average_chunks

average_chunks(arr, wts, chunksize)

Average the array arr with weights by chunks

Array len does not have to be multiple of chunksize

This version is optimised for plain numpy. It is roughly ten times faster that average_chunks_jit when used without numba jit. It cannot (yet) be used with numba because the add.reduceat is not support in numba 0.31

Parameters

• arr – 1D array of values

• wts – 1D array of weights

• chunksize – averaging size

Returns

1D array of averaged data_models, 1d array of weights