewport - The viewport of the text layer. |
33
|
|
|
* @property {PDFFindController} findController |
34
|
|
|
*/ |
35
|
|
|
|
36
|
|
|
/** |
37
|
|
|
* TextLayerBuilder provides text-selection functionality for the PDF. |
38
|
|
|
* It does this by creating overlay divs over the PDF text. These divs |
39
|
|
|
* contain text that matches the PDF text they are overlaying. This object |
40
|
|
|
* also provides a way to highlight text that is being searched for. |
41
|
|
|
* @class |
42
|
|
|
*/ |
43
|
|
|
var TextLayerBuilder = (function TextLayerBuilderClosure() { |
44
|
|
|
function TextLayerBuilder(options) { |
45
|
|
|
this.textLayerDiv = options.textLayerDiv; |
46
|
|
|
this.renderingDone = false; |
47
|
|
|
this.divContentDone = false; |
48
|
|
|
this.pageIdx = options.pageIndex; |
49
|
|
|
this.pageNumber = this.pageIdx + 1; |
50
|
|
|
this.matches = []; |
51
|
|
|
this.viewport = options.viewport; |
52
|
|
|
this.textDivs = []; |
53
|
|
|
this.findController = options.findController || null; |
54
|
|
|
} |
55
|
|
|
|
56
|
|
|
TextLayerBuilder.prototype = { |
57
|
|
|
_finishRendering: function TextLayerBuilder_finishRendering() { |
58
|
|
|
this.renderingDone = true; |
59
|
|
|
|
60
|
|
|
var event = document.createEvent('CustomEvent'); |
61
|
|
|
event.initCustomEvent('textlayerrendered', true, true, { |
62
|
|
|
pageNumber: this.pageNumber |
63
|
|
|
}); |
64
|
|
|
this.textLayerDiv.dispatchEvent(event); |
65
|
|
|
}, |
66
|
|
|
|
67
|
|
|
renderLayer: function TextLayerBuilder_renderLayer() { |
68
|
|
|
var textLayerFrag = document.createDocumentFragment(); |
69
|
|
|
var textDivs = this.textDivs; |
70
|
|
|
var textDivsLength = textDivs.length; |
71
|
|
|
var canvas = document.createElement('canvas'); |
72
|
|
|
var ctx = canvas.getContext('2d'); |
73
|
|
|
|
74
|
|
|
// No point in rendering many divs as it would make the browser |
75
|
|
|
// unusable even after the divs are rendered. |
76
|
|
|
if (textDivsLength > MAX_TEXT_DIVS_TO_RENDER) { |
77
|
|
|
this._finishRendering(); |
78
|
|
|
return; |
79
|
|
|
} |
80
|
|
|
|
81
|
|
|
var lastFontSize; |
82
|
|
|
var lastFontFamily; |
83
|
|
|
for (var i = 0; i < textDivsLength; i++) { |
84
|
|
|
var textDiv = textDivs[i]; |
85
|
|
|
if (textDiv.dataset.isWhitespace !== undefined) { |
86
|
|
|
continue; |
87
|
|
|
} |
88
|
|
|
|
89
|
|
|
var fontSize = textDiv.style.fontSize; |
90
|
|
|
var fontFamily = textDiv.style.fontFamily; |
91
|
|
|
|
92
|
|
|
// Only build font string and set to context if different from last. |
93
|
|
|
if (fontSize !== lastFontSize || fontFamily !== lastFontFamily) { |
|
|
|
|
94
|
|
|
ctx.font = fontSize + ' ' + fontFamily; |
95
|
|
|
lastFontSize = fontSize; |
96
|
|
|
lastFontFamily = fontFamily; |
97
|
|
|
} |
98
|
|
|
|
99
|
|
|
var width = ctx.measureText(textDiv.textContent).width; |
100
|
|
|
if (width > 0) { |
101
|
|
|
textLayerFrag.appendChild(textDiv); |
102
|
|
|
var transform; |
103
|
|
|
if (textDiv.dataset.canvasWidth !== undefined) { |
