[Bug]: COSINE metric type not supported

[darthjaja6]

Is there an existing issue for this?

• [X] I have searched the existing issues

Describe the bug

When I use the code below to create a collection and create index for it

        from pymilvus import Collection, DataType, FieldSchema, CollectionSchema
        collection_name = "some_collection"
        id_field = FieldSchema(name="id", dtype=DataType.INT64,
                               is_primary=True, description="auto id")
        category_field = FieldSchema(
            name="category", dtype=DataType.VARCHAR, max_length=65535, description="description of object")
        created_at_field = FieldSchema(
            name="created_at", dtype=DataType.INT64,  description="unix seconds, publish time")
        vector_field = FieldSchema(
            name="vector", dtype=DataType.FLOAT_VECTOR, dim=VISUAL_EMBEDDING_DIM)

        schema = CollectionSchema(fields=[id_field, category_field, created_at_field, vector_field],
                                  auto_id=True,
                                  description="Knowledge base that contians ")
        collection = Collection(
            name=collection_name, schema=schema)
        index_params = {"index_type": "FLAT", "metric_type": "COSINE", "params": {"nlist": self.nlist}}
        collection.create_index(
            field_name=vector_field.name, index_params=index_params)

I got error

RPC error: [create_index], <MilvusException: (code=1, message=metric type not found or not supported, supported: [L2 IP])>, <Time:{'RPC start': '2023-10-09 04:02:01.609395', 'RPC error': '2023-10-09 04:02:01.629531'}>

It seems to indicate "COSINE" is not supported.

Expected Behavior

From the doc, https://milvus.io/docs/metric.md#Cosine-Similarity, COSINE should be supported

Steps/Code To Reproduce behavior

Run the code below:
        from pymilvus import Collection, DataType, FieldSchema, CollectionSchema
        collection_name = "some_collection"
        id_field = FieldSchema(name="id", dtype=DataType.INT64,
                               is_primary=True, description="auto id")
        category_field = FieldSchema(
            name="category", dtype=DataType.VARCHAR, max_length=65535, description="description of object")
        created_at_field = FieldSchema(
            name="created_at", dtype=DataType.INT64,  description="unix seconds, publish time")
        vector_field = FieldSchema(
            name="vector", dtype=DataType.FLOAT_VECTOR, dim=VISUAL_EMBEDDING_DIM)

        schema = CollectionSchema(fields=[id_field, category_field, created_at_field, vector_field],
                                  auto_id=True,
                                  description="Knowledge base that contians ")
        collection = Collection(
            name=collection_name, schema=schema)
        index_params = {"index_type": "FLAT", "metric_type": "COSINE", "params": {"nlist": self.nlist}}
        collection.create_index(
            field_name=vector_field.name, index_params=index_params)

Environment details

- Hardware/Softward conditions (OS, CPU, GPU, Memory):
MacOS
- Method of installation (Docker, or from source):
Docker
- Milvus version (v0.3.1, or v0.4.0):
pymilvus is v2.3.1, milvus version is v2.2.9
- Milvus configuration (Settings you made in `server_config.yaml`):

Anything else?

Did more research by creating the index from Attu, got the same result: Is there anything I can do to enable COSINE metrics?

No response