Skip to content

Latest commit

 

History

History
78 lines (61 loc) · 2.75 KB

python.md

File metadata and controls

78 lines (61 loc) · 2.75 KB

##Python interface We have made an interface so you can use SPLINTER through Python. If you follow the instructions provided below you should be good to go in a matter of minutes (unless you choose to compile the library yourself). The interface has been tested and should work with both Python 2.7 and 3.x.

###Setup First, head to the releases tab and download a copy of splinter-python. Or, you can compile the library for yourself (as described in the documentation). After doing the step make install, you should have a directory structure that looks like this:

  • main

    • lib
      • windows
        • x86
          • splinter-x-y.dll
        • x86-64
          • splinter-x-y.dll
      • linux
        • x86
          • libsplinter-x-y.so
        • x86-64
          • libsplinter-x-y.so
      • osx
        • x86
          • libsplinter-x-y.so
        • x86-64
          • libsplinter-x-y.so
    • include
      • cinterface.h
    • python
      • splinter
        • All files from the matlab directory in the repository

  • The numbers in the file name (x-y) corresponds to the SPLINTER version, where x is the major and y is the minor version.

Make sure the folder called python is in your path, or that that folder is your current working directory. Then you can do import splinter and it should automatically load all classes along with the binary for you.

Basic usage

You can then start using the library by making a DataTable and populate it with samples using addSample().

import splinter

def f(x, y):
    return x**2*y + y**3

d = splinter.DataTable()

for i in range(10):
    for j in range(10):
        d.addSample([i,j], f(i,j))

bspline = splinter.BSpline(d, 3) # Make a BSpline of degree 3 in all dimensions

# Evaluate one point:
approxVal = bspline.eval([0.5,0.5])

# Evaluate three points:
approxVal = bspline.eval([0.3,0.2, 0.4,0.2, 7.8,7.9])
# or, equivalent:
approxVal = bspline.eval([[0.3,0.2], [0.4,0.2], [7.8,7.9]])

# Save BSpline to file for loading it later:
bspline.save("myfile.myextension")

# Load BSpline from file:
loadedBSpline = splinter.BSpline("myfile.myextension")

Notice that if you are going to evaluate the approximant in more than one point, it is preferred to call eval once, instead of n times. This is because you then only make a call to the binary one time, instead of n times.

When you are done populating the DataTable you can create an Approximant of your choosing:

  • BSpline
  • PSpline
  • RadialBasisFunction
  • PolynomialRegression

All these derive from the Approximant base class, and their usage mainly differ in the signature of the constructor.

Please consult the documentation for the C++ version of the library if you still have unanswered questions after reading this document.