patch.pad and patch.pad_to

[d-chambers]

I am thinking about implementing patch.pad and patch.pad_to.

The former will pretty much follow numpy.pad, with the exception that kwargs can be used to specify dimensions.

pad_to would work something like this:

Pad to a certain time/distance

# pad to a specific start/end along a dimension
# simply determine if the end value is before or after the coordinate to know to add padding to start or end.
padded = patch.pad_to(time="2021-01-01", distance=100, constant_values=0)

Pad to a certain duration

# simply determine if the end value is before or after the coordinate to know to add padding to start or end.
padded = patch.pad_to(time=100 * s, distance=100 * m, relative=True, constant_values=0)

This may need a from_end option to know if the padding should be at the start or end.

Pad to a certain array shape (in samples)

This will be useful for many machine learning models which require input data of a certain shape.

# simply determine if the end value is before or after the coordinate to know to add padding to start or end.
padded = patch.pad_to(time=10_000, distance=120, samples=True, constant_values=0)

It could also use the from_end flag.

Any thoughts or comments on the API?

[d-chambers]

I think think #373 takes care of this.