104
|
|
|
// Dataset values come of type string. |
105
|
|
|
var textScale = textDiv.dataset.canvasWidth / width; |
106
|
|
|
transform = 'scaleX(' + textScale + ')'; |
107
|
|
|
} else { |
108
|
|
|
transform = ''; |
109
|
|
|
} |
110
|
|
|
var rotation = textDiv.dataset.angle; |
111
|
|
|
if (rotation) { |
112
|
|
|
transform = 'rotate(' + rotation + 'deg) ' + transform; |
113
|
|
|
} |
114
|
|
|
if (transform) { |
115
|
|
|
CustomStyle.setProp('transform' , textDiv, transform); |
116
|
|
|
} |
117
|
|
|
} |
118
|
|
|
} |
119
|
|
|
|
120
|
|
|
this.textLayerDiv.appendChild(textLayerFrag); |
121
|
|
|
this._finishRendering(); |
122
|
|
|
this.updateMatches(); |
123
|
|
|
}, |
124
|
|
|
|
125
|
|
|
/** |
126
|
|
|
* Renders the text layer. |
127
|
|
|
* @param {number} timeout (optional) if specified, the rendering waits |
128
|
|
|
* for specified amount of ms. |
129
|
|
|
*/ |
130
|
|
|
render: function TextLayerBuilder_render(timeout) { |
131
|
|
|
if (!this.divContentDone || this.renderingDone) { |
132
|
|
|
return; |
133
|
|
|
} |
134
|
|
|
|
135
|
|
|
if (this.renderTimer) { |
136
|
|
|
clearTimeout(this.renderTimer); |
137
|
|
|
this.renderTimer = null; |
138
|
|
|
} |
139
|
|
|
|
140
|
|
|
if (!timeout) { // Render right away |
141
|
|
|
this.renderLayer(); |
142
|
|
|
} else { // Schedule |
143
|
|
|
var self = this; |
144
|
|
|
this.renderTimer = setTimeout(function() { |
145
|
|
|
self.renderLayer(); |
146
|
|
|
self.renderTimer = null; |
147
|
|
|
}, timeout); |
148
|
|
|
} |
149
|
|
|
}, |
150
|
|
|
|
151
|
|
|
appendText: function TextLayerBuilder_appendText(geom, styles) { |
152
|
|
|
var style = styles[geom.fontName]; |
153
|
|
|
var textDiv = document.createElement('div'); |
154
|
|
|
this.textDivs.push(textDiv); |
155
|
|
|
if (isAllWhitespace(geom.str)) { |
156
|
|
|
textDiv.dataset.isWhitespace = true; |
157
|
|
|
return; |
158
|
|
|
} |
159
|
|
|
var tx = PDFJS.Util.transform(this.viewport.transform, geom.transform); |
160
|
|
|
var angle = Math.atan2(tx[1], tx[0]); |
161
|
|
|
if (style.vertical) { |
162
|
|
|
angle += Math.PI / 2; |
163
|
|
|
} |
164
|
|
|
var fontHeight = Math.sqrt((tx[2] * tx[2]) + (tx[3] * tx[3])); |
165
|
|
|
var fontAscent = fontHeight; |
166
|
|
|
if (style.ascent) { |
167
|
|
|
fontAscent = style.ascent * fontAscent; |
168
|
|
|
} else if (style.descent) { |
169
|
|
|
fontAscent = (1 + style.descent) * fontAscent; |
170
|
|
|
} |
171
|
|
|
|
172
|
|
|
var left; |
173
|
|
|
var top; |
174
|
|
|
if (angle === 0) { |
175
|
|
|
left = tx[4]; |
176
|
|
|
top = tx[5] - fontAscent; |
177
|
|
|
} else { |
178
|
|
|
left = tx[4] + (fontAscent * Math.sin(angle)); |
179
|
|
|
top = tx[5] - (fontAscent * Math.cos(angle)); |
180
|
|
|
} |
181
|
|
|
textDiv.style.left = left + 'px'; |
182
|
|
|
textDiv.style.top = top + 'px'; |
183
|
|
|
textDiv.style.fontSize = fontHeight + 'px'; |
184
|
|
|
textDiv.style.fontFamily = style.fontFamily; |
185
|
|
|
|
186
|
|
|
textDiv.textContent = geom.str; |
