:Vector Search SDK

Connect to Hologres

import holo_search_sdk as holo

# Connect to the database.
client = holo.connect(
    host="<HOLO_HOST>",
    port=<HOLO_PORT>,
    database="<HOLO_DBNAME>",
    access_key_id="<ACCESS_KEY_ID>",
    access_key_secret="<ACCESS_KEY_SECRET>",
    schema="public" # Modify the schema as needed.
)

# Establish the connection.
client.connect()

The following table describes the variables.

Create a vector table

You can create the table using a Data Definition Language (DDL) statement. The following code provides an example:

create_table_sql = """
    CREATE TABLE IF NOT EXISTS <TABLE_NAME> (
        id BIGINT PRIMARY KEY,
        content TEXT,
        vector_column FLOAT4[] CHECK (array_ndims(vector_column) = 1 AND array_length(vector_column, 1) = 3)
    );
"""
_ = client.execute(create_table_sql, fetch_result=False)

Open the vector table

columns = {
    "id": ("INTEGER", "PRIMARY KEY"),
    "content": "TEXT",
    "vector_column": "FLOAT4[]"
}
table = client.open_table("<TABLE_NAME>")

Import vector data

data = [
    [1, "Hello world", [0.1, 0.2, 0.3]],
    [2, "Python SDK", [0.4, 0.5, 0.6]],
    [3, "Vector search", [0.7, 0.8, 0.9]]
]
table.insert_multi(data, ["id", "content", "vector_column"])

Set a vector index

table.set_vector_index(
    column="vector_column",
    distance_method="Cosine",
    base_quantization_type="rabitq",
    max_degree=64,
    ef_construction=400
)

Query vector data

• Basic vector search

  # Vector search
  query_vector = [0.2, 0.3, 0.4]
  
  # Limit the number of results.
  results = table.search_vector(
      vector=query_vector, 
      column="vector_column",
      distance_method="Cosine"
  ).limit(10).fetchall()
  
  # Set a minimum distance.
  results = table.search_vector(
      vector=query_vector, 
      column="vector_column",
      distance_method="Cosine"
  ).min_distance(0.5).fetchall()
  
  # Search with an output alias.
  results = table.search_vector(
      vector=query_vector,
      column="vector_column",
      output_name="similarity_score",
      distance_method="Cosine"
  ).fetchall()

• Exact query by primary key

  # Query a single record by primary key.
  result = table.get_by_key(
      key_column="id",
      key_value=1,
      return_columns=["id", "content", "vector_column"]  # Optional. If you do not specify this parameter, all columns are returned.
  ).fetchone()
  
  # Batch query records by a list of primary keys.
  results = table.get_multi_by_keys(
      key_column="id", 
      key_values=[1, 2, 3],
      return_columns=["id", "content"]  # Optional. If you do not specify this parameter, all columns are returned.
  ).fetchall()

Close the connection

# Close the connection.
client.disconnect()