1
0
Fork 0
mirror of https://github.com/mozilla/pdf.js.git synced 2025-04-26 01:58:06 +02:00

Text char codes extraction

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

View file

@ -200,10 +200,12 @@ var Page = (function PageClosure() {
if (isArray(content)) {
// fetching items
var i, n = content.length;
var streams = [];
for (i = 0; i < n; ++i)
content[i] = xref.fetchIfRef(content[i]);
content = new StreamsSequenceStream(content);
}
streams.push(xref.fetchIfRef(content[i]));
content = new StreamsSequenceStream(streams);
} else if (isStream(content))
content.pos = 0;
var pe = this.pe = new PartialEvaluator(
xref, handler, 'p' + this.pageNumber + '_');
@ -212,6 +214,36 @@ var Page = (function PageClosure() {
dependency));
},
extractTextContent: function pageExtractPageContent() {
if ('textContent' in this) {
// text content was extracted
return this.textContent;
}
var handler = {
on: function () {},
send: function() {}
};
var xref = this.xref;
var content = xref.fetchIfRef(this.content);
var resources = xref.fetchIfRef(this.resources);
if (isArray(content)) {
// fetching items
var i, n = content.length;
var streams = [];
for (i = 0; i < n; ++i)
streams.push(xref.fetchIfRef(content[i]));
content = new StreamsSequenceStream(streams);
} else if (isStream(content))
content.pos = 0;
var pe = new PartialEvaluator(
xref, handler, 'p' + this.pageNumber + '_');
var text = pe.getTextContent(content, resources);
return (this.textContent = text);
},
ensureFonts: function pageEnsureFonts(fonts, callback) {
// Convert the font names to the corresponding font obj.
for (var i = 0, ii = fonts.length; i < ii; i++) {
@ -614,6 +646,12 @@ var PDFDoc = (function PDFDocClosure() {
throw data.error;
}, this);
messageHandler.on('text_extracted', function pdfDocError(data) {
var index = data.index;
if (this.textExtracted)
this.textExtracted(index);
}, this);
setTimeout(function pdfDocFontReadySetTimeout() {
messageHandler.send('doc', this.data);
this.workerReadyPromise.resolve(true);
@ -643,6 +681,12 @@ var PDFDoc = (function PDFDocClosure() {
return (this.pageCache[n] = page);
},
extractText: function pdfDocExtractExtractText() {
this.workerReadyPromise.then(function pdfDocStartRenderingThen() {
this.messageHandler.send('extract_text');
}.bind(this));
},
destroy: function pdfDocDestroy() {
if (this.worker)
this.worker.terminate();