perda.core_data_structures.resampling#

class perda.core_data_structures.resampling.ResampleMethod(*values)[source]#

Bases: str, Enum

Interpolation strategy used when aligning a series to a target timestamp grid.

CUBIC = 'cubic'#
LINEAR = 'linear'#
NEAREST = 'nearest'#
ZOH = 'zoh'#
perda.core_data_structures.resampling.resample_to_freq(ts, val, freq_hz, timestamp_divisor, method=ResampleMethod.LINEAR)[source]#

Resample a time series onto a uniform frequency grid.

Parameters:

  • ts (NDArray) – Source timestamps (int64)

  • val (NDArray) – Source values

  • freq_hz (float) – Target sampling frequency in Hz

  • timestamp_divisor (float) – Raw timestamp units per second (e.g. 1e6 for microseconds)

  • method (ResampleMethod, optional) – Interpolation method. Default is LINEAR.

Return type:

Tuple[ndarray[tuple[Any, ...], dtype[TypeVar(_ScalarT, bound= generic)]], ndarray[tuple[Any, ...], dtype[TypeVar(_ScalarT, bound= generic)]]]

Returns:

  • target_ts (NDArray) – Uniform timestamp grid (int64)

  • resampled_val (NDArray) – Values interpolated onto the uniform grid