187
|
|
|
// |fontName| is only used by the Font Inspector. This test will succeed |
188
|
|
|
// when e.g. the Font Inspector is off but the Stepper is on, but it's |
189
|
|
|
// not worth the effort to do a more accurate test. |
190
|
|
|
if (PDFJS.pdfBug) { |
191
|
|
|
textDiv.dataset.fontName = geom.fontName; |
192
|
|
|
} |
193
|
|
|
// Storing into dataset will convert number into string. |
194
|
|
|
if (angle !== 0) { |
195
|
|
|
textDiv.dataset.angle = angle * (180 / Math.PI); |
196
|
|
|
} |
197
|
|
|
// We don't bother scaling single-char text divs, because it has very |
198
|
|
|
// little effect on text highlighting. This makes scrolling on docs with |
199
|
|
|
// lots of such divs a lot faster. |
200
|
|
|
if (textDiv.textContent.length > 1) { |
201
|
|
|
if (style.vertical) { |
202
|
|
|
textDiv.dataset.canvasWidth = geom.height * this.viewport.scale; |
203
|
|
|
} else { |
204
|
|
|
textDiv.dataset.canvasWidth = geom.width * this.viewport.scale; |
205
|
|
|
} |
206
|
|
|
} |
207
|
|
|
}, |
208
|
|
|
|
209
|
|
|
setTextContent: function TextLayerBuilder_setTextContent(textContent) { |
210
|
|
|
this.textContent = textContent; |
211
|
|
|
|
212
|
|
|
var textItems = textContent.items; |
213
|
|
|
for (var i = 0, len = textItems.length; i < len; i++) { |
214
|
|
|
this.appendText(textItems[i], textContent.styles); |
215
|
|
|
} |
216
|
|
|
this.divContentDone = true; |
217
|
|
|
}, |
218
|
|
|
|
219
|
|
|
convertMatches: function TextLayerBuilder_convertMatches(matches) { |
220
|
|
|
var i = 0; |
221
|
|
|
var iIndex = 0; |
222
|
|
|
var bidiTexts = this.textContent.items; |
223
|
|
|
var end = bidiTexts.length - 1; |
224
|
|
|
var queryLen = (this.findController === null ? |
225
|
|
|
0 : this.findController.state.query.length); |
226
|
|
|
var ret = []; |
227
|
|
|
|
228
|
|
|
for (var m = 0, len = matches.length; m < len; m++) { |
229
|
|
|
// Calculate the start position. |
230
|
|
|
var matchIdx = matches[m]; |
231
|
|
|
|
232
|
|
|
// Loop over the divIdxs. |
233
|
|
|
while (i !== end && matchIdx >= (iIndex + bidiTexts[i].str.length)) { |
234
|
|
|
iIndex += bidiTexts[i].str.length; |
235
|
|
|
i++; |
236
|
|
|
} |
237
|
|
|
|
238
|
|
|
if (i === bidiTexts.length) { |
239
|
|
|
console.error('Could not find a matching mapping'); |
240
|
|
|
} |
241
|
|
|
|
242
|
|
|
var match = { |
243
|
|
|
begin: { |
244
|
|
|
divIdx: i, |
245
|
|
|
offset: matchIdx - iIndex |
246
|
|
|
} |
247
|
|
|
}; |
248
|
|
|
|
249
|
|
|
// Calculate the end position. |
250
|
|
|
matchIdx += queryLen; |
251
|
|
|
|
252
|
|
|
// Somewhat the same array as above, but use > instead of >= to get |
253
|
|
|
// the end position right. |
254
|
|
|
while (i !== end && matchIdx > (iIndex + bidiTexts[i].str.length)) { |
255
|
|
|
iIndex += bidiTexts[i].str.length; |
256
|
|
|
i++; |
257
|
|
|
} |
258
|
|
|
|
259
|
|
|
match.end = { |
260
|
|
|
divIdx: i, |
261
|
|
|
offset: matchIdx - iIndex |
262
|
|
|
}; |
263
|
|
|
ret.push(match); |
264
|
|
