pyinterp.TemporalAxis

class pyinterp.TemporalAxis(self, points: object, epsilon: object | None = None, period: object | None = None)

Bases: object

Temporal axis for datetime64 or timedelta64 values.

Create a coordinate axis for datetime64 or timedelta64 values.

This class accepts numpy.datetime64 or numpy.timedelta64 arrays and supports periodicity and tolerance (epsilon) using timedelta64 values.

Parameters:

• points – Axis coordinate values as a numpy datetime64 or timedelta64 array.

• epsilon – Maximum allowed difference between two values to consider them equal. Must have the same or a coarser resolution than points. Defaults to None.

• period – Period for cyclic/periodic axes (for example, 24 hours or 365 days). Must have the same or a coarser resolution than points. Defaults to None.

Raises:

• ValueError – If epsilon or period have a finer resolution than points.

• TypeError – If points is not a datetime64 or timedelta64 array.

Examples

>>> import numpy as np
>>> import pyinterp

Create a regular hourly axis over one day

>>> times = np.arange('2024-01-01', '2024-01-02', dtype='datetime64[h]')
>>> axis = pyinterp.TemporalAxis(times)

Create a daily periodic axis (24-hour cycle)

>>> axis_periodic = pyinterp.TemporalAxis(
...     times,
...     epsilon=np.timedelta64(1, 'm'),  # 1 minute tolerance
...     period=np.timedelta64(24, 'h')   # 24-hour period
... )

Create an irregular axis with tolerance

>>> irregular_times = np.array([
...     '2024-01-01T00:00',
...     '2024-01-01T06:15',
...     '2024-01-01T12:30',
...     '2024-01-01T18:45'
... ], dtype='datetime64[m]')
>>> axis_irregular = pyinterp.TemporalAxis(
...     irregular_times,
...     epsilon=np.timedelta64(5, 'm')  # 5 minute tolerance
... )

Create a timedelta64 axis for durations

>>> durations = np.array([0, 3600, 7200, 10800], dtype='timedelta64[s]')
>>> axis_duration = pyinterp.TemporalAxis(durations)

Attributes

dtype	Get the numpy dtype of this axis.
is_periodic	True if this axis represents a periodic variable.
period	Get the period of this axis.

Public Methods

back(self)	Get the last value of this axis.
cast_to_temporal_axis(self, array)	Safely cast an array to the axis resolution.
find_index(self, coordinates, bounded)	Find the index of the axis element that contains a coordinate.
find_indexes(self, coordinates)	Find indexes of axis elements surrounding each coordinate.
flip(self[, inplace])	Reverse the order of elements in the axis.
front(self)	Get the first value of this axis.
increment(self)	Get the increment (step) between values in this axis.
is_ascending(self)	Check if the axis values are in ascending order.
is_regular(self)	Check if the axis is regular.
max_value(self)	Get the maximum value of this axis.
min_value(self)	Get the minimum value of this axis.

Special Methods

__copy__(self)	Implement the shallow copy operation.
__eq__(self, other)	Override the default behavior of the == operator.
__getitem__(-> object)	Overloaded function.
__getstate__(self)	Get the state for pickling.
__len__(self)	Return the length of the axis.
__ne__(self, other)	Override the default behavior of the != operator.
__new__(*args, **kwargs)
__repr__(self)	Return the string representation of this Axis.
__setstate__(self, state)	Set the state for unpickling.