TSInterpolator

I had assumed that the transformer TSInterpolator interpolated missing values. Instead, it resamples data to make it different length. To me this is resampling not interpolation. It does not have the capability to remove missing values, and even if you give it that capability, it does not remove them. So

from aeon.utils.validation import has_missing
X = np.random.random(size=(10,1,100))
X[0][0][55] = np.nan
print(has_missing(X))
from aeon.transformations.collection import TSInterpolator
interpolator = TSInterpolator(X.shape[2])
X = interpolator.fit_transform(X)
print(has_missing(X))

raises

ValueError: Data seen by instance of TSInterpolator has missing values, but TSInterpolator cannot handle missing values.

If I change "capability:missing_values": True,
it does not remove the missing, even if I resize

from aeon.transformations.collection import TSInterpolator
interpolator = TSInterpolator(X.shape[2]-10)
X = interpolator.fit_transform(X)
print(X.shape," has missing = ",has_missing(X))

even the comments refer to resizing, so its just badly named imo. I think we should rename this to Resizer then having a simple transform Imputer that simply fills in missing values through linear interpolation.