only exclude links if overlap is over 50%, add test and comments

mozilla · Jan 8, 2025 · 5103798 · 5103798
1 parent 02ff41a
commit 5103798
Show file tree

Hide file tree

Showing 4 changed files with 8 additions and 2 deletions.
diff --git a/test/pdfs/.gitignore b/test/pdfs/.gitignore
@@ -692,3 +692,4 @@
 !issue19182.pdf
 !issue18911.pdf
 !issue19207.pdf
+!link.pdf
diff --git a/test/pdfs/link.pdf b/test/pdfs/link.pdf
diff --git a/web/annotation_layer_builder.js b/web/annotation_layer_builder.js
@@ -121,10 +121,15 @@ class AnnotationLayerBuilder {
 
     const uniqueLinks = linkAnnotations.filter(link => {
       for (const annotation of annotations) {
+        function area(rect) {
+          return Math.abs(rect[2] - rect[0]) * Math.abs(rect[3] - rect[1]);
+        }
+        const intersect = Util.intersect(annotation.rect, link.rect); // Find the intersection between the annotation and the link.
         if (
           annotation.subtype === "Link" &&
           annotation.url === link.url &&
-          Util.intersect(annotation.rect, link.rect) !== null
+          intersect !== null &&
+          area(intersect) / area(link.rect) > 0.5 // If the overlap is more than 50%.
         ) {
           return false;
         }

diff --git a/web/autolinker.js b/web/autolinker.js
@@ -3,7 +3,7 @@ import { getOriginalIndex, normalize } from "./pdf_find_controller.js";
 
 class Autolinker {
   static #urlRegex =
-    /\b(?:https?:\/\/|mailto:|www.)(?:[[\S--\[]--\p{P}]|\/|[\p{P}--\[]+[[\S--\[]--\p{P}])+/gmv;
+    /\b(?:https?:\/\/|mailto:|www.)(?:[[\S--\[]--\p{P}]|\/|[\p{P}--\[]+[[\S--\[]--\p{P}])+/gmv; // Regex can be tested and verified at https://regex101.com/r/zuMeQX/3.
 
   static #addLinkAnnotations(url, index, length, pdfPageView) {
     // TODO refactor out the logic for a single match from this function