perf(ebook-search): run phrase detection in parallel with retrieval

Move protected phrase detection into the retrieval gather so it runs
concurrently with vector and BM25 candidates instead of sequentially
before them. Make the search API accept real bool form fields for
rerank/phrase_matching, gate phrase matching on both the request and
config kill switch, and reflow log f-strings for readability.

python/ebook_search/api/routes/search.py

@@ -79,8 +79,9 @@ async def search(
     engine: AppEngine,
     client: AppHttpClient,
     query: Annotated[str, Form()],
-    rerank: Annotated[str | None, Form()] = None,
-    phrase_matching: Annotated[str | None, Form()] = None,
+    *,
+    rerank: Annotated[bool, Form()] = False,
+    phrase_matching: Annotated[bool, Form()] = False,
 ) -> HTMLResponse:
     """Run a search and render HTMX results."""
     try:
@@ -89,8 +90,8 @@ async def search(
             client,
             query,
             config,
-            rerank=rerank == "true",
-            phrase_matching=phrase_matching == "true",
+            rerank=rerank,
+            phrase_matching=phrase_matching,
         )
     except Exception as error:
         logger.exception("ebook_search_request_failed")