|
|
#include "CoreWrappers.h"
|
|
|
|
|
|
#include "pywrappers_common.h"
|
|
|
|
|
|
#include <Common/debug.h>
|
|
|
#include <Data/DataSeriesType.h>
|
|
|
#include <Data/IDataProvider.h>
|
|
|
#include <Network/Downloader.h>
|
|
|
#include <Time/TimeController.h>
|
|
|
#include <Variable/Variable2.h>
|
|
|
#include <Variable/VariableController2.h>
|
|
|
#include <pybind11/chrono.h>
|
|
|
#include <pybind11/embed.h>
|
|
|
#include <pybind11/functional.h>
|
|
|
#include <pybind11/numpy.h>
|
|
|
#include <pybind11/operators.h>
|
|
|
#include <pybind11/pybind11.h>
|
|
|
#include <pybind11/stl.h>
|
|
|
#include <pybind11/stl_bind.h>
|
|
|
#include <sstream>
|
|
|
#include <string>
|
|
|
|
|
|
namespace py = pybind11;
|
|
|
using namespace std::chrono;
|
|
|
|
|
|
template<typename T, typename U, bool row_major = true>
|
|
|
void copy_vector(py::array_t<double>& t, py::array_t<double>& values, T& dest_t,
|
|
|
U& dest_values)
|
|
|
{
|
|
|
auto t_view = t.unchecked<1>();
|
|
|
auto values_view = values.unchecked<2>();
|
|
|
for(std::size_t i = 0; i < t.size(); i++)
|
|
|
{
|
|
|
dest_t[i] = t_view[i];
|
|
|
dest_values[i] = {values_view(i, 0), values_view(i, 1), values_view(i, 2)};
|
|
|
}
|
|
|
}
|
|
|
|
|
|
template<typename T, typename U>
|
|
|
void copy_scalar(py::array_t<double>& t, py::array_t<double>& values, T& dest_t,
|
|
|
U& dest_values)
|
|
|
{
|
|
|
auto t_view = t.unchecked<1>();
|
|
|
if(values.ndim() == 1)
|
|
|
{
|
|
|
auto values_view = values.unchecked<1>();
|
|
|
for(std::size_t i = 0; i < t.size(); i++)
|
|
|
{
|
|
|
dest_t[i] = t_view[i];
|
|
|
dest_values[i] = values_view[i];
|
|
|
}
|
|
|
}
|
|
|
else if(values.ndim() == 2 && values.shape(1) == 1)
|
|
|
{
|
|
|
auto values_view = values.unchecked<2>();
|
|
|
for(std::size_t i = 0; i < t.size(); i++)
|
|
|
{
|
|
|
dest_t[i] = t_view[i];
|
|
|
dest_values[i] = values_view(i, 0);
|
|
|
}
|
|
|
}
|
|
|
}
|
|
|
template<typename T, typename U>
|
|
|
void copy_multicomp(py::array_t<double>& t, py::array_t<double>& values,
|
|
|
T& dest_t, U& dest_values)
|
|
|
{
|
|
|
auto t_view = t.unchecked<1>();
|
|
|
auto values_view = values.unchecked<2>();
|
|
|
const auto width = values.shape(1);
|
|
|
for(std::size_t i = 0; i < t.size(); i++)
|
|
|
{
|
|
|
dest_t[i] = t_view[i];
|
|
|
for(int j = 0; j < width; j++)
|
|
|
{
|
|
|
dest_values[i * width + j] = values_view(i, j);
|
|
|
}
|
|
|
}
|
|
|
}
|
|
|
|
|
|
template<typename T, typename U>
|
|
|
void copy_spectro(py::array_t<double>& t, py::array_t<double>& y,
|
|
|
py::array_t<double>& values, T& dest_t, T& dest_y,
|
|
|
U& dest_values)
|
|
|
{
|
|
|
auto t_view = t.unchecked<1>();
|
|
|
auto y_view = y.unchecked<1>();
|
|
|
auto values_view = values.unchecked<2>();
|
|
|
const auto width = values.shape(1);
|
|
|
for(std::size_t i = 0; i < y.size(); i++)
|
|
|
{
|
|
|
dest_y[i] = y_view[i];
