From 5eeb81f867d93be4a6d46238556acaf6b9019f02 Mon Sep 17 00:00:00 2001
From: Zhe Yu <zcabzyu@ucl.ac.uk>
Date: Thu, 29 May 2025 10:05:10 +0800
Subject: [PATCH] feat(cli): set the default `hnsw:M` value to 64

---
 docs/cli.md              |  4 ++--
 src/vectorcode/common.py |  1 +
 tests/test_common.py     | 21 +++++++++++++++++----
 3 files changed, 20 insertions(+), 6 deletions(-)

diff --git a/docs/cli.md b/docs/cli.md
index c5931da9..4f1514f3 100644
--- a/docs/cli.md
+++ b/docs/cli.md
@@ -318,10 +318,10 @@ The JSON configuration file may hold the following values:
   queries. **It's recommended to re-vectorise the collection after modifying these
   options, because some of the options can only be set during collection
   creation.** Example:
-  ```json
+  ```json5
+  // the following is the default value.
   "hnsw": {
     "hnsw:M": 64,
-    "hnsw:construction_ef": 100
   }
   ```
 - `filetype_map`: `dict[str, list[str]]`, a dictionary where keys are
diff --git a/src/vectorcode/common.py b/src/vectorcode/common.py
index bfde24a0..f4fff1a6 100644
--- a/src/vectorcode/common.py
+++ b/src/vectorcode/common.py
@@ -193,6 +193,7 @@ async def get_collection(
                 "USER", os.environ.get("USERNAME", "DEFAULT_USER")
             ),
             "embedding_function": configs.embedding_function,
+            "hnsw:M": 64,
         }
         if configs.hnsw:
             for key in configs.hnsw.keys():
diff --git a/tests/test_common.py b/tests/test_common.py
index 16382ad2..98f1370b 100644
--- a/tests/test_common.py
+++ b/tests/test_common.py
@@ -327,7 +327,19 @@ async def test_get_collection():
             ),
             "created-by": "VectorCode",
         }
-        mock_client.get_or_create_collection.return_value = mock_collection
+
+        async def mock_get_or_create_collection(
+            self,
+            name=None,
+            configuration=None,
+            metadata=None,
+            embedding_function=None,
+            data_loader=None,
+        ):
+            mock_collection.metadata.update(metadata or {})
+            return mock_collection
+
+        mock_client.get_or_create_collection.side_effect = mock_get_or_create_collection
         MockAsyncHttpClient.return_value = mock_client
 
         collection = await get_collection(mock_client, config, make_if_missing=True)
@@ -336,6 +348,7 @@ async def test_get_collection():
             "USER", os.environ.get("USERNAME", "DEFAULT_USER")
         )
         assert collection.metadata["created-by"] == "VectorCode"
+        assert collection.metadata["hnsw:M"] == 64
         mock_client.get_or_create_collection.assert_called_once()
         mock_client.get_collection.side_effect = None
 
@@ -361,7 +374,7 @@ async def test_get_collection_hnsw():
         embedding_function="SentenceTransformerEmbeddingFunction",
         embedding_params={},
         project_root="/test_project",
-        hnsw={"ef_construction": 200, "m": 32},
+        hnsw={"ef_construction": 200, "M": 32},
     )
 
     with patch("chromadb.AsyncHttpClient") as MockAsyncHttpClient:
@@ -374,7 +387,7 @@ async def test_get_collection_hnsw():
             ),
             "created-by": "VectorCode",
             "hnsw:ef_construction": 200,
-            "hnsw:m": 32,
+            "hnsw:M": 32,
             "embedding_function": "SentenceTransformerEmbeddingFunction",
             "path": "/test_project",
         }
@@ -394,7 +407,7 @@ async def test_get_collection_hnsw():
         )
         assert collection.metadata["created-by"] == "VectorCode"
         assert collection.metadata["hnsw:ef_construction"] == 200
-        assert collection.metadata["hnsw:m"] == 32
+        assert collection.metadata["hnsw:M"] == 32
         mock_client.get_or_create_collection.assert_called_once()
         assert (
             mock_client.get_or_create_collection.call_args.kwargs["metadata"]