DuckDB Connector

KumoRFM can connect directly to DuckDB databases, automatically inferring table metadata and relationships from the schema.

Installation

The DuckDB backend requires the ADBC DuckDB driver:

pip install kumoai[duckdb]

Quick Start

Connect to a file-based DuckDB database:

import kumoai.rfm as rfm

graph = rfm.Graph.from_duckdb(connection="my_database.duckdb")

This will:

Connect to the DuckDB database
Discover all non-temporary, non-internal tables automatically
Infer column metadata (data types, semantic types, primary keys, time columns)
Detect foreign key relationships
Print a summary of the inferred metadata and links

Specifying Tables

Control which tables to include and customize their configuration:

graph = rfm.Graph.from_duckdb(
    connection="my_database.duckdb",
    tables=[
        "USERS",
        {"name": "ORDERS", "source_name": "ORDERS_SNAPSHOT"},
        {"name": "ITEMS", "primary_key": "ITEM_ID"},
    ],
)

Table configuration options:

Key	Description	Required
`name`	The table name used in PQL queries	Yes
`source_name`	The actual table name in the database (if different from `name`)	No
`primary_key`	Override the auto-detected primary key	No
`time_column`	The name of the time column for this table	No
`end_time_column`	The name of the end time column for this table	No
`columns`	Selected source columns or column specs, including expression columns	No

Connection Options

In-memory database:

from kumoai.rfm.backend.duckdb import connect

conn = connect()  # in-memory DuckDB database
graph = rfm.Graph.from_duckdb(connection=conn)

From a file path:

graph = rfm.Graph.from_duckdb(connection="path/to/database.duckdb")

From an existing ADBC connection:

from kumoai.rfm.backend.duckdb import connect

conn = connect("path/to/database.duckdb")
graph = rfm.Graph.from_duckdb(connection=conn)

From a connection config dict:

graph = rfm.Graph.from_duckdb(
    connection={"uri": "path/to/database.duckdb"},
)

Controlling Metadata Inference

graph = rfm.Graph.from_duckdb(
    connection="my_database.duckdb",
    infer_metadata=False,
    verbose=False,
)

graph.infer_metadata()
graph.infer_links()

Manual Edge Specification

Override automatic link detection by providing edges explicitly:

graph = rfm.Graph.from_duckdb(
    connection="my_database.duckdb",
    edges=[
        ("ORDERS", "USER_ID", "USERS"),
        ("ORDERS", "ITEM_ID", "ITEMS"),
    ],
)

Get Started

KumoRFM SDK

KumoRFM Fine-Tuning SDK

KumoRFM Fine-Tuning Platform

DuckDB Connector

Installation

Quick Start

Specifying Tables

Connection Options

Controlling Metadata Inference

Manual Edge Specification

​Installation

​Quick Start

​Specifying Tables

​Connection Options

​Controlling Metadata Inference

​Manual Edge Specification

Installation

Quick Start

Specifying Tables

Connection Options

Controlling Metadata Inference

Manual Edge Specification