feat(ebook-search): add load-test CLI for the search service

Add a Typer CLI script that drives POST /search on a running server at a
configurable concurrency and reports latency percentiles (p50/p90/p95/p99),
throughput, and HTTP status distribution. Queries are drawn from the shared
eval JSONL set so load testing and evaluation exercise the same questions.

python/ebook_search/eval/dataset.py

@@ -0,0 +1,47 @@
+"""Shared query set loading for evaluation and load testing.
+
+Each JSONL record has a ``query`` and an optional reference ``answer``. ``answerable``
+marks whether the query should be answerable from the library (false for out-of-corpus
+"garbage" queries used to test the refusal path). Relevance for retrieval metrics is
+labeled at source (book) granularity in ``relevant_sources``; source titles must match
+``ebook_source.title`` values for the indexed corpus.
+"""
+
+from __future__ import annotations
+
+import json
+from dataclasses import dataclass
+from pathlib import Path
+
+DEFAULT_QUERIES_PATH = Path(__file__).parent / "data" / "queries.jsonl"
+
+
+@dataclass(frozen=True)
+class GoldQuery:
+    """One labeled query shared by the eval and load-test tools."""
+
+    query: str
+    answer: str | None
+    answerable: bool
+    relevant_sources: tuple[str, ...]
+    relevant_substrings: tuple[str, ...]
+
+
+def load_gold_queries(path: Path = DEFAULT_QUERIES_PATH) -> list[GoldQuery]:
+    """Load labeled queries from a JSONL file."""
+    queries: list[GoldQuery] = []
+    for line in path.read_text(encoding="utf-8").splitlines():
+        stripped = line.strip()
+        if not stripped:
+            continue
+        record = json.loads(stripped)
+        queries.append(
+            GoldQuery(
+                query=str(record["query"]),
+                answer=record.get("answer"),
+                answerable=bool(record.get("answerable", True)),
+                relevant_sources=tuple(record.get("relevant_sources", ())),
+                relevant_substrings=tuple(record.get("relevant_substrings", ())),
+            )
+        )
+    return queries