Sakurai/pdf.js
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Split large image groups into smaller chunks

Browse Source
This commit is contained in:
Yury Delendik 2012-12-07 17:19:06 -06:00
parent d71c702dcf
commit cae62341ac
3 changed files with 122 additions and 113 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
src/canvas.js
View File

@ -330,7 +330,10 @@ var CanvasGraphics = (function CanvasGraphicsClosure() {
'setFillCMYKColor': true,
'paintJpegXObject': true,
'paintImageXObject': true,
'paintInlineImageXObject': true,
'paintInlineImageXObjectGroup': true,
'paintImageMaskXObject': true,
'paintImageMaskXObjectGroup': true,
'shadingFill': true
},

4
src/core.js
View File

@ -193,7 +193,9 @@ var Page = (function PageClosure() {
xref, handler, this.pageIndex,
'p' + this.pageIndex + '_');
return pe.getOperatorList(contentStream, resources, dependency);
var list = pe.getOperatorList(contentStream, resources, dependency);
pe.optimizeQueue(list);
return list;
},
extractTextContent: function Page_extractTextContent() {
var handler = {

228
src/evaluator.js
View File

@ -298,116 +298,6 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
}, handler, xref, resources, image, inline);
}
function optimizeQueue() {
// grouping paintInlineImageXObject's into paintInlineImageXObjectGroup
// searching for (save, transform, paintInlineImageXObject, restore)+
var MIN_IMAGES_COUNT = 10;
var MAX_WIDTH = 1000;
var IMAGE_PADDING = 1;
for (var i = 0, ii = fnArray.length; i < ii; i++) {
if (fnArray[i] === 'paintInlineImageXObject' &&
fnArray[i - 2] === 'save' && fnArray[i - 1] === 'transform' &&
fnArray[i + 1] === 'restore') {
var j = i - 2;
for (i += 2; i < ii && fnArray[i - 4] === fnArray[i]; i++) {
}
var count = (i - j) >> 2;
if (count < MIN_IMAGES_COUNT) {
continue;
}
// assuming that heights of those image is too small (~1 pixel)
// packing as much as possible by lines
var maxX = 0;
var map = [], maxLineHeight = 0;
var currentX = IMAGE_PADDING, currentY = IMAGE_PADDING;
for (var q = 0; q < count; q++) {
var transform = argsArray[j + (q << 2) + 1];
var img = argsArray[j + (q << 2) + 2][0];
if (currentX + img.width > MAX_WIDTH) {
// starting new line
maxX = Math.max(maxX, currentX);
currentY += maxLineHeight + 2 * IMAGE_PADDING;
currentX = 0;
maxLineHeight = 0;
}
map.push({
transform: transform,
x: currentX, y: currentY,
w: img.width, h: img.height
});
currentX += img.width + 2 * IMAGE_PADDING;
maxLineHeight = Math.max(maxLineHeight, img.height);
}
var imgWidth = Math.max(maxX, currentX) + IMAGE_PADDING;
var imgHeight = currentY + maxLineHeight + IMAGE_PADDING;
var imgData = new Uint8Array(imgWidth * imgHeight * 4);
var imgRowSize = imgWidth << 2;
for (var q = 0; q < count; q++) {
var data = argsArray[j + (q << 2) + 2][0].data;
// copy image by lines and extends pixels into padding
var rowSize = map[q].w << 2;
var dataOffset = 0;
var offset = (map[q].x + map[q].y * imgWidth) << 2;
imgData.set(
data.subarray(0, rowSize), offset - imgRowSize);
for (var k = 0, kk = map[q].h; k < kk; k++) {
imgData.set(
data.subarray(dataOffset, dataOffset + rowSize), offset);
dataOffset += rowSize;
offset += imgRowSize;
}
imgData.set(
data.subarray(dataOffset - rowSize, dataOffset), offset);
while (offset >= 0) {
data[offset - 4] = data[offset];
data[offset - 3] = data[offset + 1];
data[offset - 2] = data[offset + 2];
data[offset - 1] = data[offset + 3];
data[offset + rowSize] = data[offset + rowSize - 4];
data[offset + rowSize + 1] = data[offset + rowSize - 3];
data[offset + rowSize + 2] = data[offset + rowSize - 2];
data[offset + rowSize + 3] = data[offset + rowSize - 1];
offset -= imgRowSize;
}
}
// replacing queue items
fnArray.splice(j, count * 4, ['paintInlineImageXObjectGroup']);
argsArray.splice(j, count * 4,
[{width: imgWidth, height: imgHeight, data: imgData}, map]);
i = j;
ii = fnArray.length;
}
}
// grouping paintImageMaskXObject's into paintImageMaskXObjectGroup
// searching for (save, transform, paintImageMaskXObject, restore)+
for (var i = 0, ii = fnArray.length; i < ii; i++) {
if (fnArray[i] === 'paintImageMaskXObject' &&
fnArray[i - 2] === 'save' && fnArray[i - 1] === 'transform' &&
fnArray[i + 1] === 'restore') {
var j = i - 2;
for (i += 2; i < ii && fnArray[i - 4] === fnArray[i]; i++) {
}
var count = (i - j) >> 2;
if (count < MIN_IMAGES_COUNT) {
continue;
}
var images = [];
for (var q = 0; q < count; q++) {
var transform = argsArray[j + (q << 2) + 1];
var maskParams = argsArray[j + (q << 2) + 2];
