Open moorepants opened 6 years ago
We may be able to use numpy.vectorize() so that students only have to make their functions work with floats and don't have to construct them to work with 1D arrays/broadcasting, etc. Forgot that this existed. Might simplify a lot.
I think that this would incur a speed penalty though.
This doesn't seem to work because we rely on the functions argument names to stay in tact. vectorize() seems to muck with the argument names.
vectorize()
We may be able to use numpy.vectorize() so that students only have to make their functions work with floats and don't have to construct them to work with 1D arrays/broadcasting, etc. Forgot that this existed. Might simplify a lot.