聊天和回忆录证据检索都走 pgvector，去掉 Postgres FTS/content_tsv，新迁移删掉 content_tsv 列（部署要先 alembic upgrade）。

Embedding 端口增加 is_available()，聊天和回忆录日志用统一方式表示向量是否真能调用。

记忆整理（compaction）支持 Beat 定期扫用户；

事实抽取提示与 subject 归一化，减少同一人多种称呼；

api/app/features/memory/retriever.py

@@ -1,31 +1,17 @@
-"""Hybrid retriever — metadata filter + FTS + vector retrieval + score fusion."""
+"""Hybrid retriever — 向量检索 + 元数据证据包。"""
 
 from sqlalchemy.ext.asyncio import AsyncSession
 
+from app.core.logging import get_logger
 from app.features.memory.evidence import retrieve_evidence_bundle_async
-from app.features.memory.repo import search_chunks_fts, search_chunks_vector
+from app.features.memory.repo import search_chunks_vector
 from app.ports.embedding import EmbeddingProvider
 
-
-def _rrf_merge(
-    fts_items: list[dict], vector_items: list[dict], k: int = 60
-) -> list[dict]:
-    """Reciprocal Rank Fusion. Merge FTS and vector results by id."""
-    scores: dict[str, float] = {}
-    for rank, item in enumerate(fts_items):
-        cid = item["id"]
-        scores[cid] = scores.get(cid, 0) + 1 / (k + rank + 1)
-    for rank, item in enumerate(vector_items):
-        cid = item["id"]
-        scores[cid] = scores.get(cid, 0) + 1 / (k + rank + 1)
-
-    all_items = {x["id"]: x for x in fts_items + vector_items}
-    sorted_ids = sorted(scores.keys(), key=lambda i: scores[i], reverse=True)
-    return [all_items[i] for i in sorted_ids]
+logger = get_logger(__name__)
 
 
 class HybridRetriever:
-    """Combine FTS, vector, and metadata filter into evidence bundle."""
+    """向量 chunk 检索 + facts/timeline/summaries/stories。"""
 
     def __init__(
         self,
@@ -51,27 +37,27 @@ class HybridRetriever:
             )
 
         q = query.strip()
-        fts_chunks = await search_chunks_fts(
-            self._db, user_id=user_id, query=query, limit=top_k * 2
-        )
-
-        vector_chunks: list[dict] = []
-        if self._embedding and q:
+        merged_chunk_dicts: list[dict] = []
+        if self._embedding:
             q_emb = await self._embedding.embed_text(q)
             if q_emb:
-                vector_chunks = await search_chunks_vector(
-                    self._db, user_id=user_id, query_embedding=q_emb, limit=top_k * 2
+                vector_rows = await search_chunks_vector(
+                    self._db, user_id, q_emb, limit=top_k
                 )
-
-        merged = _rrf_merge(fts_chunks, vector_chunks)[:top_k]
-        merged_chunk_dicts = [
-            {
-                "id": c["id"],
-                "content": c["content"],
-                "chunk_index": c.get("chunk_index", 0),
-            }
-            for c in merged
-        ]
+                merged_chunk_dicts = [
+                    {
+                        "id": c["id"],
+                        "content": c["content"],
+                        "chunk_index": c.get("chunk_index", 0),
+                    }
+                    for c in vector_rows
+                ]
+            else:
+                logger.warning(
+                    "HybridRetriever empty_query_embedding user_id={}", user_id
+                )
+        else:
+            logger.warning("HybridRetriever no_embedding_provider user_id={}", user_id)
 
         return await retrieve_evidence_bundle_async(
             self._db,