early and informative exception when passing a non-scalar func to solver.optimize

[joostvanzwieten]

solver.optimize requires a scalar function. If provided with a function that has one dimension or more, than solver.optimize (or actually sparse.take) spits out an error message that is not very helpful to end users. Example:

from nutils import function, solver
solver.optimize('x', function.Argument('x', (1,))**2)

Traceback:

Traceback (most recent call last):
  File "solver.py", line 834, in optimize
    val, res, jac = _integrate_blocks(functional, residual, jacobian, arguments=lhs, mask=mask)
  File "nutils/solver.py", line 981, in _integrate_blocks
    res = [sparse.take(next(data), [m]) for m in mask]
  File "nutils/solver.py", line 981, in <listcomp>
    res = [sparse.take(next(data), [m]) for m in mask]
  File "nutils/sparse.py", line 290, in take
    assert axis == len(shape)
AssertionError

Let's raise an early (don't traverse into _integrate_blocks) and informative exception instead.

[Sbhat92]

I would like to work on this issue

[gertjanvanzwieten]

@Sbhat92 that would be great!

[Sbhat92]

I'm having trouble understanding how to check the dimension (if it is greater than or equal to 1) or if is a scalar. Can you please direct me?

[gertjanvanzwieten]

Sorry for not noticing your question here before! I'll reply in the PR.