Support order= in ndarray bound function reshape.

[TomMaullin]

Reporting a bug

• I am using the latest released version of Numba (most recent is visible in
  the change log (https://github.com/numba/numba/blob/master/CHANGE_LOG).
• I have included below a minimal working reproducer (if you are unsure how
  to write one see http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports).

I have written the following function, which works out a permutation I need for my code, it works well without numba but fails with it.

def permOfIkKkI2D(k1,k2,n1,n2):

  # First we need the permutation represented by matrix K in vector format
  permP = np.arange(n1*k2).reshape((n1, k2), order='F').ravel()

  # Now we work out the permutation obtained by the first kronecker product (i.e. I kron K)
  permKron1 = np.repeat(np.arange(k1),n1*k2)*n1*k2+np.tile(permP,k1)

  # Now we work out the final permutation by appplying the second kronecker product (i.e. I kron K kron I)
  p = np.repeat(permKron1,n2)*n2+np.tile(np.arange(n2),n1*k1*k2)

  # Return the permutation
  return(p)

With numba I get this error:

    perm = permOfIkKkI2D(n2,n1,n2,n1)
  File "/home/tommaullin/.local/lib/python3.6/site-packages/numba/dispatcher.py", line 401, in _compile_for_args
    error_rewrite(e, 'typing')
  File "/home/tommaullin/.local/lib/python3.6/site-packages/numba/dispatcher.py", line 344, in error_rewrite
    reraise(type(e), e, None)
  File "/home/tommaullin/.local/lib/python3.6/site-packages/numba/six.py", line 668, in reraise
    raise value.with_traceback(tb)
numba.errors.TypingError: Failed in nopython mode pipeline (step: nopython frontend)
Internal error at <numba.typeinfer.CallConstraint object at 0x7f0202c460f0>.
�[1m�[1m�[1m�[0m
�[0m�[1m[1] During: resolving callee type: BoundFunction(array.reshape for array(int64, 1d, C))�[0m
�[0m�[1m[2] During: typing of call at /home/tommaullin/Documents/BLMM_creation/BLMM/lib/tools2d.py (409)
�[0m
Enable logging at debug level for details.
�[1m
File "../../lib/tools2d.py", line 409:�[0m
�[1mdef permOfIkKkI2D(k1,k2,n1,n2):
    <source elided>
  # First we need the permutation represented by matrix K in vector format
�[1m  permP = np.arange(n1*k2).reshape((n1, k2), order='F').ravel()
�[0m  �[1m^�[0m�[0m

[stuartarchibald]

Thanks for the report. Please could you provide more information about what caused this problem:

1. The code that you were running (as a minimal working reproducer, if you are unsure how to write one see http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports).
2. Any data needed to run the code.
3. The full trace back of the exception.

Once we have a reproducer someone can take a look and work on a fix. Thanks for your help.

[TomMaullin]

Sure!

# Generate random numbers
n1=np.random.randint(1,10)
n2=np.random.randint(1,10)
k1=np.random.randint(1,10)
k2=np.random.randint(1,10)

# Work out the permutation 
perm = permOfIkKkI2D(k1,k2,n1,n2)

[esc]

@TomMaullin thanks for submitting the code to run the example. For convenience, I have patched up a self contained runnable version, (including imports). As a recommendation for the future, if you can provide such an example that we can copy and paste it makes debugging much quicker and easier.

from numba import njit
import numpy as np

@njit
def permOfIkKkI2D(k1,k2,n1,n2):

  # First we need the permutation represented by matrix K in vector format
  permP = np.arange(n1*k2).reshape((n1, k2), order='F').ravel()

  # Now we work out the permutation obtained by the first kronecker product (i.e. I kron K)
  permKron1 = np.repeat(np.arange(k1),n1*k2)*n1*k2+np.tile(permP,k1)

  # Now we work out the final permutation by appplying the second kronecker product (i.e. I kron K kron I)
  p = np.repeat(permKron1,n2)*n2+np.tile(np.arange(n2),n1*k1*k2)

  # Return the permutation
  return(p)

# Generate random numbers
n1=np.random.randint(1,10)
n2=np.random.randint(1,10)
k1=np.random.randint(1,10)
k2=np.random.randint(1,10)

# Work out the permutation
perm = permOfIkKkI2D(k1,k2,n1,n2)

Also, I can reproduce this locally but not yet sure what the problem is, so I will label it as needing triage.

