WaveletCoherence.compute#

WaveletCoherence.compute(first, second)[source]#

Compute coherence and phase between two aligned series.

Builds the scale grid for the sample length, transforms both series, smooths and normalizes the cross-wavelet, marks the cone of influence, and – if surrogates were requested – draws the AR(1) surrogates and their per-scale significance level.

Parameters:

  • first (ArrayLike) – The (T,) reference series; a positive phase means it leads.

  • second (ArrayLike) – The other (T,) series, of the same length.

Return type:

WaveletCoherenceResult

Returns:

The WaveletCoherenceResult.

Raises:

  • DimensionError – If either series is not one-dimensional or the second does not match the first’s length.

  • SpecificationError – If they are shorter than 32.

Example

>>> rng = np.random.default_rng(0)
>>> a, b = rng.standard_normal(128), rng.standard_normal(128)
>>> res = WaveletCoherence(per_octave=2, surrogates=0).compute(a, b)
>>> res.coherence.shape, res.significance is None, bool(res.cone[-1].all())
((13, 128), True, True)
>>> WaveletCoherence(surrogates=0).compute(a, b[:100])
Traceback (most recent call last):
    ...
cultivars.exceptions.DimensionError: second must be one-dimensional with length 128; ...