|
} |
265
|
|
|
|
266
|
|
|
return ret; |
267
|
|
|
}, |
268
|
|
|
|
269
|
|
|
renderMatches: function TextLayerBuilder_renderMatches(matches) { |
270
|
|
|
// Early exit if there is nothing to render. |
271
|
|
|
if (matches.length === 0) { |
272
|
|
|
return; |
273
|
|
|
} |
274
|
|
|
|
275
|
|
|
var bidiTexts = this.textContent.items; |
276
|
|
|
var textDivs = this.textDivs; |
277
|
|
|
var prevEnd = null; |
278
|
|
|
var pageIdx = this.pageIdx; |
279
|
|
|
var isSelectedPage = (this.findController === null ? |
280
|
|
|
false : (pageIdx === this.findController.selected.pageIdx)); |
281
|
|
|
var selectedMatchIdx = (this.findController === null ? |
282
|
|
|
-1 : this.findController.selected.matchIdx); |
283
|
|
|
var highlightAll = (this.findController === null ? |
284
|
|
|
false : this.findController.state.highlightAll); |
285
|
|
|
var infinity = { |
286
|
|
|
divIdx: -1, |
287
|
|
|
offset: undefined |
288
|
|
|
}; |
289
|
|
|
|
290
|
|
|
function beginText(begin, className) { |
291
|
|
|
var divIdx = begin.divIdx; |
292
|
|
|
textDivs[divIdx].textContent = ''; |
293
|
|
|
appendTextToDiv(divIdx, 0, begin.offset, className); |
294
|
|
|
} |
295
|
|
|
|
296
|
|
|
function appendTextToDiv(divIdx, fromOffset, toOffset, className) { |
297
|
|
|
var div = textDivs[divIdx]; |
298
|
|
|
var content = bidiTexts[divIdx].str.substring(fromOffset, toOffset); |
299
|
|
|
var node = document.createTextNode(content); |
300
|
|
|
if (className) { |
301
|
|
|
var span = document.createElement('span'); |
302
|
|
|
span.className = className; |
303
|
|
|
span.appendChild(node); |
304
|
|
|
div.appendChild(span); |
305
|
|
|
return; |
306
|
|
|
} |
307
|
|
|
div.appendChild(node); |
308
|
|
|
} |
309
|
|
|
|
310
|
|
|
var i0 = selectedMatchIdx, i1 = i0 + 1; |
311
|
|
|
if (highlightAll) { |
312
|
|
|
i0 = 0; |
313
|
|
|
i1 = matches.length; |
314
|
|
|
} else if (!isSelectedPage) { |
315
|
|
|
// Not highlighting all and this isn't the selected page, so do nothing. |
316
|
|
|
return; |
317
|
|
|
} |
318
|
|
|
|
319
|
|
|
for (var i = i0; i < i1; i++) { |
320
|
|
|
var match = matches[i]; |
321
|
|
|
var begin = match.begin; |
322
|
|
|
var end = match.end; |
323
|
|
|
var isSelected = (isSelectedPage && i === selectedMatchIdx); |
324
|
|
|
var highlightSuffix = (isSelected ? ' selected' : ''); |
325
|
|
|
|
326
|
|
|
if (this.findController) { |
327
|
|
|
this.findController.updateMatchPosition(pageIdx, i, textDivs, |
328
|
|
|
begin.divIdx, end.divIdx); |
329
|
|
|
} |
330
|
|
|
|
331
|
|
|
// Match inside new div. |
332
|
|
|
if (!prevEnd || begin.divIdx !== prevEnd.divIdx) { |
333
|
|
|
// If there was a previous div, then add the text at the end. |
334
|
|
|
if (prevEnd !== null) { |
335
|
|
|
appendTextToDiv(prevEnd.divIdx, prevEnd.offset, infinity.offset); |
336
|
|
|
} |
337
|
|
|
// Clear the divs and set the content until the starting point. |
338
|
|
|
beginText(begin); |
339
|
|
|
} else { |
340
|
|
|
