Homec4science
Diffusion akantu (features/68-automatic-differentiation) Browse

rAKA/python/swig/aka_array.i
0de4dd0d2bbdfeatures/68-automatic…

Last Change
Raw File

aka_array.i
View Options

/**
* @file aka_array.i
*
* @author Guillaume Anciaux <guillaume.anciaux@epfl.ch>
* @author Nicolas Richart <nicolas.richart@epfl.ch>
*
* @date creation: Fri Dec 12 2014
* @date last modification: Wed Nov 11 2015
*
* @brief wrapper for arrays
*
* @section LICENSE
*
* Copyright (©) 2015 EPFL (Ecole Polytechnique Fédérale de Lausanne) Laboratory
* (LSMS - Laboratoire de Simulation en Mécanique des Solides)
*
* Akantu is free software: you can redistribute it and/or modify it under the
* terms of the GNU Lesser General Public License as published by the Free
* Software Foundation, either version 3 of the License, or (at your option) any
* later version.
*
* Akantu is distributed in the hope that it will be useful, but WITHOUT ANY
* WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
* A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more
* details.
*
* You should have received a copy of the GNU Lesser General Public License
* along with Akantu. If not, see <http://www.gnu.org/licenses/>.
*
*/
%{
#define SWIG_FILE_WITH_INIT
#include "aka_array.hh"
#include "aka_types.hh"
%}
%include "typemaps.i"
namespace akantu {
%ignore Array::operator=;
%ignore Array::operator[];
%ignore Array::operator();
%ignore Array::set;
%ignore Array::begin;
%ignore Array::end;
%ignore Array::begin_reinterpret;
%ignore Array::end_reinterpret;
%ignore ArrayBase::getSize;
%ignore Array::operator*=;
%ignore Array::operator*;
};
%include "aka_array.hh"
namespace akantu {
%ignore TensorProxy::operator=;
%ignore TensorProxy::operator[];
%ignore TensorProxy::operator();
%ignore Tensor3Proxy::operator=;
%ignore Tensor3Proxy::operator[];
%ignore Tensor3Proxy::operator();
%ignore TensorStorage::operator=;
%ignore TensorStorage::operator[];
%ignore TensorStorage::operator();
%ignore VectorProxy::operator=;
%ignore VectorProxy::operator[];
%ignore VectorProxy::operator();
%ignore MatrixProxy::operator=;
%ignore MatrixProxy::operator[];
%ignore MatrixProxy::operator();
%ignore Matrix::operator=;
%ignore Matrix::operator[];
%ignore Matrix::operator();
%ignore Tensor3::operator=;
%ignore Tensor3::operator[];
%ignore Tensor3::operator();
%ignore Vector::operator=;
%ignore Vector::operator[];
%ignore Vector::operator();
%ignore Vector::solve;
};
%include "aka_types.hh"
namespace akantu {
%template(RArray) Array<akantu::Real, true>;
%template(UArray) Array<akantu::UInt, true>;
%template(BArray) Array<bool, true>;
%template(RVector) Vector<akantu::Real>;
};
%include "numpy.i"
%init %{
import_array();
%}
%inline %{
namespace akantu {
template <typename T> class ArrayForPython : public Array<T> {
public:
ArrayForPython(T * wrapped_memory, UInt size_ = 0, UInt nb_component = 1,
const ID & id = "")
: Array<T>(0, nb_component, id) {
this->values = wrapped_memory;
this->size_ = size_;
};
~ArrayForPython() { this->values = NULL; };
void resize(UInt new_size) {
AKANTU_DEBUG_ASSERT(this->size_ == new_size,
"cannot resize a temporary vector");
}
};
}
template <typename T> int getPythonDataTypeCode() {
AKANTU_EXCEPTION("undefined type");
}
template <> int getPythonDataTypeCode<bool>() {
int data_typecode = NPY_NOTYPE;
size_t s = sizeof(bool);
switch (s) {
case 1:
data_typecode = NPY_BOOL;
break;
case 2:
data_typecode = NPY_UINT16;
break;
case 4:
data_typecode = NPY_UINT32;
break;
case 8:
data_typecode = NPY_UINT64;
break;
}
return data_typecode;
}
template <> int getPythonDataTypeCode<double>() { return NPY_DOUBLE; }
template <> int getPythonDataTypeCode<long double>() { return NPY_LONGDOUBLE; }
