But there's always another problem right?
I need to prepare to use some pretty big matrices, so I'd like to make
Pooling_Matrix be a subclass of either
- sage.matrix.matrix_integer_dense.Matrix_integer_dense,
if the call is through Pooling_Matrix(parent, entries, copy, coerce)
where parent is that of a
sage.matrix.matrix_integer_dense.Matrix_integer_dense object; or
- sage.matrix.matrix_integer_sparse.Matrix_integer_sparse,
if the call is through Pooling_Matrix(parent, entries, copy, coerce)
where parent is that of a
sage.matrix.matrix_integer_sparse.Matrix_integer_sparse object.
Currently I am using the call (through a constructor pooling_matrix):
Pooling_Matrix(parent, entries, copy, coerce)
with the class definition starting below (a lot of methods I don't want to
write twice are omitted). The important line is the first line of __init__:
sage.matrix.matrix_integer_dense.Matrix_integer_dense.__init__(self,
parent, entries, copy, coerce)
I'd like for Pooling_Matrix to inherit from the sparse class and use the
following as first line of __init__ instead when the parent in the
Pooling_Matrix call is that of a sparse matrix.
sage.matrix.matrix_integer_sparse.Matrix_integer_sparse.__init__(self,
parent, entries, copy, coerce)
The Pooling_Matrix class is adding bookkeeping and diagnostics on top of
matrices, but I need to keep the underlying matrix dense or sparse,
respectively and access the matrix methods/data/etc. through new methods of
Pooling_Matrix that are agnostic to whether the underlying matrix is dense
or sparse, and takes advantage of sparsity to make the operations tractable
in the huge sparse matrix case.
Thanks for any answers!
-Rob
class
Pooling_Matrix_Dense(sage.matrix.matrix_integer_dense.Matrix_integer_dense):
"""
SUMMARY:
Pooling_Matrix extends sage.matrix.constructor.MatrixFactory to
provide functionality for
evaluating the disjunctness of the matrix.
BASIC CONSTRUCTION:
Pooling_Matrix(m), where m is a 2D-array or a nonnegative matrix
over '\mathbf{Z}'.
EXAMPLE CONSTRUCTION::
sage: m = Pooling_Matrix(matrix(ZZ,
[[1,0,0],[0,1,0],[0,0,1],[0,0,1]])
sage: m
[1 0 1]
[0 1 0]
[0 0 1]
INPUT:
- ``mat`` -- a nonnegative integer matrix (over '\mathbb{Z}')
"""
def __init__(self, parent, entries, copy, coerce):
sage.matrix.matrix_integer_dense.Matrix_integer_dense.__init__(self,
parent, entries, copy, coerce)
self.d_lower_bound = -1
self.d_upper_bound = self.ncols()
self.d_max = None
...
On Wednesday, June 5, 2013 12:02:40 PM UTC-7, Rob wrote:
>
> Thanks David, I am getting the following to work.
>
> def make_pool(data):
> rows = len(data)
> cols = len(data[0])
> parent_arg = parent(matrix(ZZ, data))
> return Pool(parent_arg, flatten(data), False, False)
>
> class Pool(sage.matrix.matrix_integer_dense.Matrix_integer_dense):
> def __init__(self, parent, entries, coerce, copy):
>
> sage.matrix.matrix_integer_dense.Matrix_integer_dense.__init__(self,
> parent, entries, coerce, copy)
> self.d_lower_bound = -1
>
>
> m = make_pool([[1,0],[0,1]])
> print m.rows()
> m.d_lower_bound
>
> On Wednesday, June 5, 2013 11:45:39 AM UTC-7, David Roe wrote:
>>
>> The issue is that Matrix_integer_dense has a __cinit__ method, which
>> means that all subclasses must conform to the same inputs for their
>> __init__ methods. So you need to do
>>
>> def __init__(self, parent, entries, coerce, copy):
>> sage.matrix.matrix_integer_dense.Matrix_integer_dense.__init__(self,
>> parent, entries, coerce, copy)
>> ....
>>
>> I would suggest having another function (e.g. pooling_matrix) which takes
>> arguments in the way you want to handle them and creates the appropriate
>> class; this is what the function matrix does. Note that you may need to
>> have a separate parent class if you want to do arithmetic with these
>> matrices, since by default MatrixSpace will create Matrix_integer_dense
>> objects rather than PoolingMatrix objects. See sage.matrix.constructor and
>> sage.matrix.matrix_space.
>> David
>>
>>
>> On Wed, Jun 5, 2013 at 1:53 PM, Rob <ulam...@gmail.com> wrote:
>>
>>> Here is a kluge that is closer to what I want. Can be copied into and
>>> run in a Sage cell. The deficiency in the construction is on lines 6-8
>>> (within commented section).
