Add property flags to dpnp_array

[densmirn]

No description provided.

[PokhodenkoSA]

Possibly, adding a test would be great too.

[Alexander-Makaryev]

Tests should check the correctness/expected results of the functionality.
But these tests check something else... maybe they check that we can create 2 objects with the equal property, is not it?)
The example below demonstrates this.
As you can see, the current implementation does not provide the expected functionality, but tests pass.

In [1]: import dpctl.tensor as dpt

In [2]: shape = (5, 2)

In [3]: X = dpt.usm_ndarray(shape)

In [4]: X.flags
Out[4]: 1

In [5]: type(X.flags)
Out[5]: int

In [6]: import numpy

In [7]: Y = numpy.ndarray((2,5))

In [8]: Y.flags
Out[8]:
  C_CONTIGUOUS : True
  F_CONTIGUOUS : False
  OWNDATA : True
  WRITEABLE : True
  ALIGNED : True
  WRITEBACKIFCOPY : False
  UPDATEIFCOPY : False

In [9]: type(Y.flags)
Out[9]: numpy.flagsobj

[densmirn]

Tests should check the correctness/expected results of the functionality. But these tests check something else... maybe they check that we can create 2 objects with the equal property, is not it?) The example below demonstrates this. As you can see, the current implementation does not provide the expected functionality, but tests pass.

In [1]: import dpctl.tensor as dpt

In [2]: shape = (5, 2)

In [3]: X = dpt.usm_ndarray(shape)

In [4]: X.flags
Out[4]: 1

In [5]: type(X.flags)
Out[5]: int

In [6]: import numpy

In [7]: Y = numpy.ndarray((2,5))

In [8]: Y.flags
Out[8]:
  C_CONTIGUOUS : True
  F_CONTIGUOUS : False
  OWNDATA : True
  WRITEABLE : True
  ALIGNED : True
  WRITEBACKIFCOPY : False
  UPDATEIFCOPY : False

In [9]: type(Y.flags)
Out[9]: numpy.flagsobj

dpnp is based on dpctl in case of property flags where flags differ numpy. Tests check dpnp flags are equal to dpctl flags.

[densmirn]

In case if we dpnp should work like numpy in case of flags, I vote for closing the PR.

[antonwolfy]

Done in #1265