added proper cache invalidation to load_bm25_corpus

python/ebook_search/api/bm25_tasks.py

@@ -8,7 +8,7 @@ from typing import TYPE_CHECKING
 
 from sqlalchemy.orm import Session
 
-from python.ebook_search.bm25_corpus import refresh_bm25_corpus
+from python.ebook_search.bm25_corpus import load_bm25_corpus, refresh_bm25_corpus
 
 if TYPE_CHECKING:
     from fastapi import FastAPI
@@ -56,3 +56,5 @@ def refresh_bm25_for_engine(engine: Engine, config: EbookSearchConfig) -> None:
     """Refresh the BM25 corpus using a SQLAlchemy engine."""
     with Session(engine) as session:
         refresh_bm25_corpus(session, config)
+    load_bm25_corpus.cache_clear()
+    logger.info("ebook_bm25_corpus_cache_cleared_after_refresh")

python/ebook_search/bm25_corpus.py

@@ -108,11 +108,10 @@ def refresh_bm25_corpus(
     )
     write_bm25_corpus(index_path, records, manifest)
     logger.info(
-        "ebook_bm25_index_refreshed path=%s chunks=%s created_at=%s note=%s",
+        "ebook_bm25_index_refreshed path=%s chunks=%s created_at=%s",
         index_path,
         manifest.chunk_count,
         manifest.created_at.isoformat(),
-        "restart_service_to_use_refreshed_bm25_cache",
     )
     return manifest
 
@@ -121,15 +120,10 @@ def refresh_bm25_corpus(
 def load_bm25_corpus(config: EbookSearchConfig) -> BM25Corpus:
     """Load the BM25 corpus into memory once per process.
 
-    This cache intentionally does not notice later on-disk corpus refreshes. Restart the service after rebuilding the
-    BM25 corpus for searches to use the new index.
+    Background refresh tasks clear this cache after rebuilding the on-disk corpus.
     """
     index_path = bm25_index_path(config)
-    logger.info(
-        "ebook_bm25_corpus_cache_load path=%s note=%s",
-        index_path,
-        "restart_service_after_bm25_refresh",
-    )
+    logger.info("ebook_bm25_corpus_cache_load path=%s", index_path)
     manifest = read_bm25_manifest(index_path)
     if manifest is None or not bm25_index_exists(index_path, manifest):
         msg = f"BM25 corpus is not available: {index_path}"

tests/test_ebook_search_ui.py

@@ -5,6 +5,7 @@ from __future__ import annotations
 from fastapi.testclient import TestClient
 from sqlalchemy import create_engine
 
+from python.ebook_search.api.bm25_tasks import refresh_bm25_for_engine
 from python.ebook_search.api.main import create_app
 from python.ebook_search.config import EbookSearchConfig, RerankConfig
 from python.ebook_search.embeddings import EmbeddingModelStats
@@ -232,6 +233,29 @@ def test_ui_scan_schedules_bm25_refresh_after_database_change(monkeypatch) -> No
     assert scheduled is True
 
 
+def test_bm25_refresh_clears_loaded_corpus_cache(monkeypatch) -> None:
+    refreshed: list[object] = []
+    cache_cleared = False
+
+    def fake_refresh_bm25_corpus(session, config):
+        refreshed.append((session, config))
+
+    def fake_cache_clear():
+        nonlocal cache_cleared
+        cache_cleared = True
+
+    monkeypatch.setattr("python.ebook_search.api.bm25_tasks.refresh_bm25_corpus", fake_refresh_bm25_corpus)
+    monkeypatch.setattr("python.ebook_search.api.bm25_tasks.load_bm25_corpus.cache_clear", fake_cache_clear)
+    engine = create_engine("sqlite+pysqlite:///:memory:", future=True)
+    config = EbookSearchConfig(rerank=RerankConfig(enabled=False))
+
+    refresh_bm25_for_engine(engine, config)
+
+    assert len(refreshed) == 1
+    assert refreshed[0][1] == config
+    assert cache_cleared is True
+
+
 def test_admin_page_shows_embedding_counts_by_model(monkeypatch) -> None:
     def fake_embedding_model_stats(_session):
         return [