[stuartarchibald]

Doing: NUMBA_DEVELOPER_MODE=1 python issue5435.py
Gives:

Traceback (most recent call last):
  File "<path>/numba/numba/core/errors.py", line 769, in new_error_context
    yield
  File "<path>/numba/numba/core/typeinfer.py", line 487, in __call__
    self.resolve(typeinfer, typevars, fnty)
  File "<path>/numba/numba/core/typeinfer.py", line 507, in resolve
    sig = typeinfer.resolve_call(fnty, pos_args, kw_args)
  File "<path>/numba/numba/core/typeinfer.py", line 1451, in resolve_call
    return self.context.resolve_function_type(fnty, pos_args, kw_args)
  File "<path>/numba/numba/core/typing/context.py", line 197, in resolve_function_type
    res = self._resolve_user_function_type(func, args, kws)
  File "<path>/numba/numba/core/typing/context.py", line 249, in _resolve_user_function_type
    return func.get_call_type(self, args, kws)
  File "<path>/numba/numba/core/types/functions.py", line 215, in get_call_type
    out = template.apply(args, kws)
  File "<path>/numba/numba/core/typing/templates.py", line 290, in apply
    sig = generic(args, kws)
  File "<path>/numba/numba/core/typing/templates.py", line 908, in generic
    sig = method_resolver(self, ty, args, kws)
  File "<path>/numba/numba/core/typing/arraydecl.py", line 375, in resolve_reshape
    assert not kws
AssertionError

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<path>/numba/numba/core/typeinfer.py", line 149, in propagate
    constraint(typeinfer)
  File "<path>/numba/numba/core/typeinfer.py", line 487, in __call__
    self.resolve(typeinfer, typevars, fnty)
  File "<py_path>/lib/python3.8/contextlib.py", line 131, in __exit__
    self.gen.throw(type, value, traceback)
  File "<path>/numba/numba/core/errors.py", line 776, in new_error_context
    reraise(type(newerr), newerr, tb)
  File "<path>/numba/numba/core/utils.py", line 80, in reraise
    raise value
numba.core.errors.InternalError: 
[1] During: resolving callee type: BoundFunction(array.reshape for array(int64, 1d, C))
[2] During: typing of call at issue5435.py (8)


The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "issue5435.py", line 26, in <module>
    perm = permOfIkKkI2D(k1,k2,n1,n2)
  File "<path>/numba/numba/core/dispatcher.py", line 401, in _compile_for_args
    error_rewrite(e, 'typing')
  File "<path>/numba/numba/core/dispatcher.py", line 344, in error_rewrite
    reraise(type(e), e, None)
  File "<path>/numba/numba/core/utils.py", line 79, in reraise
    raise value.with_traceback(tb)
numba.core.errors.TypingError: Failed in nopython mode pipeline (step: nopython frontend)
Internal error at <numba.core.typeinfer.CallConstraint object at 0x7f8c197c8a60>.

[1] During: resolving callee type: BoundFunction(array.reshape for array(int64, 1d, C))
[2] During: typing of call at issue5435.py (8)

Enable logging at debug level for details.

File "issue5435.py", line 8:
def permOfIkKkI2D(k1,k2,n1,n2):
    <source elided>
  # First we need the permutation represented by matrix K in vector format
  permP = np.arange(n1*k2).reshape((n1, k2), order='F').ravel()
  ^

the:

    assert not kws
AssertionError

is the important bit. The bound function ndarray.reshape doesn't support keyword arguments, hence the problem is the order='F'.

[stuartarchibald]

Updated title and made into feature request.