|
|
|
}
|
|
|
for(std::size_t i = 0; i < t.size(); i++)
|
|
|
{
|
|
|
dest_t[i] = t_view[i];
|
|
|
for(int j = 0; j < width; j++)
|
|
|
{
|
|
|
dest_values[i * width + j] = values_view(i, j);
|
|
|
}
|
|
|
}
|
|
|
}
|
|
|
|
|
|
PYBIND11_MODULE(pysciqlopcore, m)
|
|
|
{
|
|
|
pybind11::bind_vector<std::vector<double>>(m, "VectorDouble");
|
|
|
|
|
|
py::enum_<DataSeriesType>(m, "DataSeriesType")
|
|
|
.value("SCALAR", DataSeriesType::SCALAR)
|
|
|
.value("SPECTROGRAM", DataSeriesType::SPECTROGRAM)
|
|
|
.value("VECTOR", DataSeriesType::VECTOR)
|
|
|
.value("MULTICOMPONENT", DataSeriesType::MULTICOMPONENT)
|
|
|
.value("NONE", DataSeriesType::NONE)
|
|
|
.export_values();
|
|
|
|
|
|
py::class_<Response>(m, "Response")
|
|
|
.def("status_code", &Response::status_code);
|
|
|
|
|
|
py::class_<Downloader>(m, "Downloader")
|
|
|
.def_static("get", Downloader::get)
|
|
|
.def_static("getAsync", Downloader::getAsync)
|
|
|
.def_static("downloadFinished", Downloader::downloadFinished);
|
|
|
|
|
|
py::class_<IDataProvider, std::shared_ptr<IDataProvider>>(m, "IDataProvider");
|
|
|
|
|
|
py::class_<TimeSeries::ITimeSerie, std::shared_ptr<TimeSeries::ITimeSerie>>(
|
|
|
m, "ITimeSerie")
|
|
|
.def_property_readonly(
|
|
|
"size", [](const TimeSeries::ITimeSerie& ts) { return ts.size(); })
|
|
|
.def("__len__",
|
|
|
[](const TimeSeries::ITimeSerie& ts) { return ts.size(); })
|
|
|
.def_property_readonly(
|
|
|
"shape", [](const TimeSeries::ITimeSerie& ts) { return ts.shape(); })
|
|
|
.def_property_readonly(
|
|
|
"t",
|
|
|
[](TimeSeries::ITimeSerie& ts) -> decltype(ts.axis(0))& {
|
|
|
return ts.axis(0);
|
|
|
},
|
|
|
py::return_value_policy::reference)
|
|
|
.def(
|
|
|
"axis",
|
|
|
[](TimeSeries::ITimeSerie& ts, unsigned int index)
|
|
|
-> decltype(ts.axis(0))& { return ts.axis(index); },
|
|
|
py::return_value_policy::reference);
|
|
|
|
|
|
py::class_<ScalarTimeSerie, TimeSeries::ITimeSerie,
|
|
|
std::shared_ptr<ScalarTimeSerie>>(m, "ScalarTimeSerie")
|
|
|
.def(py::init<>())
|
|
|
.def(py::init<std::size_t>())
|
|
|
.def(py::init([](py::array_t<double> t, py::array_t<double> values) {
|
|
|
assert(t.size() == values.size());
|
|
|
ScalarTimeSerie::axis_t _t(t.size());
|
|
|
ScalarTimeSerie::axis_t _values(t.size());
|
|
|
copy_scalar(t, values, _t, _values);
|
|
|
return ScalarTimeSerie(_t, _values);
|
|
|
}))
|
|
|
.def("__getitem__",
|
|
|
[](ScalarTimeSerie& ts, std::size_t key) { return ts[key]; })
|
|
|
.def("__setitem__", [](ScalarTimeSerie& ts, std::size_t key,
|
|
|
double value) { *(ts.begin() + key) = value; });
|
|
|
|
|
|
py::class_<VectorTimeSerie::raw_value_type>(m, "vector")
|
|
|
.def(py::init<>())
|
|
|
.def(py::init<double, double, double>())
|
|
|
.def("__repr__", __repr__<VectorTimeSerie::raw_value_type>)
|
|
|
.def_readwrite("x", &VectorTimeSerie::raw_value_type::x)
|
|
|
