Data processing basics¶
The examples below detail pre-processing functionality in spm1d.util.
Warning
spm1d currently offers only very basic data processing functions.
Please consider using other software tools like SciPy and Visual3D for data processing.
Interpolation¶
./spm1d/examples/processing/ex_interpolate.py
Linear interpolation to n points:
>>> y0 = np.random.rand(51)
>>> y1 = np.random.rand(87)
>>> y2 = np.random.rand(68)
>>> Y = [y0, y1, y2]
>>> Y = spm1d.util.interp(Y, Q=101)
(Source code, png, hires.png, pdf)

Smoothing¶
./spm1d/examples/processing/ex_smooth.py
Convolution with a Gaussian kernel having a given FWHM (full-width at half-maximum):
>>> Y_smoothed = spm1d.util.smooth(Y0, fwhm=5.0)
(Source code, png, hires.png, pdf)
