From 1c57845ebd174729709c09131ee1843cbe3f7f48 Mon Sep 17 00:00:00 2001
From: Kyle Isom <kyle@imap.cc>
Date: Wed, 23 Sep 2026 16:04:45 -0700
Subject: [PATCH] p3: keep original URLs on RRF/dedupe; rerank=false is not a
 merge

CanonicalURL is the dedupe key only. Hits keep the first source's
original URL, including a trailing slash. Dedupe has the same rule.
---
 docs/P3.md                     | 16 ++++++++++------
 internal/search/engine_test.go | 20 ++++++++++++++++++++
 internal/search/rank.go        |  6 ++++--
 internal/search/types.go       |  3 ++-
 4 files changed, 36 insertions(+), 9 deletions(-)

diff --git a/docs/P3.md b/docs/P3.md
index e9a1ced..dc8a273 100644
--- a/docs/P3.md
+++ b/docs/P3.md
@@ -146,10 +146,10 @@ placeholder line the same way (empty, on-box fill).
 3. Fan-in: one goroutine per enabled source, each with `search.timeout`. Collect whatever
    arrived; a timeout or non-2xx is a `warnings` entry (`searxng: timeout`, `tavily: 429`)
    and that source contributes nothing.
-4. Dedupe on **canonical URL**: scheme+host lowercased, default ports stripped, fragment
-   dropped, trailing slash dropped on the path, `www.` left as-is (different sites). First
-   snippet wins; `source` becomes the first source that produced it; later sources only
-   affect rerank input if they brought a longer snippet (keep the longer one).
+4. Dedupe on **canonical URL as a key only** (scheme+host lowercased, default ports
+   stripped, fragment dropped, trailing slash dropped on the path; `www.` left as-is).
+   The hit returned to the agent keeps the **first source's original URL**. First
+   source wins attribution; a later longer snippet replaces the text only.
 5. Merge: reciprocal-rank fusion across source lists unless `rerank: false`.
    A source with `no_rerank = true` keeps its own order as a prefix (first-source
    wins on URL). Cosine/honcho-embed is not used for search.
@@ -165,8 +165,12 @@ Fan-in uses reciprocal-rank fusion (`1/(60+rank)`) over per-source lists. Search
 does not call `honcho-embed`; Dixie saturation cannot fail a search.
 
 `[search.tavily] no_rerank = true` keeps Tavily's own order at the front of the
-result list; remaining sources still RRF. `rerank = false` skips RRF and keeps
-source-fan-in order (SearXNG, then Brave, then Tavily).
+result list; remaining sources still RRF.
+
+**`rerank = false` is not a merge.** It concatenates in source order (SearXNG,
+then Brave, then Tavily) and dedupes. Do not enable two sources with
+`rerank = false` expecting RRF or Tavily-first. That only happens with
+`rerank = true` (and `no_rerank` on Tavily if you want Tavily's order).
 
 ## 6. Guard boundary (SSRF)
 
diff --git a/internal/search/engine_test.go b/internal/search/engine_test.go
index cabdc2d..4040f97 100644
--- a/internal/search/engine_test.go
+++ b/internal/search/engine_test.go
@@ -157,6 +157,23 @@ func TestRRFInterleaveAndNoRerank(t *testing.T) {
 	}
 }
 
+func TestRRFKeepsOriginalURL(t *testing.T) {
+	orig := "https://ex.com/docs/"
+	merged := rrfMerge([][]Hit{
+		{{URL: orig, Title: "slash", Source: NameSearxNG}},
+		{{URL: "https://ex.com/docs", Title: "stripped", Snippet: "longer", Source: NameTavily}},
+	})
+	if len(merged) != 1 {
+		t.Fatalf("%+v", merged)
+	}
+	if merged[0].URL != orig {
+		t.Fatalf("want original URL %q got %q", orig, merged[0].URL)
+	}
+	if CanonicalURL(orig) != "https://ex.com/docs" {
+		t.Fatalf("canonical %q", CanonicalURL(orig))
+	}
+}
+
 func TestDailyCapSkipsTavily(t *testing.T) {
 	db, err := dcache.Open(filepath.Join(t.TempDir(), "c.db"), 1<<20, nil)
 	if err != nil {
@@ -187,4 +204,7 @@ func TestDedupeLongerSnippet(t *testing.T) {
 	if len(out) != 1 || out[0].Source != NameSearxNG || out[0].Snippet != "aabbcc" {
 		t.Fatalf("%+v", out)
 	}
+	if out[0].URL != "https://Ex.com/a#x" {
+		t.Fatalf("dedupe must keep first original URL, got %q", out[0].URL)
+	}
 }
diff --git a/internal/search/rank.go b/internal/search/rank.go
index c939328..8cea60c 100644
--- a/internal/search/rank.go
+++ b/internal/search/rank.go
@@ -97,18 +97,20 @@ func rrfMerge(lists [][]Hit) []Hit {
 	var order []string
 	for _, list := range lists {
 		for r, h := range list {
-			c := CanonicalURL(h.URL)
+			orig := h.URL
+			c := CanonicalURL(orig)
 			if c == "" {
 				continue
 			}
-			h.URL = c
 			sc := 1.0 / float64(rrfK+r+1)
 			if a, ok := seen[c]; ok {
 				a.s += sc
 				if len(h.Snippet) > len(a.h.Snippet) {
 					src := a.h.Source
+					keep := a.h.URL
 					a.h = h
 					a.h.Source = src
+					a.h.URL = keep
 				}
 				continue
 			}
diff --git a/internal/search/types.go b/internal/search/types.go
index 5648104..03b3379 100644
--- a/internal/search/types.go
+++ b/internal/search/types.go
@@ -79,13 +79,14 @@ func Dedupe(in []Hit) []Hit {
 		if canon == "" {
 			continue
 		}
-		h.URL = canon
 		if s, ok := seen[canon]; ok {
 			if len(h.Snippet) > len(s.h.Snippet) {
 				src := s.h.Source
+				keep := s.h.URL
 				sc := s.h.Score
 				s.h = h
 				s.h.Source = src
+				s.h.URL = keep
 				if sc > s.h.Score {
 					s.h.Score = sc
 				}