fix serialization issues

This commit is contained in:
Hans Dembinski
2017-04-27 13:48:21 +02:00
parent 784e482037
commit ce7046d1a4
8 changed files with 245 additions and 220 deletions

View File

@@ -27,16 +27,138 @@
#endif
namespace boost {
namespace histogram {
namespace histogram {
using dynamic_histogram = histogram<Dynamic, default_axes, adaptive_storage<>>;
} // namespace histogram
namespace python {
#ifdef HAVE_NUMPY
auto array_cast = [](python::handle<>& h) {
return python::downcast<PyArrayObject>(h.get());
auto array_cast = [](handle<>& h) {
return downcast<PyArrayObject>(h.get());
};
#endif
#ifdef HAVE_NUMPY
class access {
public:
using mp_int = histogram::adaptive_storage<>::mp_int;
using weight = histogram::adaptive_storage<>::weight;
template <typename T>
using array = histogram::adaptive_storage<>::array<T>;
struct dtype_visitor : public static_visitor<std::pair<int, object>> {
template <typename Array>
std::pair<int, object> operator()(const Array& /*unused*/) const {
std::pair<int, object> p;
p.first = sizeof(typename Array::value_type);
p.second = str("|u") + str(p.first);
return p;
}
std::pair<int, object> operator()(const array<void>& /*unused*/) const {
std::pair<int, object> p;
p.first = sizeof(uint8_t);
p.second = str("|u") + str(p.first);
return p;
}
std::pair<int, object> operator()(const array<mp_int>& /*unused*/) const {
std::pair<int, object> p;
p.first = sizeof(double);
p.second = str("|f") + str(p.first);
return p;
}
std::pair<int, object> operator()(const array<weight>& /*unused*/) const {
std::pair<int, object> p;
p.first = 0; // communicate that the type was array<weight>
p.second = str("|f") + str(sizeof(double));
return p;
}
};
struct data_visitor : public static_visitor<object> {
const list& shapes;
const list& strides;
data_visitor(const list& sh, const list& st) : shapes(sh), strides(st) {}
template <typename Array>
object operator()(const Array& b) const {
return make_tuple(reinterpret_cast<uintptr_t>(b.begin()), true);
}
object operator()(const array<void>& b) const {
// cannot pass non-existent memory to numpy; make new
// zero-initialized uint8 array, and pass it
int dim = len(shapes);
npy_intp shapes2[BOOST_HISTOGRAM_AXIS_LIMIT];
for (int i = 0; i < dim; ++i) {
shapes2[i] = extract<npy_intp>(shapes[i]);
}
handle<> a(PyArray_SimpleNew(dim, shapes2, NPY_UINT8));
for (int i = 0; i < dim; ++i) {
PyArray_STRIDES(array_cast(a))[i] = extract<npy_intp>(strides[i]);
}
auto *buf = static_cast<uint8_t *>(PyArray_DATA(array_cast(a)));
std::fill(buf, buf + b.size, uint8_t(0));
PyArray_CLEARFLAGS(array_cast(a), NPY_ARRAY_WRITEABLE);
return object(a);
}
object operator()(const array<mp_int>& b) const {
// cannot pass cpp_int to numpy; make new
// double array, fill it and pass it
int dim = len(shapes);
npy_intp shapes2[BOOST_HISTOGRAM_AXIS_LIMIT];
for (int i = 0; i < dim; ++i) {
shapes2[i] = extract<npy_intp>(shapes[i]);
}
handle<> a(PyArray_SimpleNew(dim, shapes2, NPY_DOUBLE));
for (int i = 0; i < dim; ++i) {
PyArray_STRIDES(array_cast(a))[i] = extract<npy_intp>(strides[i]);
}
auto *buf = static_cast<double *>(PyArray_DATA(array_cast(a)));
for (std::size_t i = 0; i < b.size; ++i) {
buf[i] = static_cast<double>(b[i]);
}
PyArray_CLEARFLAGS(array_cast(a), NPY_ARRAY_WRITEABLE);
return object(a);
}
};
static object array_interface(const histogram::dynamic_histogram &self) {
dict d;
list shapes;
list strides;
auto &b = self.storage_.buffer_;
