Add API for extractor scripts to query PDF links in a specific sub rect

autotests/pdfdocumenttest.cpp

@@ -86,10 +86,17 @@ private Q_SLOTS:
         QCOMPARE(doc->pageCount(), 1);
         const auto page = doc->page(0);
         QCOMPARE(page.linkCount(), 1);
-        const auto link = page.link(0);
+        auto link = page.link(0);
         QCOMPARE(link.url(), QLatin1String("https://kde.org"));
         qDebug() << link.area();
         QVERIFY(link.area().isValid());
+
+        QCOMPARE(page.linksInRect(0.0, 0.0, 1.0, 1.0).size(), 1);
+        QCOMPARE(page.linksInRect(0.0, 0.0, 1.0, 0.5).size(), 1);
+        link = page.linksInRect(0.0, 0.0, 0.5, 0.5).at(0).value<PdfLink>();
+        QCOMPARE(link.url(), QLatin1String("https://kde.org"));
+        QCOMPARE(page.linksInRect(0.0, 0.0, 0.2, 0.5).size(), 1);
+        QCOMPARE(page.linksInRect(0.0, 0.5, 1.0, 1.0).size(), 0);
     }
 
     void testInvalidPdfDocument()

src/lib/pdf/pdfdocument.cpp

@@ -175,6 +175,31 @@ QVariantList PdfPage::linksVariant() const
     return l;
 }
 
+QVariantList PdfPage::linksInRect(double left, double top, double right, double bottom) const
+{
+    QRectF bbox(QPointF(left, top), QPointF(right, bottom));
+    d->load();
+
+    QVariantList l;
+    for (const auto &link : d->m_links) {
+        if (!link.area().intersects(bbox)) {
+            continue;
+        }
+        l.push_back(QVariant::fromValue(link));
+    }
+
+    std::sort(l.begin(), l.end(), [](const auto &lhs, const auto &rhs) {
+        const auto lhsLink = lhs.template value<PdfLink>();
+        const auto rhsLink = rhs.template value<PdfLink>();
+        if (lhsLink.area().top() == rhsLink.area().top()) {
+            return lhsLink.area().left() < rhsLink.area().left();
+        }
+        return lhsLink.area().top() < rhsLink.area().top();
+    });
+
+    return l;
+}
+
 
 PdfDocument::PdfDocument(QObject *parent)
     : QObject(parent)

src/lib/pdf/pdfdocument.h

@@ -62,6 +62,12 @@ public:
     /** The n-th link found in this document. */
     PdfLink link(int index) const;
 
+    /** Returns all links in the specified sub-rect of this page.
+     *  All parameters are relative values between @c 0 and @c 1 of the entire page size.
+     *  Links are sorted top to bottom / left to right based on their top left corner.
+     */
+    Q_INVOKABLE QVariantList linksInRect(double left, double top, double right, double bottom) const;
+
 private:
     QVariantList imagesVariant() const;
     QVariantList linksVariant() const;