numpy - 3D Extrapolation in python (basically, scipy.griddata extended to extrapolate) -
i using griddata function in scipy interpolate 3 , 4 dimensional data. works champ, except returns bunch of nans because of points need outside range of input data. given n-d data works "linear" mode interpolation anyway, should snap have griddata extrapolation instead of returning nan. has done or found workaround? clarify: have unstructured data, can't use of functions require regular grid. thanks! alex not quite sure work , not available yet, in development version of numpy there 'pad' array function... https://github.com/numpy/numpy/blob/master/numpy/lib/arraypad.py one of options 'linear_ramp' extrapolates (pads) outward starting @ edge value , linearly increase/decreases specified end value. it pure python function copy path , import (untested me though)