auto dtype = apply_visitor(dtype_visitor(), b);
auto stride = dtype.first;
if (stride == 0) { // buffer is weight, needs special treatment
stride = sizeof(double);
strides.append(stride);
stride *= 2;
shapes.append(2);
}
for (unsigned i = 0; i < self.dim(); ++i) {
const auto s = shape(self.axis(i));
shapes.append(s);
strides.append(stride);
stride *= s;
}
if (self.dim() == 0) {
shapes.append(0);
strides.append(stride);
}
d["shape"] = tuple(shapes);
d["strides"] = tuple(strides);
d["typestr"] = dtype.second;
d["data"] = apply_visitor(data_visitor(shapes, strides), b);
return d;
}
};
#endif
} // namespace python
namespace histogram {
struct axis_visitor : public static_visitor<python::object> {
template <typename T> python::object operator()(const T &t) const {
return python::object(t);
@@ -123,8 +245,8 @@ python::object histogram_fill(python::tuple args, python::dict kwargs) {
// exception is thrown automatically if
python::handle<> a(PyArray_FROM_OTF(o.ptr(), NPY_DOUBLE, NPY_ARRAY_IN_ARRAY));
npy_intp *dims = PyArray_DIMS(array_cast(a));
switch (PyArray_NDIM(array_cast(a))) {
npy_intp *dims = PyArray_DIMS(python::array_cast(a));
switch (PyArray_NDIM(python::array_cast(a))) {
case 1:
if (self.dim() > 1) {
PyErr_SetString(PyExc_ValueError, "array has to be two-dimensional");
@@ -150,20 +272,20 @@ python::object histogram_fill(python::tuple args, python::dict kwargs) {
python::handle<> aw(
PyArray_FROM_OTF(ow.ptr(), NPY_DOUBLE, NPY_ARRAY_IN_ARRAY));
if (PyArray_NDIM(array_cast(aw)) != 1) {
if (PyArray_NDIM(python::array_cast(aw)) != 1) {
PyErr_SetString(PyExc_ValueError,
"array has to be one-dimensional");
python::throw_error_already_set();
}
if (PyArray_DIMS(array_cast(aw))[0] != dims[0]) {
if (PyArray_DIMS(python::array_cast(aw))[0] != dims[0]) {
PyErr_SetString(PyExc_ValueError, "sizes do not match");
python::throw_error_already_set();
}
for (unsigned i = 0; i < dims[0]; ++i) {
double *v = reinterpret_cast<double *>(PyArray_GETPTR1(array_cast(a), i));
double *w = reinterpret_cast<double *>(PyArray_GETPTR1(array_cast(aw), i));
double *v = reinterpret_cast<double *>(PyArray_GETPTR1(python::array_cast(a), i));
double *w = reinterpret_cast<double *>(PyArray_GETPTR1(python::array_cast(aw), i));
self.wfill(*w, v, v + self.dim());
}
@@ -173,7 +295,7 @@ python::object histogram_fill(python::tuple args, python::dict kwargs) {
}
} else {
for (unsigned i = 0; i < dims[0]; ++i) {
double *v = reinterpret_cast<double *>(PyArray_GETPTR1(array_cast(a), i));
double *v = reinterpret_cast<double *>(PyArray_GETPTR1(python::array_cast(a), i));
self.fill(v, v + self.dim());
}
}
@@ -273,120 +395,6 @@ std::string histogram_repr(const dynamic_histogram &h) {
return os.str();
}
#ifdef HAVE_NUMPY
struct storage_access {
using mp_int = adaptive_storage<>::mp_int;
using weight = adaptive_storage<>::weight;
template <typename T>
using array = adaptive_storage<>::array<T>;
struct dtype_visitor : public static_visitor<std::pair<int, python::object>> {
template <typename Array>
std::pair<int, python::object> operator()(const Array& /*unused*/) const {
std::pair<int, python::object> p;
p.first = sizeof(typename Array::value_type);
p.second = python::str("|u") + python::str(p.first);
return p;
}
std::pair<int, python::object> operator()(const array<void>& /*unused*/) const {
std::pair<int, python::object> p;
p.first = sizeof(uint8_t);
p.second = python::str("|u") + python::str(p.first);
return p;
