From ef4c69d1283980c08061e53246efc90558ac993f Mon Sep 17 00:00:00 2001
From: Andy Lee <andylizf@outlook.com>
Date: Mon, 25 Aug 2025 16:08:16 -0700
Subject: [PATCH] chore(ci): remove paru-bin submodule and config to fix
 checkout --recurse-submodules

---
 .gitmodules                                   |  2 ++
 benchmarks/bm25_diskann_baselines/README.md   |  4 ++--
 .../bm25_diskann_baselines/run_diskann.py     | 23 +++++--------------
 paru-bin                                      |  1 -
 4 files changed, 10 insertions(+), 20 deletions(-)
 delete mode 160000 paru-bin

diff --git a/.gitmodules b/.gitmodules
index c1cd540..813256e 100644
--- a/.gitmodules
+++ b/.gitmodules
@@ -14,3 +14,5 @@
 [submodule "packages/leann-backend-hnsw/third_party/libzmq"]
 	path = packages/leann-backend-hnsw/third_party/libzmq
 	url = https://github.com/zeromq/libzmq.git
+
+# Ensure CI can update this submodule; used only for Arch packaging and not required for builds.
diff --git a/benchmarks/bm25_diskann_baselines/README.md b/benchmarks/bm25_diskann_baselines/README.md
index adf4736..297a067 100644
--- a/benchmarks/bm25_diskann_baselines/README.md
+++ b/benchmarks/bm25_diskann_baselines/README.md
@@ -9,12 +9,12 @@ aws s3 sync s3://powerrag-diskann-rpj-wiki-20250824-224037-194d640c/diskann_rpj_
 - Machine-specific; results measured locally with the current repo.
 
 DiskANN (NQ queries, search-only)
-- Command: `uv run benchmarks/bm25_diskann_baselines/run_diskann.py`
+- Command: `uv run --script benchmarks/bm25_diskann_baselines/run_diskann.py`
 - Settings: `recompute_embeddings=False`, embeddings precomputed (excluded from timing), batching off, caching off (`cache_mechanism=2`, `num_nodes_to_cache=0`)
 - Result: avg 0.019339 s/query, QPS 51.71 (p50 ~0.018936 s, p95 ~0.023573 s)
 
 BM25
-- Command: `uv run --script ./benchmarks/run_bm25.py`
+- Command: `uv run --script benchmarks/bm25_diskann_baselines/run_bm25.py`
 - Settings: `k=10`, `k1=0.9`, `b=0.4`, queries=100
 - Result: avg 0.026976 s/query, QPS 37.07 (p50 0.024729 s, p90 0.042158 s, p95 0.047099 s, p99 0.053520 s)
 
diff --git a/benchmarks/bm25_diskann_baselines/run_diskann.py b/benchmarks/bm25_diskann_baselines/run_diskann.py
index 2173af3..a4dd558 100644
--- a/benchmarks/bm25_diskann_baselines/run_diskann.py
+++ b/benchmarks/bm25_diskann_baselines/run_diskann.py
@@ -1,19 +1,8 @@
-#!/usr/bin/env python3
-"""
-Run DiskANN with real NQ queries (search-only timing).
-
-Steps:
-- Load queries from nq_open.jsonl
-- Compute embeddings (facebook/contriever-msmarco) once upfront
-- Search via DiskANN (no recompute, no batching), measure per-query latency
-
-Example:
-  python benchmarks/bm25_diskann_baselines/run_diskann_nq.py \
-    --index-dir benchmarks/data/indices/diskann_rpj_wiki \
-    --index-prefix ann \
-    --queries-file benchmarks/data/queries/nq_open.jsonl \
-    --num-queries 200 --top-k 10 --complexity 120 --threads 1 --beam-width 1
-"""
+# /// script
+# dependencies = [
+#   "leann-backend-diskann"
+# ]
+# ///
 
 import argparse
 import json
@@ -47,7 +36,7 @@ def main() -> None:
     ap.add_argument("--queries-file", default="benchmarks/data/queries/nq_open.jsonl")
     ap.add_argument("--num-queries", type=int, default=200)
     ap.add_argument("--top-k", type=int, default=10)
-    ap.add_argument("--complexity", type=int, default=120)
+    ap.add_argument("--complexity", type=int, default=62)
     ap.add_argument("--threads", type=int, default=1)
     ap.add_argument("--beam-width", type=int, default=1)
     ap.add_argument("--cache-mechanism", type=int, default=2)
diff --git a/paru-bin b/paru-bin
deleted file mode 160000
index 92a5542..0000000
--- a/paru-bin
+++ /dev/null
@@ -1 +0,0 @@
-Subproject commit 92a55429afbec4fceeb2cef843245105307444d2