Data Designer Configuration

DataDesignerConfig is the top-level configuration object passed to Data Designer. It declares the columns to generate and may include model configs, tool configs, seed settings, sampler constraints, processors, and profiler configs.

Prefer DataDesignerConfigBuilder for programmatic construction. Direct DataDesignerConfig instantiation is also supported.

Classes:

Name	Description
`DataDesignerConfig`	Configuration for NeMo Data Designer.

`DataDesignerConfig`

Bases: ExportableConfigBase

Configuration for NeMo Data Designer.

This class defines the main configuration structure for NeMo Data Designer, which the engine consumes when generating synthetic data.

Attributes:

Name	Type	Description
`columns`	`list[Annotated[ColumnConfigT, Field(discriminator='column_type')]]`	Required list of column configurations defining how each column should be generated. Must contain at least one column.
`model_configs`	`list[ModelConfig] \| None`	Optional list of model configurations for LLM-based generation. Each model config defines the model, provider, and inference parameters.
`tool_configs`	`list[ToolConfig] \| None`	Optional list of tool configurations for MCP tool calling. Each tool config defines the provider, allowed tools, and execution limits.
`seed_config`	`SeedConfig \| None`	Optional seed dataset settings to use for generation.
`constraints`	`list[ColumnConstraintInputT] \| None`	Optional list of column constraints.
`profilers`	`list[ColumnProfilerConfigT] \| None`	Optional list of column profilers for analyzing generated data characteristics.
`processors`	`list[Annotated[ProcessorConfigT, Field(discriminator='processor_type')]] \| None`	Optional list of processor configurations for post-generation transformations.

Methods:

Name	Description
`fingerprint`	Compute a deterministic content-addressable fingerprint of this config.

`fingerprint()`

Compute a deterministic content-addressable fingerprint of this config.

See data_designer.config.fingerprint.fingerprint_config for the full list of identity-relevant and excluded fields, and how custom column generators are identified.

Returns:

Type	Description
`dict[str, str \| int]`	A dict with `config_hash`, `config_hash_algo`, and
`dict[str, str \| int]`	`config_hash_version`.

Source code in packages/data-designer-config/src/data_designer/config/data_designer_config.py

def fingerprint(self) -> dict[str, str | int]:
    """Compute a deterministic content-addressable fingerprint of this config.

    See `data_designer.config.fingerprint.fingerprint_config` for the full
    list of identity-relevant and excluded fields, and how custom column
    generators are identified.

    Returns:
        A dict with `config_hash`, `config_hash_algo`, and
        `config_hash_version`.
    """
    return fingerprint_config(self)