Function run_suite

tools/bench_src_index.py:145–190

Run all benchmark queries and return stats dict.

(
    conn: sqlite3.Connection,
    label: str,
    srcs_single: list[str],
    srcs_multi: list[str],
    srcs_subcmd: list[str],
    iterations: int,
)

def run_suite(
    conn: sqlite3.Connection,
    label: str,
    srcs_single: list[str],
    srcs_multi: list[str],
    srcs_subcmd: list[str],
    iterations: int,
) -> dict[str, dict[str, float]]:
    """Run all benchmark queries and return stats dict."""
    print(f"\n{'=' * 60}")
    print(f"  {label}")
    print(f"{'=' * 60}")

    results: dict[str, dict[str, float]] = {}

    # Warm up the page cache
    conn.execute("SELECT COUNT(*) FROM mappings").fetchone()
    conn.execute("SELECT COUNT(*) FROM parsed_manpages").fetchone()

    # 1) find_man_page — single destination
    params = [(s,) for s in srcs_single]
    times = bench_query(conn, QUERY_FIND, params, iterations)
    results["find_man_page (1 dst)"] = report("find_man_page (1 dst)", times)

    # 2) find_man_page — multiple destinations
    params = [(s,) for s in srcs_multi]
    times = bench_query(conn, QUERY_FIND, params, iterations)
    results["find_man_page (N dst)"] = report("find_man_page (N dst)", times)

    # 3) distros_for_name
    params = [(s,) for s in srcs_single + srcs_multi]
    random.shuffle(params)
    times = bench_query(conn, QUERY_DISTROS, params, iterations)
    results["distros_for_name"] = report("distros_for_name", times)

    # 4) IN query (subcommand suggestions)
    # Build groups of 2-5 src values to simulate real usage
    src_pool = srcs_single + srcs_multi + srcs_subcmd
    groups = []
    for _ in range(max(iterations, 200)):
        k = random.randint(2, min(5, len(src_pool)))
        groups.append(random.sample(src_pool, k))
    times = bench_in_query(conn, groups, iterations)
    results["suggestions IN(...)"] = report("suggestions IN(...)", times)

    return results


def main() -> None: