From 6ca089286c14e247b70a50d83b3fa278099add79 Mon Sep 17 00:00:00 2001
From: altair823 <dlsrks0734@gmail.com>
Date: Sun, 3 May 2026 06:25:13 +0000
Subject: [PATCH] =?UTF-8?q?review(p9-fb-18):=20=ED=9A=8C=EC=B0=A8=201=20ni?=
 =?UTF-8?q?t=20=EB=B0=98=EC=98=81?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- `App::build_retriever(mode) -> Result<Arc<dyn Retriever>>` 추출.
  `ask` 와 `ask_with_session` 모두 사용. 35+ 줄 retriever stack 중복
  제거 — 미래 retriever 변경이 한 곳만.
- V005 migration `chat_sessions.sql` 의 `citations_json` doc 수정:
  `Vec<Citation>` → `Vec<AnswerCitation>` (실제 stored type 과 일치).
  AnswerCitation 가 marker + Citation 등 포함하므로 deserialize 시
  type mismatch 회피.

15 app lib + 9 store chat_sessions + clippy 통과.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---
 crates/kebab-app/src/app.rs        | 65 +++++++++---------------------
 migrations/V005__chat_sessions.sql |  7 ++--
 2 files changed, 22 insertions(+), 50 deletions(-)

diff --git a/crates/kebab-app/src/app.rs b/crates/kebab-app/src/app.rs
index e1cc135..e538efe 100644
--- a/crates/kebab-app/src/app.rs
+++ b/crates/kebab-app/src/app.rs
@@ -254,7 +254,20 @@ impl App {
     /// Run a RAG `ask` against the configured retriever + LLM. Reuses
     /// the memoized embedder / vector / LLM where applicable.
     pub fn ask(&self, query: &str, opts: AskOpts) -> Result<Answer> {
-        let retriever: Arc<dyn Retriever> = match opts.mode {
+        let retriever = self.build_retriever(opts.mode)?;
+        let llm = self.llm()?;
+        let pipeline =
+            RagPipeline::new(self.config.clone(), retriever, llm, self.sqlite.clone());
+        pipeline.ask(query, opts)
+    }
+
+    /// p9-fb-18: shared retriever-stack builder used by [`Self::ask`]
+    /// and [`Self::ask_with_session`]. Lexical mode uses the FTS5
+    /// retriever directly; vector / hybrid require embeddings (and
+    /// surface the same "switch to --mode lexical" error from
+    /// [`Self::require_embeddings`] when disabled).
+    fn build_retriever(&self, mode: SearchMode) -> Result<Arc<dyn Retriever>> {
+        Ok(match mode {
             SearchMode::Lexical => Arc::new(LexicalRetriever::with_settings(
                 self.sqlite.clone(),
                 lexical_index_version(&self.config),
@@ -292,12 +305,7 @@ impl App {
                 )) as Arc<dyn Retriever>;
                 Arc::new(HybridRetriever::new(&self.config, lex, vec_retr))
             }
-        };
-
-        let llm = self.llm()?;
-        let pipeline =
-            RagPipeline::new(self.config.clone(), retriever, llm, self.sqlite.clone());
-        pipeline.ask(query, opts)
+        })
     }
 
     /// p9-fb-18: ask under a persistent chat session. Loads the
@@ -353,46 +361,9 @@ impl App {
             })
             .collect();
 
-        // Build the retriever stack the same way `ask` does.
-        let retriever: Arc<dyn Retriever> = match opts.mode {
-            SearchMode::Lexical => Arc::new(LexicalRetriever::with_settings(
-                self.sqlite.clone(),
-                lexical_index_version(&self.config),
-                self.config.search.snippet_chars,
-            )),
-            SearchMode::Vector => {
-                let (emb, vec_store) = self.require_embeddings()?;
-                let vec_iv = vector_index_version(emb.as_ref());
-                let vec_dyn: Arc<dyn VectorStore + Send + Sync> = vec_store;
-                let emb_dyn: Arc<dyn Embedder> = emb;
-                Arc::new(VectorRetriever::with_settings(
-                    vec_dyn,
-                    emb_dyn,
-                    self.sqlite.clone(),
-                    vec_iv,
-                    self.config.search.snippet_chars,
-                ))
-            }
-            SearchMode::Hybrid => {
-                let lex = Arc::new(LexicalRetriever::with_settings(
-                    self.sqlite.clone(),
-                    lexical_index_version(&self.config),
-                    self.config.search.snippet_chars,
-                )) as Arc<dyn Retriever>;
-                let (emb, vec_store) = self.require_embeddings()?;
-                let vec_iv = vector_index_version(emb.as_ref());
-                let vec_dyn: Arc<dyn VectorStore + Send + Sync> = vec_store;
-                let emb_dyn: Arc<dyn Embedder> = emb;
-                let vec_retr = Arc::new(VectorRetriever::with_settings(
-                    vec_dyn,
-                    emb_dyn,
-                    self.sqlite.clone(),
-                    vec_iv,
-                    self.config.search.snippet_chars,
-                )) as Arc<dyn Retriever>;
-                Arc::new(HybridRetriever::new(&self.config, lex, vec_retr))
-            }
-        };
+        // p9-fb-18 R1: shared retriever builder removes the prior
+        // copy of `ask`'s 35-line stack — see [`Self::build_retriever`].
+        let retriever = self.build_retriever(opts.mode)?;
         let llm = self.llm()?;
         let pipeline =
             RagPipeline::new(self.config.clone(), retriever, llm, self.sqlite.clone());
diff --git a/migrations/V005__chat_sessions.sql b/migrations/V005__chat_sessions.sql
index 080d442..9fc9403 100644
--- a/migrations/V005__chat_sessions.sql
+++ b/migrations/V005__chat_sessions.sql
@@ -25,9 +25,10 @@
 --   max_context_tokens that produced the session so a retroactive
 --   answer-quality regression can be re-traced.
 --
--- * `citations_json` carries `Vec<Citation>` so the answer can be
---   redisplayed with the same citation markers a future session
---   sees on resume.
+-- * `citations_json` carries `Vec<AnswerCitation>` (per p9-fb-18) —
+--   each AnswerCitation holds a `Citation` plus `marker`, so the
+--   answer can be redisplayed with the same citation markers a
+--   future session sees on resume.
 --
 -- * `INTEGER` timestamps (unix epoch seconds) — same convention the
 --   rest of the schema uses (P1-7 baselines this).