.def_readwrite("y", &VectorTimeSerie::raw_value_type::y)
|
|
|
.def_readwrite("z", &VectorTimeSerie::raw_value_type::z);
|
|
|
|
|
|
py::class_<VectorTimeSerie, TimeSeries::ITimeSerie,
|
|
|
std::shared_ptr<VectorTimeSerie>>(m, "VectorTimeSerie")
|
|
|
.def(py::init<>())
|
|
|
.def(py::init<std::size_t>())
|
|
|
.def(py::init([](py::array_t<double> t, py::array_t<double> values) {
|
|
|
assert(t.size() * 3 == values.size());
|
|
|
VectorTimeSerie::axis_t _t(t.size());
|
|
|
VectorTimeSerie::container_type<VectorTimeSerie::raw_value_type>
|
|
|
_values(t.size());
|
|
|
copy_vector(t, values, _t, _values);
|
|
|
return VectorTimeSerie(_t, _values);
|
|
|
}))
|
|
|
.def(
|
|
|
"__getitem__",
|
|
|
[](VectorTimeSerie& ts, std::size_t key)
|
|
|
-> VectorTimeSerie::raw_value_type& { return ts[key]; },
|
|
|
py::return_value_policy::reference)
|
|
|
.def("__setitem__", [](VectorTimeSerie& ts, std::size_t key,
|
|
|
VectorTimeSerie::raw_value_type value) {
|
|
|
*(ts.begin() + key) = value;
|
|
|
});
|
|
|
|
|
|
py::class_<MultiComponentTimeSerie::iterator_t>(m,
|
|
|
"MultiComponentTimeSerieItem")
|
|
|
.def("__getitem__", [](MultiComponentTimeSerie::iterator_t& self,
|
|
|
std::size_t key) { return (*self)[key]; })
|
|
|
.def("__setitem__",
|
|
|
[](MultiComponentTimeSerie::iterator_t& self, std::size_t key,
|
|
|
double value) { (*self)[key] = value; });
|
|
|
|
|
|
py::class_<MultiComponentTimeSerie, TimeSeries::ITimeSerie,
|
|
|
std::shared_ptr<MultiComponentTimeSerie>>(
|
|
|
m, "MultiComponentTimeSerie")
|
|
|
.def(py::init<>())
|
|
|
.def(py::init<const std::vector<std::size_t>>())
|
|
|
.def(py::init([](py::array_t<double> t, py::array_t<double> values) {
|
|
|
assert((t.size() < values.size()) |
|
|
|
(t.size() == 0)); // TODO check geometry
|
|
|
MultiComponentTimeSerie::axis_t _t(t.size());
|
|
|
MultiComponentTimeSerie::container_type<
|
|
|
MultiComponentTimeSerie::raw_value_type>
|
|
|
_values(values.size());
|
|
|
copy_multicomp(t, values, _t, _values);
|
|
|
std::vector<std::size_t> shape;
|
|
|
shape.push_back(values.shape(0));
|
|
|
shape.push_back(values.shape(1));
|
|
|
return MultiComponentTimeSerie(_t, _values, shape);
|
|
|
}))
|
|
|
.def("__getitem__",
|
|
|
[](MultiComponentTimeSerie& ts,
|
|
|
std::size_t key) -> MultiComponentTimeSerie::iterator_t {
|
|
|
return ts.begin() + key;
|
|
|
});
|
|
|
|
|
|
py::class_<SpectrogramTimeSerie::iterator_t>(m, "SpectrogramTimeSerieItem")
|
|
|
.def("__getitem__", [](SpectrogramTimeSerie::iterator_t& self,
|
|
|
std::size_t key) { return (*self)[key]; })
|
|
|
.def("__setitem__",
|
|
|
[](SpectrogramTimeSerie::iterator_t& self, std::size_t key,
|
|
|
double value) { (*self)[key] = value; });
|
|
|
|
|
|
py::class_<SpectrogramTimeSerie, TimeSeries::ITimeSerie,
|
|
|
std::shared_ptr<SpectrogramTimeSerie>>(m, "SpectrogramTimeSerie")
|
|
|
.def(py::init<>())
