From 7a8f261ec50b9c15bc37ba8ae1f0022c392111d3 Mon Sep 17 00:00:00 2001
From: Igor Soarez <igor@soarez.org>
Date: Mon, 3 Aug 2026 22:09:48 +0100
Subject: Take Brave snippets by position rather than by class name

Brave's source name was leaking into every snippet ("Medium March 27,
2025 - ..."), because the .sitename/.netloc selectors guessed for it match
nothing. Rather than guess again, use the ordering: an engine rendering
"source / breadcrumb / title / description" puts all its metadata before
the title, so everything after the title line is the description. Class
names churn; that ordering does not. The named selectors stay as the
fallback for when the title is not on a line of its own.

Verified against Brave, and Google/DuckDuckGo/Bing re-checked for
regressions.
---
 src/engines.ts |  8 +++++---
 src/extract.ts | 45 ++++++++++++++++++++++++++++++++-------------
 2 files changed, 37 insertions(+), 16 deletions(-)

diff --git a/src/engines.ts b/src/engines.ts
index 79462fc..a71b9c8 100644
--- a/src/engines.ts
+++ b/src/engines.ts
@@ -190,9 +190,11 @@ const BRAVE: EngineDefinition = {
     item: '.snippet[data-type="web"]',
     link: "a[href]",
     title: ".title",
-    // No dedicated description element; the item's text is
-    // "source | breadcrumb | title | description", so subtract the first three.
-    subtract: [".title", "cite", ".sitename", ".netloc"],
+    // No dedicated description element. The item's text is
+    // "source | breadcrumb | title | description", so the extractor keeps only
+    // the lines after the title. These selectors are the fallback for when the
+    // title is not on a line of its own.
+    subtract: [".title", "cite"],
     unwrap: "none",
     selfHostPattern: String.raw`^https?://(search\.)?brave\.com/`,
   },
diff --git a/src/extract.ts b/src/extract.ts
index 282fa07..8b5c57f 100644
--- a/src/extract.ts
+++ b/src/extract.ts
@@ -177,21 +177,40 @@ export const buildProbeScript = (config: ExtractionConfig): string => String.raw
       if (snippetEl) {
         snippet = norm(snippetEl.innerText);
       } else {
-        // No dedicated description element: subtract the parts we can name
-        // (title, breadcrumb, source) and keep what is left.
-        var drop = {};
-        for (var s = 0; CFG.subtract && s < CFG.subtract.length; s++) {
-          var parts = item.querySelectorAll(CFG.subtract[s]);
-          for (var p = 0; p < parts.length; p++) {
-            var pl = linesOf(parts[p]);
-            for (var q = 0; q < pl.length; q++) drop[norm(pl[q])] = true;
+        // No dedicated description element. Rather than guess at class names
+        // for the source and breadcrumb, use their position: an engine that
+        // renders "source / breadcrumb / title / description" puts every piece
+        // of metadata *before* the title, so everything after the title line is
+        // the description. Class names churn; that ordering does not.
+        var itemLines = linesOf(item);
+        var titleIndex = -1;
+        for (var t = 0; t < itemLines.length; t++) {
+          if (norm(itemLines[t]) === title) { titleIndex = t; break; }
+        }
+
+        var candidate;
+        if (titleIndex >= 0) {
+          candidate = itemLines.slice(titleIndex + 1);
+        } else {
+          // Title is not its own line (it may be inline with other text).
+          // Fall back to subtracting the parts we can name.
+          var drop = {};
+          for (var s = 0; CFG.subtract && s < CFG.subtract.length; s++) {
+            var parts = item.querySelectorAll(CFG.subtract[s]);
+            for (var p = 0; p < parts.length; p++) {
+              var pl = linesOf(parts[p]);
+              for (var q = 0; q < pl.length; q++) drop[norm(pl[q])] = true;
+            }
           }
+          candidate = itemLines.filter(function (l) { return !drop[norm(l)]; });
         }
-        var kept = linesOf(item).filter(function (l) {
-          var n = norm(l);
-          return !drop[n] && n !== title && !/^https?:\/\//.test(n) && n.indexOf("›") === -1;
-        });
-        snippet = norm(kept.join(" "));
+
+        snippet = norm(
+          candidate.filter(function (l) {
+            var n = norm(l);
+            return n !== title && !/^https?:\/\//.test(n) && n.indexOf("›") === -1;
+          }).join(" "),
+        );
       }
 
       seen[url] = true;
-- 
cgit v1.3.1