Polyval handles non-array as second argument

[HugoNumworks]

This would allow the evaluation of a polynomial at single value, as allowed in numpy.
https://numpy.org/doc/stable/reference/generated/numpy.polyval.html

[v923z]

What happens, if o_x can't be converted to a list?

[v923z]

Whenever there is a user-facing change, increment the version number in

micropython-ulab/code/ulab.c

Line 36 in 47ad73a

#define ULAB_VERSION 6.0.7

. This would be 6.0.8 now.

[v923z]

I'm trying to understand what this PR wants to solve: you call the function

micropython-ulab/code/ndarray.c

Line 217 in 47ad73a

bool ndarray_object_is_array_like(mp_obj_t o_in) {

which basically tells you, whether an object is iterable. If so, then you iterate over it in

micropython-ulab/code/numpy/poly.c

Lines 238 to 247 in 45d6375

	mp_obj_t x_item, x_iterable = mp_getiter(o_x, &x_buf);
	while ((x_item = mp_iternext(x_iterable)) != MP_OBJ_STOP_ITERATION) {
	mp_float_t _x = mp_obj_get_float(x_item);
	mp_float_t y = p[0];
	for(uint8_t j=0; j < plen-1; j++) {
	y *= _x;
	y += p[j+1];
	}
	*array++ = y;
	}

Why is a conversion to a list going to help?

[HugoNumworks]

I'm trying to understand what this PR wants to solve: you call the function

I just needed Polyval to handle scalar and not only lists. Otherwise,

np.polyval([3,0,1], 5)

raises a inputs are not iterable message.
Converting to a list beforehand was an easy way to handle np.polyval([3,0,1], 5) as if the user inputted np.polyval([3,0,1], [5])

The big flaw is that it returns a list in both case which is not Numpy's behavior, so I have to rework this anyway.

[v923z]

You could call the

micropython-ulab/code/ndarray.c

Line 1690 in 47ad73a

ndarray_obj_t *ndarray_from_mp_obj(mp_obj_t obj, uint8_t other_type) {

function in that case.

[HugoNumworks]

@v923z , thanks for the suggestion.
I went with a different implementation where I no longer need to convert to a list.
Now a scalar is returned if the second argument is a scalar.
I also incremented the version number and factorized a method.
I Hope this helps !

[v923z]

What happens, if the input is not array-like, but not a float, either? E.g., what do you do with a complex scalar?

[HugoNumworks]

It will happen the same as if the input is array-like, with non-float elements: a type error such as

TypeError: can't convert complex to float