appendTextToDiv(prevEnd.divIdx, prevEnd.offset, begin.offset); |
341
|
|
|
} |
342
|
|
|
|
343
|
|
|
if (begin.divIdx === end.divIdx) { |
344
|
|
|
appendTextToDiv(begin.divIdx, begin.offset, end.offset, |
345
|
|
|
'highlight' + highlightSuffix); |
346
|
|
|
} else { |
347
|
|
|
appendTextToDiv(begin.divIdx, begin.offset, infinity.offset, |
348
|
|
|
'highlight begin' + highlightSuffix); |
349
|
|
|
for (var n0 = begin.divIdx + 1, n1 = end.divIdx; n0 < n1; n0++) { |
350
|
|
|
textDivs[n0].className = 'highlight middle' + highlightSuffix; |
351
|
|
|
} |
352
|
|
|
beginText(end, 'highlight end' + highlightSuffix); |
353
|
|
|
} |
354
|
|
|
prevEnd = end; |
355
|
|
|
} |
356
|
|
|
|
357
|
|
|
if (prevEnd) { |
358
|
|
|
appendTextToDiv(prevEnd.divIdx, prevEnd.offset, infinity.offset); |
359
|
|
|
} |
360
|
|
|
}, |
361
|
|
|
|
362
|
|
|
updateMatches: function TextLayerBuilder_updateMatches() { |
363
|
|
|
// Only show matches when all rendering is done. |
364
|
|
|
if (!this.renderingDone) { |
365
|
|
|
return; |
366
|
|
|
} |
367
|
|
|
|
368
|
|
|
// Clear all matches. |
369
|
|
|
var matches = this.matches; |
370
|
|
|
var textDivs = this.textDivs; |
371
|
|
|
var bidiTexts = this.textContent.items; |
372
|
|
|
var clearedUntilDivIdx = -1; |
373
|
|
|
|
374
|
|
|
// Clear all current matches. |
375
|
|
|
for (var i = 0, len = matches.length; i < len; i++) { |
376
|
|
|
var match = matches[i]; |
377
|
|
|
var begin = Math.max(clearedUntilDivIdx, match.begin.divIdx); |
378
|
|
|
for (var n = begin, end = match.end.divIdx; n <= end; n++) { |
379
|
|
|
var div = textDivs[n]; |
380
|
|
|
div.textContent = bidiTexts[n].str; |
381
|
|
|
div.className = ''; |
382
|
|
|
} |
383
|
|
|
clearedUntilDivIdx = match.end.divIdx + 1; |
384
|
|
|
} |
385
|
|
|
|
386
|
|
|
if (this.findController === null || !this.findController.active) { |
387
|
|
|
return; |
388
|
|
|
} |
389
|
|
|
|
390
|
|
|
// Convert the matches on the page controller into the match format |
391
|
|
|
// used for the textLayer. |
392
|
|
|
this.matches = this.convertMatches(this.findController === null ? |
393
|
|
|
[] : (this.findController.pageMatches[this.pageIdx] || [])); |
394
|
|
|
this.renderMatches(this.matches); |
395
|
|
|
} |
396
|
|
|
}; |
397
|
|
|
return TextLayerBuilder; |
398
|
|
|
})(); |
399
|
|
|
|
400
|
|
|
/** |
401
|
|
|
* @constructor |
402
|
|
|
* @implements IPDFTextLayerFactory |
403
|
|
|
*/ |
404
|
|
|
function DefaultTextLayerFactory() {} |
405
|
|
|
DefaultTextLayerFactory.prototype = { |
406
|
|
|
/** |
407
|
|
|
* @param {HTMLDivElement} textLayerDiv |
408
|
|
|
* @param {number} pageIndex |
409
|
|
|
* @param {PageViewport} viewport |
410
|
|
|
* @returns {TextLayerBuilder} |
411
|
|
|
*/ |
412
|
|
|
createTextLayerBuilder: function (textLayerDiv, pageIndex, viewport) { |
413
|
|
|
return new TextLayerBuilder({ |
414
|
|
|
textLayerDiv: textLayerDiv, |
415
|
|
|
pageIndex: pageIndex, |
416
|
|
|
viewport: viewport |
417
|
|
|
}); |
418
|
|
|
} |
419
|
|
|
}; |
420
|
|
|
|