BaseGenerator - Nixtla

`BaseGenerator`

Bases: BaseModel, ABC Base class for all time series generators. Parameters:

Name	Type	Description	Default
`min_length`	`int`	Minimum length of each series	required
`max_length`	`int`	Maximum length of each series	required
`freq`	`str \| int`	Frequency of the data. Either a pandas offset alias (e.g. ‘D’, ‘h’, ‘5min’, ‘MS’, ‘W-MON’) or an integer for an integer time index	required
`engine`	`str`	Output dataframe library (default: ‘pandas’). Options are ‘pandas’, ‘polars’, ‘cudf’, ‘modin’, ‘pyarrow’	required
`alias`	`str \| None`	Name of the generator (default: class name)	required
`id_col`	`str`	Name of the ID column (default: ‘unique_id’)	required
`time_col`	`str`	Name of the timestamp column (default: ‘ds’)	required
`target_col`	`str`	Name of the value column (default: ‘y’)	required
`start_datetime`	`str`	First timestamp of every series, in any format accepted by pandas.Timestamp (default: ‘2000-01-01’). Ignored when freq is an integer	required
`seed`	`int \| None`	Random seed for reproducibility (default: None)	required
`Exogenous parameters`			required
`exogenous`	`ExogenousConfig \| None`	Configuration for exogenous variable generation. None = no exogenous columns (default: None)	required
`Missingness parameters`			required
`missing_data`	`bool`	Enable missing data patterns (default: False)	required
`missing_pattern`	`str`	Pattern: ‘random’, ‘block’, ‘seasonal’ (default: ‘random’)	required
`missing_rate`	`float`	Proportion of missing values 0-1 (default: 0.1)	required
`missing_block_size`	`int`	Size of missing blocks for ‘block’ pattern (default: 3)	required
`missing_seasonal_period`	`int`	Period for ‘seasonal’ pattern (default: 7)	required
`Anomaly parameters`			required
`anomalies`	`bool`	Enable anomaly injection (default: False)	required
`anomaly_fraction`	`float`	Fraction of points that are anomalies (default: 0.05)	required
`anomaly_types`	`list[str]`	Types: ‘spike’, ‘dip’, ‘level_shift’ (default: [‘spike’, ‘dip’])	required
`spike_magnitude`	`float`	Magnitude of spikes (default: 10.0)	required
`dip_magnitude`	`float`	Magnitude of dips (default: -10.0)	required
`level_shift_magnitude`	`float`	Magnitude of level shifts (default: 20.0)	required
`level_shift_duration`	`int`	Duration of level shifts in time steps (default: 10)	required
`Changepoint parameters`			required
`changepoints`	`bool`	Enable changepoint injection (default: False)	required
`num_changepoints`	`int`	Number of changepoints (default: 2)	required
`changepoint_type`	`str`	Type: ‘level’, ‘trend’, ‘variance’, ‘mixed’ (default: ‘level’)	required
`changepoint_level_changes`	`list[float] \| None`	Size of level changes (default: random)	required
`changepoint_trend_changes`	`list[float] \| None`	Size of trend changes (default: random)	required
`changepoint_variance_changes`	`list[float] \| None`	Size of variance changes (default: random)	required
`changepoint_locations`	`list[float] \| None`	Relative positions 0-1 (default: random)	required

`BaseGenerator.generate`

generate(n_series, start_id=0, n_jobs=-1)

Generate synthetic time series data. Parameters:

Name	Type	Description	Default
`n_series`	`int`	Number of time series to generate	required
`start_id`	`int`	Starting ID for the series numbering (default: 0) Series will be numbered from start_id to start_id + n_series - 1	`0`
`n_jobs`	`int`	Number of parallel workers. -1 (default) uses `RAYON_NUM_THREADS` if set, otherwise all logical cores. Results are seed-deterministic and do not depend on n_jobs.	`-1`

Returns:

Type	Description
`IntoDataFrameT`	DataFrame in long format with columns [id_col, time_col, target_col] (default [‘unique_id’, ‘ds’, ‘y’]), plus any exogenous or flag columns.

`BaseGenerator.generate_single_series`

generate_single_series(length)

Generate values for a single time series. Parameters:

Name	Type	Description	Default
`length`	`int`	The length of the series to generate	required

Returns:

Type	Description
`ndarray`	Array of time series values

⌘I