API Reference / agent / tools / data /

generate.py

Synthetic data generation.

Functions

Func

execute_generate_data

Line 11
execute_generate_data(**kwargs) -> Dict[str, Any]

Generate synthetic data using a generator class.

Backs the tuiml_generate_data tool. The generated dataset is written to a temporary CSV whose path can be passed to other tools.

Parameters

generator
str
Generator class name, one of: RandomRBF, Agrawal, LED, Hyperplane, Friedman, MexicanHat, Sine, Blobs, Moons, Circles, SwissRoll (arrives via **kwargs, like all parameters below).
n_samples
int = None
Number of samples to generate.
n_features
int = None
Number of features (generator dependent).
n_classes
int = None
Number of classes (classification generators).
n_clusters
int = None
Number of clusters (clustering generators).
noise
float = None
Noise level (generator dependent).
random_seed
int = None
Random seed; mapped to the generator's random_state.
generator_params
dict = None
Additional constructor parameters merged on top.

Returns

result
dict
On success: status ('success'), generator, file_path (CSV on disk), shape, feature_names, preview (first 5 rows of up to 6 columns) and optionally target_names. On failure: status ('error'), error and optionally suggestion / error_type.