}
std::pair<int, python::object> operator()(const array<mp_int>& /*unused*/) const {
std::pair<int, python::object> p;
p.first = sizeof(double);
p.second = python::str("|f") + python::str(p.first);
return p;
}
std::pair<int, python::object> operator()(const array<weight>& /*unused*/) const {
std::pair<int, python::object> p;
p.first = 0; // communicate that the type was array<weight>
p.second = python::str("|f") + python::str(sizeof(double));
return p;
}
};
struct data_visitor : public static_visitor<python::object> {
const python::list& shapes;
const python::list& strides;
data_visitor(const python::list& sh, const python::list& st) : shapes(sh), strides(st) {}
template <typename Array>
python::object operator()(const Array& b) const {
return python::make_tuple(reinterpret_cast<uintptr_t>(b.begin()), true);
}
python::object operator()(const array<void>& b) const {
// cannot pass non-existent memory to numpy; make new
// zero-initialized uint8 array, and pass it
int dim = python::len(shapes);
npy_intp shapes2[BOOST_HISTOGRAM_AXIS_LIMIT];
for (int i = 0; i < dim; ++i) {
shapes2[i] = python::extract<npy_intp>(shapes[i]);
}
python::handle<> a(PyArray_SimpleNew(dim, shapes2, NPY_UINT8));
for (int i = 0; i < dim; ++i) {
PyArray_STRIDES(array_cast(a))[i] = python::extract<npy_intp>(strides[i]);
}
auto *buf = static_cast<uint8_t *>(PyArray_DATA(array_cast(a)));
std::fill(buf, buf + b.size, uint8_t(0));
PyArray_CLEARFLAGS(array_cast(a), NPY_ARRAY_WRITEABLE);
return python::object(a);
}
python::object operator()(const array<mp_int>& b) const {
// cannot pass cpp_int to numpy; make new
// double array, fill it and pass it
int dim = python::len(shapes);
npy_intp shapes2[BOOST_HISTOGRAM_AXIS_LIMIT];
for (int i = 0; i < dim; ++i) {
shapes2[i] = python::extract<npy_intp>(shapes[i]);
}
python::handle<> a(PyArray_SimpleNew(dim, shapes2, NPY_DOUBLE));
for (int i = 0; i < dim; ++i) {
PyArray_STRIDES(array_cast(a))[i] = python::extract<npy_intp>(strides[i]);
}
auto *buf = static_cast<double *>(PyArray_DATA(array_cast(a)));
for (std::size_t i = 0; i < b.size; ++i) {
buf[i] = static_cast<double>(b[i]);
}
PyArray_CLEARFLAGS(array_cast(a), NPY_ARRAY_WRITEABLE);
return python::object(a);
}
};
static python::object array_interface(const dynamic_histogram &self) {
python::dict d;
python::list shapes;
python::list strides;
auto &b = self.storage_.buffer_;
auto dtype = apply_visitor(dtype_visitor(), b);
auto stride = dtype.first;
if (stride == 0) { // buffer is weight, needs special treatment
stride = sizeof(double);
strides.append(stride);
stride *= 2;
shapes.append(2);
}
for (unsigned i = 0; i < self.dim(); ++i) {
const auto s = shape(self.axis(i));
shapes.append(s);
strides.append(stride);
stride *= s;
}
if (self.dim() == 0) {
shapes.append(0);
strides.append(stride);
}
d["shape"] = python::tuple(shapes);
d["strides"] = python::tuple(strides);
d["typestr"] = dtype.second;
d["data"] = apply_visitor(data_visitor(shapes, strides), b);
return d;
}
};
#endif
void register_histogram() {
python::docstring_options dopt(true, true, false);
@@ -399,7 +407,7 @@ void register_histogram() {
// shadowed C++ ctors
.def(python::init<const dynamic_histogram &>())
#ifdef HAVE_NUMPY
.add_property("__array_interface__", &storage_access::array_interface)
.add_property("__array_interface__", &python::access::array_interface)
#endif
.def("__len__", &dynamic_histogram::dim)
.def("__getitem__", histogram_axis)