@@ -1,41 +1,43 | |||
|
1 | 1 | #ifndef SCIQLOP_SPECTROGRAMTIMESERIE_H |
|
2 | 2 | #define SCIQLOP_SPECTROGRAMTIMESERIE_H |
|
3 | 3 | |
|
4 | 4 | #include "CoreGlobal.h" |
|
5 | 5 | |
|
6 | 6 | #include <TimeSeries.h> |
|
7 | 7 | #include <cmath> |
|
8 | 8 | |
|
9 | 9 | class SCIQLOP_CORE_EXPORT SpectrogramTimeSerie |
|
10 | 10 | : public TimeSeries::TimeSerie<double, SpectrogramTimeSerie, 2> |
|
11 | 11 | { |
|
12 | 12 | public: |
|
13 | 13 | double min_sampling = std::nan(""); |
|
14 | 14 | double max_sampling = std::nan(""); |
|
15 | bool y_is_log = true; | |
|
15 | 16 | using item_t = |
|
16 | 17 | decltype(std::declval< |
|
17 | 18 | TimeSeries::TimeSerie<double, SpectrogramTimeSerie, 2>>()[0]); |
|
18 | 19 | |
|
19 | 20 | using iterator_t = decltype( |
|
20 | 21 | std::declval<TimeSeries::TimeSerie<double, SpectrogramTimeSerie, 2>>() |
|
21 | 22 | .begin()); |
|
22 | 23 | |
|
23 | 24 | SpectrogramTimeSerie() {} |
|
24 | 25 | SpectrogramTimeSerie(SpectrogramTimeSerie::axis_t&& t, |
|
25 | 26 | SpectrogramTimeSerie::axis_t&& y, |
|
26 | 27 | SpectrogramTimeSerie::container_type< |
|
27 | 28 | SpectrogramTimeSerie::raw_value_type>&& values, |
|
28 | 29 | std::vector<std::size_t>& shape, double min_sampling, |
|
29 | double max_sampling) | |
|
30 | double max_sampling, bool y_is_log = true) | |
|
30 | 31 | : TimeSeries::TimeSerie<double, SpectrogramTimeSerie, 2>(t, values, |
|
31 | 32 | shape), |
|
32 | min_sampling{min_sampling}, max_sampling{max_sampling} | |
|
33 | min_sampling{min_sampling}, max_sampling{max_sampling}, y_is_log{ | |
|
34 | y_is_log} | |
|
33 | 35 | { |
|
34 | 36 | _axes[1] = y; |
|
35 | 37 | } |
|
36 | 38 | |
|
37 | 39 | ~SpectrogramTimeSerie() = default; |
|
38 | 40 | using TimeSerie::TimeSerie; |
|
39 | 41 | }; |
|
40 | 42 | |
|
41 | 43 | #endif // SCIQLOP_SPECTROGRAMTIMESERIE_H |
@@ -1,302 +1,302 | |||
|
1 | 1 | #include "CoreWrappers.h" |
|
2 | 2 | |
|
3 | 3 | #include "pywrappers_common.h" |
|
4 | 4 | |
|
5 | 5 | #include <Common/debug.h> |
|
6 | 6 | #include <Data/DataSeriesType.h> |
|
7 | 7 | #include <Data/IDataProvider.h> |
|
8 | 8 | #include <Network/Downloader.h> |
|
9 | 9 | #include <Time/TimeController.h> |
|
10 | 10 | #include <Variable/Variable2.h> |
|
11 | 11 | #include <Variable/VariableController2.h> |
|
12 | 12 | #include <pybind11/chrono.h> |
|
13 | 13 | #include <pybind11/embed.h> |
|
14 | 14 | #include <pybind11/functional.h> |
|
15 | 15 | #include <pybind11/numpy.h> |
|
16 | 16 | #include <pybind11/operators.h> |
|
17 | 17 | #include <pybind11/pybind11.h> |
|
18 | 18 | #include <pybind11/stl.h> |
|
19 | 19 | #include <pybind11/stl_bind.h> |
|
20 | 20 | #include <sstream> |
|
21 | 21 | #include <string> |
|
22 | 22 | |
|
23 | 23 | namespace py = pybind11; |
|
24 | 24 | using namespace std::chrono; |
|
25 | 25 | |
|
