ENH: add njit interpolation functions

[mmcky]

This PR adds the following interpolation functions

• interp that accepts x as float or np.array

and the following internal functions:

• _interpf that accepts x as float
• _interpa that accepts x as np.array

[pep8speaks]

Hello @mmcky! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file quantecon/_interpolation.py:

Line 3:1: E302 expected 2 blank lines, found 1
Line 7:1: W293 blank line contains whitespace
Line 45:1: W293 blank line contains whitespace
Line 84:1: W293 blank line contains whitespace
Line 107:13: W292 no newline at end of file

• In the file quantecon/tests/test_interpolation.py:

Line 6:1: E302 expected 2 blank lines, found 1
Line 24:36: W292 no newline at end of file

Comment last updated at 2024-03-04 00:48:39 UTC

[oyamad]

np.interp is available through Numba: see https://numba.pydata.org/numba-doc/dev/reference/numpysupported.html.
Any advantage over that here?

[mmcky]

np.interp is available through Numba: see https://numba.pydata.org/numba-doc/dev/reference/numpysupported.html. Any advantage over that here?

thanks @oyamad probably not -- i didn't realise it was supported.

As we have decided to revert the order of arguments to match the numpy order we should use the numba built-in.

[mmcky]

cc @jstac

[jstac]

Thanks @mmcky heading to the hospital now.

[mmcky]

Thanks @mmcky heading to the hospital now.

Hope all goes well. Thanks @jstac

[jstac]

Thanks @oyamad .

Sorry @mmcky , my bad. perhaps we can test if swapping for the numpy version degrades performance or causes any issues in one or two lectures that are currently using jitted functions plus the version from interpolation. If all is good we can close this PR and swap every 1d interp to the numpy version.