Re-embed the knowledge base when the embedding model changes

Vectors from two different embedding models are not comparable — Titan v2 has
1024 dimensions where nomic-embed-text has 768 — and _cosine_similarity zips
them, so a mismatch produced quietly meaningless scores rather than an error.
Nothing forced a re-embed either: sync_local_kb only looked at the content
hash, so switching models left every cached vector in place.

local_knowledge_docs now records the model each vector came from, and a row is
re-embedded when its content hash OR its model differs. Saving a new
LOCAL_KB_EMBEDDING_MODEL from the settings panel starts a background pass over
every file; the knowledge base panel also has a button for it. While that pass
runs, the per-request sync steps aside so no file is embedded twice.

The panel now shows what the embedding model is doing: model name, vector
size, how many files are embedded with the current model, how many are still
on an older one, a progress bar while a pass runs, and the last error. A
"接続テスト" button embeds one short string so a broken or unreachable model
can be told apart from one that is merely slow to warm up.

Also read TAVILY_API_KEY from os.environ at call time. It was imported as a
config constant, which is evaluated before load_into_env() runs, so a key
saved from the settings panel was stored but never used — not even after a
restart.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

backend/app/api/settings.py

@@ -1,9 +1,12 @@
+import os
+
 from fastapi import APIRouter, Depends, HTTPException
 from pydantic import BaseModel
 from sqlalchemy.orm import Session as DBSession
 
 from app.core import settings_store
 from app.core.auth import require_settings_password
+from app.core.memory_kb import local_knowledge
 from app.db.database import get_db
 
 router = APIRouter(prefix="/api/settings", tags=["settings"], dependencies=[Depends(require_settings_password)])
@@ -29,7 +32,8 @@ def list_settings(db: DBSession = Depends(get_db)):
 
 
 @router.put("", response_model=list[SettingEntry])
-def save_settings(req: SaveSettingsRequest, db: DBSession = Depends(get_db)):
+async def save_settings(req: SaveSettingsRequest, db: DBSession = Depends(get_db)):
+    before_embedding_model = os.environ.get("LOCAL_KB_EMBEDDING_MODEL")
     try:
         for key, value in req.values.items():
             if not value:
@@ -37,13 +41,29 @@ def save_settings(req: SaveSettingsRequest, db: DBSession = Depends(get_db)):
             settings_store.save(db, key, value)
     except ValueError as e:
         raise HTTPException(status_code=400, detail=str(e)) from e
+
+    _reembed_if_model_changed(before_embedding_model)
     return settings_store.list_masked(db)
 
 
 @router.delete("/{key}", response_model=list[SettingEntry])
-def clear_setting(key: str, db: DBSession = Depends(get_db)):
+async def clear_setting(key: str, db: DBSession = Depends(get_db)):
+    before_embedding_model = os.environ.get("LOCAL_KB_EMBEDDING_MODEL")
     try:
         settings_store.clear(db, key)
     except ValueError as e:
         raise HTTPException(status_code=400, detail=str(e)) from e
+
+    _reembed_if_model_changed(before_embedding_model)
     return settings_store.list_masked(db)
+
+
+def _reembed_if_model_changed(before: str | None) -> None:
+    """埋め込みモデルが変わったら、ローカルKBの再埋め込みをバックグラウンドで走らせる。
+
+    モデルが違うと次元数も意味空間も変わり、古いベクトルは比較しても無意味なので、
+    設定を変えた時点で全件やり直す（完了を待たずに応答は返す）。
+    """
+    if os.environ.get("LOCAL_KB_EMBEDDING_MODEL") == before:
+        return
+    local_knowledge.start_reembed_all()