>>>
>>> class
>>> PoolingMatrix(sage.matrix.matrix_integer_dense.Matrix_integer_dense):
>>> # Example construction:
>>> # a = matrix(ZZ, [[1,1],[2,2]])
>>> # m=PoolingMatrix(parent(a), [1,2,3,4], False, False)
>>> #
>>> # But the construction above is silly. You have to make a matrix of
>>> same dimensions and get its parent before
>>> # constructing the matrix you want.
>>> # I would prefer these construction capabilities:
>>> # m = PoolingMatrix(ZZ, [[1,0,0],[0,1,0]])
>>> # m = PoolingMatrix(ZZ, 2, [1,0,0,0,1,0])
>>> # m = PoolingMatrix(ZZ, 2, 3, [1,0,0,0,1,0])
>>> # And since the ring is always ZZ, maybe the ZZ argument should be
>>> optional.
>>> def __init__(self, parent, data, arg1, arg2):
>>> print data
>>>
>>> sage.matrix.matrix_integer_dense.Matrix_integer_dense.__init__(self,
>>> parent, data, arg1, arg2)
>>> self.d_lower_bound = -1
>>> # parent class, maybe dense integer matrices, provides the
>>> method ncols()
>>> self.d_upper_bound = self.ncols()
>>>
>>> a = matrix(ZZ, [[1,1],[2,2]])
>>> m=PoolingMatrix(parent(a), [1,2,3,4], False, False)
>>> print m.ncols()
>>>
>>>
>>> On Wednesday, June 5, 2013 10:09:55 AM UTC-7, Rob wrote:
>>>>
>>>> Thanks for the responses. Probably the answer is I don't know what
>>>> __init__ method to call within the inheriting __init__ method.
>>>>
>>>> Maybe I'd like to say:
>>>>
>>>> class PoolingMatrix(parent_class):
>>>> def __init__(self, ring_arg, 2D_list_arg):
>>>> parent_class.__init__(self, ring_arg, 2D_list_arg)
>>>> self.my_variable1 = ...
>>>> self.my_variable2 = ...
>>>>
>>>> I'm getting confused by what's a class, what's a constructor, what's a
>>>> parent going into a Matrix_integer_dense.__init__ call, etc. (this is for
>>>> example, not saying Matrix_integer_dense is right). I do know that I want
>>>> dense integer matrices of whatever shape the the 2D_list_arg determines.
>>>>
>>>> I'd like to know what to put in for parent_class in both places above,
>>>> and whether the same thing goes in both places.
>>>>
>>>> Thanks,
>>>> Rob
>>>>
>>>>
>>>>
>>>> On Wednesday, June 5, 2013 9:37:01 AM UTC-7, David Roe wrote:
>>>>>
>>>>> Are you calling some_matrix_thingy.__init__ inside your __init__
>>>>> method?
>>>>> David
>>>>>
>>>>>
>>>>> On Tue, Jun 4, 2013 at 8:10 PM, Rob <ulam...@gmail.com> wrote:
>>>>>
>>>>>> I am trying to make a class PoolingMatrix, which needs to be an
>>>>>> (binary) integer matrix with extra attributes and functions. For
>>>>>> example, I'd like to say:
>>>>>>
>>>>>> sage: m = PoolingMatrix(ZZ, [[1,0,0],[0,1,0],[0,0,1]])
>>>>>> sage: m.nrows()
>>>>>> 2
>>>>>> sage: m.is_disjunct(2) # the 3x3 identity matrix is 2-disjunct
>>>>>> True
>>>>>>
>>>>>> But the init specification for PoolingMatrix is tripping me up. Can
>>>>>> anyone provide a suggestion?
>>>>>>
>>>>>> I'm trying something like:
>>>>>> class PoolingMatrix(some_matrix_**thingy):
>>>>>> def __init__(self, input_ring, input_array):
>>>>>> ...
>>>>>>
>>>>>> Thanks for any assistance
>>>>>> -Rob
>>>>>>
>>>>>> --
>>>>>> You received this message because you are subscribed to the Google
>>>>>> Groups "sage-devel" group.
>>>>>> To unsubscribe from this group and stop receiving emails from it,
>>>>>> send an email to sage-devel+...@googlegroups.**com.
>>>>>> To post to this group, send email to sage-...@googlegroups.com.
>>>>>> Visit this group at
>>>>>> http://groups.google.com/**group/sage-devel?hl=en<http://groups.google.com/group/sage-devel?hl=en>
>>>>>> .
>>>>>> For more options, visit
>>>>>> https://groups.google.com/**groups/opt_out<https://groups.google.com/groups/opt_out>
>>>>>> .
>>>>>>
>>>>>>
>>>>>>
>>>>>
>>
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.