template <> int getPythonDataTypeCode<float>() { return NPY_FLOAT; }
template <> int getPythonDataTypeCode<unsigned long>() {
int data_typecode = NPY_NOTYPE;
size_t s = sizeof(unsigned long);
switch (s) {
case 2:
data_typecode = NPY_UINT16;
break;
case 4:
data_typecode = NPY_UINT32;
break;
case 8:
data_typecode = NPY_UINT64;
break;
}
return data_typecode;
}
template <> int getPythonDataTypeCode<akantu::UInt>() {
int data_typecode = NPY_NOTYPE;
size_t s = sizeof(akantu::UInt);
switch (s) {
case 2:
data_typecode = NPY_UINT16;
break;
case 4:
data_typecode = NPY_UINT32;
break;
case 8:
data_typecode = NPY_UINT64;
break;
}
return data_typecode;
}
template <> int getPythonDataTypeCode<int>() {
int data_typecode = NPY_NOTYPE;
size_t s = sizeof(int);
switch (s) {
case 2:
data_typecode = NPY_INT16;
break;
case 4:
data_typecode = NPY_INT32;
break;
case 8:
data_typecode = NPY_INT64;
break;
}
return data_typecode;
}
int getSizeOfPythonType(int type_num) {
switch (type_num) {
case NPY_INT16:
return 2;
break;
case NPY_UINT16:
return 2;
break;
case NPY_INT32:
return 4;
break;
case NPY_UINT32:
return 4;
break;
case NPY_INT64:
return 8;
break;
case NPY_UINT64:
return 8;
break;
case NPY_FLOAT:
return sizeof(float);
break;
case NPY_DOUBLE:
return sizeof(double);
break;
case NPY_LONGDOUBLE:
return sizeof(long double);
break;
}
return 0;
}
std::string getPythonTypeName(int type_num) {
switch (type_num) {
case NPY_INT16:
return "NPY_INT16";
break;
case NPY_UINT16:
return "NPY_UINT16";
break;
case NPY_INT32:
return "NPY_INT32";
break;
case NPY_UINT32:
return "NPY_UINT32";
break;
case NPY_INT64:
return "NPY_INT64";
break;
case NPY_UINT64:
return "NPY_UINT64";
break;
case NPY_FLOAT:
return "NPY_FLOAT";
break;
case NPY_DOUBLE:
return "NPY_DOUBLE";
break;
case NPY_LONGDOUBLE:
return "NPY_LONGDOUBLE";
break;
}
return 0;
}
template <typename T> void checkDataType(int type_num) {
AKANTU_DEBUG_ASSERT(
type_num == getPythonDataTypeCode<T>(),
"incompatible types between numpy and input function: "
<< type_num << " != " << getPythonDataTypeCode<T>() << std::endl
<< getSizeOfPythonType(type_num) << " != " << sizeof(T) << std::endl
<< "The numpy array is of type " << getPythonTypeName(type_num));
}
%}
%define %akantu_array_typemaps(DATA_TYPE)
%typemap(out, fragment="NumPy_Fragments") (akantu::Array< DATA_TYPE > &)
{
int data_typecode = getPythonDataTypeCode<DATA_TYPE>();
npy_intp dims[2] = {npy_intp($1->size()), npy_intp($1->getNbComponent())};
PyObject * obj =
PyArray_SimpleNewFromData(2, dims, data_typecode, $1->storage());
PyArrayObject * array = (PyArrayObject *)obj;
if (!array)
SWIG_fail;
$result = SWIG_Python_AppendOutput($result, obj);
}
%typemap(in) akantu::Array< DATA_TYPE > &
{
if (!PyArray_Check($input)) {
AKANTU_EXCEPTION("incompatible input which is not a numpy");
} else {
PyArray_Descr * numpy_type =
(PyArray_Descr *)PyArray_DESCR((PyArrayObject *)$input);
int type_num = numpy_type->type_num;
checkDataType<DATA_TYPE>(type_num);
UInt _n = PyArray_NDIM((PyArrayObject *)$input);
if (_n != 2)
AKANTU_EXCEPTION("incompatible numpy dimension " << _n);
npy_intp * ndims = PyArray_DIMS((PyArrayObject *)$input);
akantu::UInt sz = ndims[0];
akantu::UInt nb_components = ndims[1];
PyArrayIterObject * iter = (PyArrayIterObject *)PyArray_IterNew($input);
if (iter == NULL)
AKANTU_EXCEPTION("Python internal error");
$1 = new akantu::ArrayForPython<DATA_TYPE>((DATA_TYPE *)(iter->dataptr), sz,
nb_components,
"tmp_array_for_python");
}
}
%enddef
%akantu_array_typemaps(double )
%akantu_array_typemaps(float )
%akantu_array_typemaps(unsigned int)
%akantu_array_typemaps(unsigned long)
%akantu_array_typemaps(int )
%akantu_array_typemaps(bool )
c4science · Help