|
|
|
.def(py::init<const std::vector<std::size_t>>())
|
|
|
.def(py::init([](py::array_t<double> t, py::array_t<double> y,
|
|
|
py::array_t<double> values, double min_sampling,
|
|
|
double max_sampling, bool y_is_log) {
|
|
|
if(t.size() >= values.size() and t.size() != 0)
|
|
|
SCIQLOP_ERROR(decltype(py::self), "Doesn't look like a Spectrogram");
|
|
|
if(y.size() != values.shape(1))
|
|
|
SCIQLOP_ERROR(decltype(py::self),
|
|
|
"Y axis size and data shape are incompatible");
|
|
|
SpectrogramTimeSerie::axis_t _t(t.size());
|
|
|
SpectrogramTimeSerie::axis_t _y(y.size());
|
|
|
SpectrogramTimeSerie::container_type<
|
|
|
SpectrogramTimeSerie::raw_value_type>
|
|
|
_values(values.size());
|
|
|
copy_spectro(t, y, values, _t, _y, _values);
|
|
|
std::vector<std::size_t> shape;
|
|
|
shape.push_back(values.shape(0));
|
|
|
shape.push_back(values.shape(1));
|
|
|
return SpectrogramTimeSerie(std::move(_t), std::move(_y),
|
|
|
std::move(_values), shape, min_sampling,
|
|
|
max_sampling, y_is_log);
|
|
|
}))
|
|
|
.def("__getitem__",
|
|
|
[](SpectrogramTimeSerie& ts,
|
|
|
std::size_t key) -> SpectrogramTimeSerie::iterator_t {
|
|
|
return ts.begin() + key;
|
|
|
});
|
|
|
|
|
|
py::class_<Variable2, std::shared_ptr<Variable2>>(m, "Variable2")
|
|
|
.def(py::init<const QString&>())
|
|
|
.def_property("name", &Variable2::name, &Variable2::setName)
|
|
|
.def_property_readonly("range", &Variable2::range)
|
|
|
.def_property_readonly("nbPoints", &Variable2::nbPoints)
|
|
|
.def_property_readonly(
|
|
|
"data",
|
|
|
[](Variable2& var) -> std::shared_ptr<TimeSeries::ITimeSerie> {
|
|
|
return var.data();
|
|
|
})
|
|
|
.def("set_data",
|
|
|
[](Variable2& var, std::vector<TimeSeries::ITimeSerie*> ts_list,
|
|
|
const DateTimeRange& range) { var.setData(ts_list, range); })
|
|
|
.def("__len__", &Variable2::nbPoints)
|
|
|
.def("__repr__", __repr__<Variable2>);
|
|
|
|
|
|
py::class_<DateTimeRange>(m, "SqpRange")
|
|
|
//.def("fromDateTime", &DateTimeRange::fromDateTime,
|
|
|
// py::return_value_policy::move)
|
|
|
.def(py::init([](double start, double stop) {
|
|
|
return DateTimeRange{start, stop};
|
|
|
}))
|
|
|
.def(py::init(
|
|
|
[](system_clock::time_point start, system_clock::time_point stop) {
|
|
|
double start_ =
|
|
|
0.001 *
|
|
|
duration_cast<milliseconds>(start.time_since_epoch()).count();
|
|
|
double stop_ =
|
|
|
0.001 *
|
|
|
duration_cast<milliseconds>(stop.time_since_epoch()).count();
|
|
|
return DateTimeRange{start_, stop_};
|
|
|
}))
|
|
|
.def_property_readonly("start",
|
|
|
[](const DateTimeRange& range) {
|
|
|
return system_clock::from_time_t(range.m_TStart);
|
|
|
})
|
|
|
.def_property_readonly("stop",
|
|
|
[](const DateTimeRange& range) {
|
|
|
return system_clock::from_time_t(range.m_TEnd);
|
|
|
})
|
|
|
.def("__repr__", __repr__<DateTimeRange>);
|
|
|
}
|
|
|
|