fix(ebook-search): skip comment lines in gold query loader and realign tests

load_gold_queries now skips blank and `//` comment lines so the committed
section separator in queries.jsonl no longer breaks dataset/load-test loading.

Update tests left stale by the search refactor (6bc3011):
- pass the now-required rank_constant to reciprocal_rank_fusion
- expect bm25_candidates to receive the full query and drop the removed
  "BM25 query preparation" timing step
- assert reranking is enabled by default

python/ebook_search/eval/dataset.py

@@ -28,11 +28,11 @@ class GoldQuery:
 
 
 def load_gold_queries(path: Path = DEFAULT_QUERIES_PATH) -> list[GoldQuery]:
-    """Load labeled queries from a JSONL file."""
+    """Load labeled queries from a JSONL file. Blank lines and ``//`` comment lines are skipped."""
     queries: list[GoldQuery] = []
     for line in path.read_text(encoding="utf-8").splitlines():
         stripped = line.strip()
-        if not stripped:
+        if not stripped or stripped.startswith("//"):
             continue
         record = json.loads(stripped)
         queries.append(

tests/ebook_search/test_core.py

@@ -68,7 +68,7 @@ def test_reciprocal_rank_fusion_combines_vector_and_bm25_rankings() -> None:
         SearchResult(chunk_id=3, text="c", source_title="C", score=2.1, bm25_score=2.1),
     ]
 
-    fused = reciprocal_rank_fusion(vector_results, lexical_results)
+    fused = reciprocal_rank_fusion(vector_results, lexical_results, rank_constant=60)
 
     assert [result.chunk_id for result in fused] == [2, 1, 3]
     assert fused[0].rank_source == "Hybrid"
@@ -146,7 +146,7 @@ def test_reciprocal_rank_fusion_marks_hybrid_source() -> None:
     vector_results = [SearchResult(chunk_id=1, text="a", source_title="A")]
     lexical_results = [SearchResult(chunk_id=2, text="b", source_title="B")]
 
-    fused = reciprocal_rank_fusion(vector_results, lexical_results)
+    fused = reciprocal_rank_fusion(vector_results, lexical_results, rank_constant=60)
 
     assert {result.rank_source for result in fused} == {"Hybrid"}
 

tests/ebook_search/test_rag_pipeline.py

@@ -30,7 +30,7 @@ def test_search_ebooks_runs_vector_and_bm25_in_parallel(mocker: MockerFixture) -
 
     def fake_bm25_candidates(query, _config):
         """Return BM25 candidates after confirming vector search has started."""
-        assert query == "parallel"
+        assert query == "what is parallel"
         bm25_started.set()
         assert vector_started.wait(timeout=2)
         return [SearchResult(chunk_id=2, text="bm25", source_title="BM25", bm25_score=2.0)]
@@ -46,5 +46,4 @@ def test_search_ebooks_runs_vector_and_bm25_in_parallel(mocker: MockerFixture) -
     assert timings["Embedding + vector search"].counts_toward_total is False
     assert timings["BM25 search"].counts_toward_total is False
     assert timings["Hybrid retrieval"].counts_toward_total is True
-    assert timings["BM25 query preparation"].counts_toward_total is True
     assert received_engines == [engine]

tests/ebook_search/test_rerank.py

@@ -33,7 +33,7 @@ def rerank_response(payload: dict[str, object] | None = None, *, content: bytes
     )
 
 
-def test_config_defaults_keep_reranking_optional(mocker: MockerFixture) -> None:
+def test_config_defaults_enable_reranking(mocker: MockerFixture) -> None:
     mocker.patch.dict(environ, {}, clear=False)
     environ.pop("EBOOK_SEARCH_RERANK_ENABLED", None)
     environ.pop("EBOOK_SEARCH_RERANK_BASE_URL", None)
@@ -43,7 +43,7 @@ def test_config_defaults_keep_reranking_optional(mocker: MockerFixture) -> None:
 
     config = load_rerank_config()
 
-    assert config.enabled is False
+    assert config.enabled is True
     assert config.base_url == "http://192.168.90.25:8001"
     assert config.model == "qwen3-reranker-06b"
     assert config.candidates == 24