Combine text items is off #8963
Comments
|
Closing as duplicate of #2140 |
|
I am trying to use a regex with the text of a PDF. Is it posible? |
It's possible. You need to a) preprocess text items based on their positions, or b) build regex without relying on the line breaks. |
|
You are right, the y-coordinate is the same for items on the same line. But I still can´t deal with the random spaces it adds splitting words in halves. Well, I guess I could disable combine text items and concatenate each symbol but... shouldn´t it be just as easy to fix directly in the combine text items implementation? |
|
Actually the random spaces are not solved by disableCombineTextItems. :( Update: That was because of weird symbols. The y-coordinate solution should work in general but it looks kind of easy to fix for everybody. Plus it would also solve #2140 |
|
Workaround: <!doctype html>
<html>
<head>
<meta charset="utf-8">
<title>PDF.js test</title>
<script src="https://mozilla.github.io/pdf.js/build/pdf.js"></script>
</head>
<body>
<input type="file" onchange="readText(this)" autofocus>
<pre id="text"></pre>
<script>
var pre = document.getElementById("text");
function readText(filePath) {
var reader = new FileReader();
var fileName = filePath.files[0].name;
reader.onload = function(event) {
PDFJS.getDocument(new Uint8Array(event.target.result)).then(function(pdf) {
var maxPages = pdf.pdfInfo.numPages;
var countPromises = [];
for (var i = 1; i <= maxPages; i++) {
countPromises.push(pdf.getPage(i).then(function(page){
return page.getTextContent().then(function(text){
"use strict";
let lastY, str = '';
for (let item of text.items) {
if (lastY == item.transform[5] || !lastY)
str += item.str;
else
str += '\n' + item.str;
lastY = item.transform[5];
}
return str;
});
}));
}
Promise.all(countPromises).then(function(texts) {
pre.innerHTML = texts.join('\n');
});
});
};
reader.readAsArrayBuffer(filePath.files[0]);
}
</script>
</body>
</html> |
|
Even though the "duplicate" is fixed, this issue still isn't. Since the API has changed in the last year, this is now the code to test it: <!doctype html>
<html>
<head>
<meta charset="utf-8">
<title>PDF.js test</title>
<script src="https://unpkg.com/pdfjs-dist/build/pdf.min.js"></script>
</head>
<body>
<input type="file" onchange="readText(this)" autofocus>
<pre id="text"></pre>
<script>
var pre = document.getElementById("text");
function readText(filePath) {
var reader = new FileReader();
var fileName = filePath.files[0].name;
reader.onload = function(event) {
pdfjsLib.getDocument(event.target.result).promise.then(function(pdf) {
var maxPages = pdf.numPages;
var countPromises = [];
for (var i = 1; i <= maxPages; i++) {
countPromises.push(pdf.getPage(i).then(function(page){
return page.getTextContent().then(function(text){
return text.items.map(function(item) { return item.str; }).join('\n');
});
}));
}
Promise.all(countPromises).then(function(texts) {
pre.innerHTML = texts.join('\n');
});
});
};
reader.readAsArrayBuffer(filePath.files[0]);
}
</script>
</body>
</html> |
|
After all this time, the issue is still not fixed. This is the updated code to test it: <!doctype html>
<html>
<head>
<meta charset="utf-8">
<title>PDF.js test</title>
<meta name="viewport" content="width=device-width, initial-scale=1">
<script src="https://unpkg.com/pdfjs-dist/build/pdf.min.js"></script>
</head>
<body>
<input type="file" onchange="readText(this)" autofocus>
<pre id="text"></pre>
<script>
"use strict";
let pre = document.getElementById("text");
function readText(filePath) {
let reader = new FileReader();
reader.onload = async function(event) {
let pdf = await pdfjsLib.getDocument(event.target.result).promise;
let pages = Array.from({length: pdf.numPages}, (_, i) => i + 1)
.map(index => pdf.getPage(index).then(page => page.getTextContent())
.then(text => text.items.map(item => item.str).join('\n')));
pre.innerHTML = (await Promise.all(pages)).join('\n');
};
reader.readAsArrayBuffer(filePath.files[0]);
}
</script>
</body>
</html> |
Link to PDF file: Hyphenator.pdf
Configuration:
Steps to reproduce the problem:
What is the expected behavior?
A plain text version of the original PDF.
What went wrong?
Text from the same line is not combined properly.
The text was updated successfully, but these errors were encountered: