resolve_lag_sequence

resolve_lag_sequence(lags, index, clamp=True)

Normalise lag specifications into a sorted numpy array of non-negative integers.

Parameters

Name	Type	Description	Default
lags	(str, int, Sequence[int], range, slice)	- String durations (e.g. `"30 days"`, `"3 months"`) are converted to the number of observation lags implied by `index`. - Integers produce a `range(0, lags)` style sequence (inclusive). - Sequences/ranges/slices are materialised and sorted.	required
index	array - like	A date/time index used to translate duration strings into row counts. The index is sorted internally.	required
clamp	bool	If `True` (default), lags greater than `len(index) - 1` are clipped.	`True`

Returns

Name	Type	Description
	np.ndarray	Sorted array of lag integers starting at zero.

Raises

Name	Type	Description
	ValueError	When the input cannot be interpreted or the index is empty.

Examples

import numpy as np
import pandas as pd
import pytimetk as tk

idx = pd.date_range("2020-01-01", periods=5, freq="D")
tk.resolve_lag_sequence("3 days", idx)

array([0, 1, 2, 3])

tk.resolve_lag_sequence([0, 2, 4], idx)

array([0, 2, 4])