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

Attempt to improve the detection of (metric) page names, by fuzzy matching the dimensions

Browse Source 
The following is an example of a PDF file (the specification to be exact) where the page size (compare the size displayed in Adobe Reader), and the page name, isn't correctly displayed without the patch: https://www.adobe.com/content/dam/acom/en/devnet/acrobat/pdfs/PDF32000_2008.pdf
This commit is contained in:
Jonas Jenwald 2018-03-21 13:39:07 +01:00
parent d86b816c2b
commit b004308f88
1 changed files with 35 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
web/pdf_document_properties.js
View File

@ -287,19 +287,50 @@ class PDFDocumentProperties {
} }
const isPortrait = isPortraitOrientation(pageSizeInches); const isPortrait = isPortraitOrientation(pageSizeInches);
const sizeInches = { let sizeInches = {
width: Math.round(pageSizeInches.width * 100) / 100, width: Math.round(pageSizeInches.width * 100) / 100,
height: Math.round(pageSizeInches.height * 100) / 100, height: Math.round(pageSizeInches.height * 100) / 100,
}; };
// 1in == 25.4mm; no need to round to 2 decimals for millimeters. // 1in == 25.4mm; no need to round to 2 decimals for millimeters.
const sizeMillimeters = { let sizeMillimeters = {
width: Math.round(pageSizeInches.width * 25.4 * 10) / 10, width: Math.round(pageSizeInches.width * 25.4 * 10) / 10,
height: Math.round(pageSizeInches.height * 25.4 * 10) / 10, height: Math.round(pageSizeInches.height * 25.4 * 10) / 10,
}; };
let pageName = null; let pageName = null;
const name = getPageName(sizeInches, isPortrait, US_PAGE_NAMES) || let name = getPageName(sizeInches, isPortrait, US_PAGE_NAMES) ||
getPageName(sizeMillimeters, isPortrait, METRIC_PAGE_NAMES); getPageName(sizeMillimeters, isPortrait, METRIC_PAGE_NAMES);
if (!name && !(Number.isInteger(sizeMillimeters.width) &&
Number.isInteger(sizeMillimeters.height))) {
// Attempt to improve the page name detection by falling back to fuzzy
// matching of the metric dimensions, to account for e.g. rounding errors
// and/or PDF files that define the page sizes in an imprecise manner.
const exactMillimeters = {
width: pageSizeInches.width * 25.4,
height: pageSizeInches.height * 25.4,
};
const intMillimeters = {
width: Math.round(sizeMillimeters.width),
height: Math.round(sizeMillimeters.height),
};
// Try to avoid false positives, by only considering "small" differences.
if (Math.abs(exactMillimeters.width - intMillimeters.width) < 0.1 &&
Math.abs(exactMillimeters.height - intMillimeters.height) < 0.1) {
name = getPageName(intMillimeters, isPortrait, METRIC_PAGE_NAMES);
if (name) {
// Update *both* sizes, computed above, to ensure that the displayed
// dimensions always correspond to the detected page name.
sizeInches = {
width: Math.round(intMillimeters.width / 25.4 * 100) / 100,
height: Math.round(intMillimeters.height / 25.4 * 100) / 100,
};
sizeMillimeters = intMillimeters;
}
}
}
if (name) { if (name) {
pageName = this.l10n.get('document_properties_page_size_name_' + pageName = this.l10n.get('document_properties_page_size_name_' +
name.toLowerCase(), null, name); name.toLowerCase(), null, name);