Adding more functionality in Array and iop (to match Array API)

[nicola-bastianello]

there are several dunder methods and functions in the interop layer that need to be implemented to match the Array API

when I have time I will populate this issue with a list, divided by priority

[Simpag]

If all frameworks implement the specific method, just pass through the dunders, dont go through iop

[nicola-bastianello]

done in #7

for Array:

• arithmetic operations: __pos__, __floordiv__, __mod__
• bitwise operations: __invert__, __or__; low priority: __xor__, __lshift__, __rshift__
• in-place: __ifloordiv__, __ipow__, __imod__, __imatmul__, __iand__, __ior__; low priority: __ixor__, __ilshift__, __irshift__
• reflected: __rfloordiv__, __rpow__, __rmod__, __ror__; low priority: __rxor__, __rlshift__, __rrshift__
• attributes: Array.dtype, mT (matrix transpose)
• methods: __bool__, __index__, __int__

[nicola-bastianello]

for iop:

• constants: inf, nan, newaxis; low priority: e, pi
• creation functions: arange, empty, empty_like, match signature for eye, linspace, meshgrid; low priority: full, full_like, tril, triu
• data type functions (all low priority): can_cast, finfo, iinfo, isdtype, result_type
• element-wise functions: ceil, clip, exp, floor, isfinite, isinf, isnan, log, log10, logaddexp, logical_and, logical_not, logical_or, logical_xor, minimum, reciprocal, round; low priority/maybe not: acos, acosh, asin, asinh, atan, atan2, atanh, conj, copysign, cos, cosh, expm1, hypot, imag, log1p, log2, nextafter, real, signbit, sin, sinh, square, tan, tanh, trunc
• indexing functions: take, take_along_axis
• inspection: maybe: default_device, default_dtypes, devices, dtypes
• linear algebra: tensordot
• manipulation functions: concat, repeat, tile, unstack; low priority: broadcast_arrays, broadcast_shapes, broadcast_to, flip, moveaxis, permute_dims, roll
• searching functions: count_nonzero, nonzero, where; low priority: searchsorted
• set functions: isin; low priority: unique_all, unique_counts, unique_inverse, unique_values
• sorting: argsort, sort
• statistical: prod, std; low priority: cumulative_prod, cumulative_sum, var
• diff

done in #7

• element-wise functions: positive, remainder, bitwise_left_shift, bitwise_invert, bitwise_or, bitwise_right_shift, bitwise_xor, floor_divide
• linear algebra: matrix_transpose

[nicola-bastianello]

for iop, linear algebra extension:

• high priority: eig, eigh, eigvals, matrix_norm, solve
• low priority: cholesky, cross, det, inv, matrix_power, matrix_rank, outer, pinv, qr, slogdet, svd, svdvals, trace