mirrors/pdf.js
1
0
Fork 0
mirror of https://github.com/mozilla/pdf.js.git synced 2025-04-26 10:08:06 +02:00
Code Activity

Text char codes extraction

Browse source
This commit is contained in:
notmasteryet 2011-12-10 17:24:54 -06:00
parent 853f16085f
commit 3b72c6063c
4 changed files with 152 additions and 13 deletions
Download patch file Download diff file Expand all files Collapse all files

82
src/evaluator.js
View file

@ -144,7 +144,7 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
fontRef = fontRef || fontRes.get(fontName);
var font = xref.fetchIfRef(fontRef);
assertWellFormed(isDict(font));
if (!font.translated) {
if (!font.loadedName) {
font.translated = self.translateFont(font, xref, resources,
dependency);
if (font.translated) {
@ -464,6 +464,65 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
};
},
getTextContent: function partialEvaluatorGetIRQueue(stream, resources) {
var self = this;
var xref = this.xref;
function handleSetFont(fontName, fontRef) {
var fontRes = resources.get('Font');
// TODO: TOASK: Is it possible to get here? If so, what does
// args[0].name should be like???
assert(fontRes, 'fontRes not available');
fontRes = xref.fetchIfRef(fontRes);
fontRef = fontRef || fontRes.get(fontName);
var font = xref.fetchIfRef(fontRef), tra;
assertWellFormed(isDict(font));
if (!font.translated) {
font.translated = self.translateFont(font, xref, resources);
}
return font;
}
resources = xref.fetchIfRef(resources) || new Dict();
var parser = new Parser(new Lexer(stream), false);
var res = resources;
var args = [], obj;
var text = '';
var font = null;
while (!isEOF(obj = parser.getObj())) {
if (isCmd(obj)) {
var cmd = obj.cmd;
switch (cmd) {
case 'Tf':
font = handleSetFont(args[0].name);
break;
case 'TJ':
var items = args[0];
for (var j = 0, jj = items.length; j < jj; j++) {
if (typeof items[j] === 'string')
text += items[j];
}
break;
case 'Tj':
text += args[0];
break;
} // switch
args = [];
} else if (obj != null) {
assertWellFormed(args.length <= 33, 'Too many arguments');
args.push(obj);
}
}
return text;
},
extractDataStructures: function
partialEvaluatorExtractDataStructures(dict, baseDict,
xref, properties) {
@ -837,15 +896,18 @@ var PartialEvaluator = (function PartialEvaluatorClosure() {
if (type.name === 'Type3') {
properties.coded = true;
var charProcs = xref.fetchIfRef(dict.get('CharProcs'));
var fontResources = xref.fetchIfRef(dict.get('Resources')) || resources;
properties.resources = fontResources;
properties.charProcIRQueues = {};
for (var key in charProcs.map) {
var glyphStream = xref.fetchIfRef(charProcs.map[key]);
var queueObj = {};
properties.charProcIRQueues[key] =
this.getIRQueue(glyphStream, fontResources, queueObj, dependency);
// read char procs only if dependency is specified
if (dependency) {
var charProcs = xref.fetchIfRef(dict.get('CharProcs'));
var fontResources = xref.fetchIfRef(dict.get('Resources')) || resources;
properties.resources = fontResources;
properties.charProcIRQueues = {};
for (var key in charProcs.map) {
var glyphStream = xref.fetchIfRef(charProcs.map[key]);
var queueObj = {};
properties.charProcIRQueues[key] =
this.getIRQueue(glyphStream, fontResources, queueObj, dependency);
}
}
}