images.push({data: maskParams[0], width: maskParams[2],
height: maskParams[3], transform: transform,
inverseDecode: maskParams[1]});
}
// replacing queue items
fnArray.splice(j, count * 4, ['paintImageMaskXObjectGroup']);
argsArray.splice(j, count * 4, [images]);
i = j;
ii = fnArray.length;
}
}
}
if (!queue)
queue = {};
@ -624,11 +514,125 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
}
}
optimizeQueue();
return queue;
},
optimizeQueue: function PartialEvaluator_optimizeQueue(queue) {
var fnArray = queue.fnArray, argsArray = queue.argsArray;
// grouping paintInlineImageXObject's into paintInlineImageXObjectGroup
// searching for (save, transform, paintInlineImageXObject, restore)+
var MIN_IMAGES_IN_INLINE_IMAGES_BLOCK = 10;
var MAX_IMAGES_IN_INLINE_IMAGES_BLOCK = 200;
var MAX_WIDTH = 1000;
var IMAGE_PADDING = 1;
for (var i = 0, ii = fnArray.length; i < ii; i++) {
if (fnArray[i] === 'paintInlineImageXObject' &&
fnArray[i - 2] === 'save' && fnArray[i - 1] === 'transform' &&
fnArray[i + 1] === 'restore') {
var j = i - 2;
for (i += 2; i < ii && fnArray[i - 4] === fnArray[i]; i++) {
}
var count = Math.min((i - j) >> 2,
MAX_IMAGES_IN_INLINE_IMAGES_BLOCK);
if (count < MIN_IMAGES_IN_INLINE_IMAGES_BLOCK) {
continue;
}
// assuming that heights of those image is too small (~1 pixel)
// packing as much as possible by lines
var maxX = 0;
var map = [], maxLineHeight = 0;
var currentX = IMAGE_PADDING, currentY = IMAGE_PADDING;
for (var q = 0; q < count; q++) {
var transform = argsArray[j + (q << 2) + 1];
var img = argsArray[j + (q << 2) + 2][0];
if (currentX + img.width > MAX_WIDTH) {
// starting new line
maxX = Math.max(maxX, currentX);
currentY += maxLineHeight + 2 * IMAGE_PADDING;
currentX = 0;
maxLineHeight = 0;
}
map.push({
transform: transform,
x: currentX, y: currentY,
w: img.width, h: img.height
});
currentX += img.width + 2 * IMAGE_PADDING;
maxLineHeight = Math.max(maxLineHeight, img.height);
}
var imgWidth = Math.max(maxX, currentX) + IMAGE_PADDING;
var imgHeight = currentY + maxLineHeight + IMAGE_PADDING;
var imgData = new Uint8Array(imgWidth * imgHeight * 4);
var imgRowSize = imgWidth << 2;
for (var q = 0; q < count; q++) {
var data = argsArray[j + (q << 2) + 2][0].data;
// copy image by lines and extends pixels into padding
var rowSize = map[q].w << 2;
var dataOffset = 0;
var offset = (map[q].x + map[q].y * imgWidth) << 2;
imgData.set(
data.subarray(0, rowSize), offset - imgRowSize);
for (var k = 0, kk = map[q].h; k < kk; k++) {
imgData.set(
data.subarray(dataOffset, dataOffset + rowSize), offset);
dataOffset += rowSize;
offset += imgRowSize;
}
imgData.set(
data.subarray(dataOffset - rowSize, dataOffset), offset);
while (offset >= 0) {
data[offset - 4] = data[offset];
data[offset - 3] = data[offset + 1];
data[offset - 2] = data[offset + 2];
data[offset - 1] = data[offset + 3];
data[offset + rowSize] = data[offset + rowSize - 4];
data[offset + rowSize + 1] = data[offset + rowSize - 3];
data[offset + rowSize + 2] = data[offset + rowSize - 2];
data[offset + rowSize + 3] = data[offset + rowSize - 1];
offset -= imgRowSize;
}
}
// replacing queue items
fnArray.splice(j, count * 4, ['paintInlineImageXObjectGroup']);
argsArray.splice(j, count * 4,
[{width: imgWidth, height: imgHeight, data: imgData}, map]);
i = j;
ii = fnArray.length;
}
}
// grouping paintImageMaskXObject's into paintImageMaskXObjectGroup
// searching for (save, transform, paintImageMaskXObject, restore)+
var MIN_IMAGES_IN_MASKS_BLOCK = 10;
var MAX_IMAGES_IN_MASKS_BLOCK = 100;
for (var i = 0, ii = fnArray.length; i < ii; i++) {
if (fnArray[i] === 'paintImageMaskXObject' &&
fnArray[i - 2] === 'save' && fnArray[i - 1] === 'transform' &&
fnArray[i + 1] === 'restore') {
var j = i - 2;
for (i += 2; i < ii && fnArray[i - 4] === fnArray[i]; i++) {
}
var count = Math.min((i - j) >> 2,
MAX_IMAGES_IN_MASKS_BLOCK);
if (count < MIN_IMAGES_IN_MASKS_BLOCK) {
continue;
}
var images = [];
for (var q = 0; q < count; q++) {
var transform = argsArray[j + (q << 2) + 1];
var maskParams = argsArray[j + (q << 2) + 2];
images.push({data: maskParams[0], width: maskParams[2],
height: maskParams[3], transform: transform,
inverseDecode: maskParams[1]});
}
// replacing queue items
fnArray.splice(j, count * 4, ['paintImageMaskXObjectGroup']);
argsArray.splice(j, count * 4, [images]);
i = j;
ii = fnArray.length;
}
}
},
getTextContent: function PartialEvaluator_getTextContent(
stream, resources, state) {
var bidiTexts;