26 | 26 | template<typename T, typename U, bool row_major = true> |
|
27 | 27 | void copy_vector(py::array_t<double>& t, py::array_t<double>& values, T& dest_t, |
|
28 | 28 | U& dest_values) |
|
29 | 29 | { |
|
30 | 30 | auto t_view = t.unchecked<1>(); |
|
31 | 31 | auto values_view = values.unchecked<2>(); |
|
32 | 32 | for(std::size_t i = 0; i < t.size(); i++) |
|
33 | 33 | { |
|
34 | 34 | dest_t[i] = t_view[i]; |
|
35 | 35 | dest_values[i] = {values_view(i, 0), values_view(i, 1), values_view(i, 2)}; |
|
36 | 36 | } |
|
37 | 37 | } |
|
38 | 38 | |
|
39 | 39 | template<typename T, typename U> |
|
40 | 40 | void copy_scalar(py::array_t<double>& t, py::array_t<double>& values, T& dest_t, |
|
41 | 41 | U& dest_values) |
|
42 | 42 | { |
|
43 | 43 | auto t_view = t.unchecked<1>(); |
|
44 | 44 | if(values.ndim() == 1) |
|
45 | 45 | { |
|
46 | 46 | auto values_view = values.unchecked<1>(); |
|
47 | 47 | for(std::size_t i = 0; i < t.size(); i++) |
|
48 | 48 | { |
|
49 | 49 | dest_t[i] = t_view[i]; |
|
50 | 50 | dest_values[i] = values_view[i]; |
|
51 | 51 | } |
|
52 | 52 | } |
|
53 | 53 | else if(values.ndim() == 2 && values.shape(1) == 1) |
|
54 | 54 | { |
|
55 | 55 | auto values_view = values.unchecked<2>(); |
|
56 | 56 | for(std::size_t i = 0; i < t.size(); i++) |
|
57 | 57 | { |
|
58 | 58 | dest_t[i] = t_view[i]; |
|
59 | 59 | dest_values[i] = values_view(i, 0); |
|
60 | 60 | } |
|
61 | 61 | } |
|
62 | 62 | } |
|
63 | 63 | template<typename T, typename U> |
|
64 | 64 | void copy_multicomp(py::array_t<double>& t, py::array_t<double>& values, |
|
65 | 65 | T& dest_t, U& dest_values) |
|
66 | 66 | { |
|
67 | 67 | auto t_view = t.unchecked<1>(); |
|
68 | 68 | auto values_view = values.unchecked<2>(); |
|
69 | 69 | const auto width = values.shape(1); |
|
70 | 70 | for(std::size_t i = 0; i < t.size(); i++) |
|
71 | 71 | { |
|
72 | 72 | dest_t[i] = t_view[i]; |
|
73 | 73 | for(int j = 0; j < width; j++) |
|
74 | 74 | { |
|
75 | 75 | dest_values[i * width + j] = values_view(i, j); |
|
76 | 76 | } |
|
77 | 77 | } |
|
78 | 78 | } |
|
79 | 79 | |
|
80 | 80 | template<typename T, typename U> |
|
81 | 81 | void copy_spectro(py::array_t<double>& t, py::array_t<double>& y, |
|
82 | 82 | py::array_t<double>& values, T& dest_t, T& dest_y, |
|
83 | 83 | U& dest_values) |
|
84 | 84 | { |
|
85 | 85 | auto t_view = t.unchecked<1>(); |
|
86 | 86 | auto y_view = y.unchecked<1>(); |
|
87 | 87 | auto values_view = values.unchecked<2>(); |
|
88 | 88 | const auto width = values.shape(1); |
|
89 | 89 | for(std::size_t i = 0; i < y.size(); i++) |
|
90 | 90 | { |
|
91 | 91 | dest_y[i] = y_view[i]; |
|
92 | 92 | } |
|
93 | 93 | for(std::size_t i = 0; i < t.size(); i++) |
|
94 | 94 | { |
|
95 | 95 | dest_t[i] = t_view[i]; |
|
96 | 96 | for(int j = 0; j < width; j++) |
|
97 | 97 | { |
|
98 | 98 | dest_values[i * width + j] = values_view(i, j); |
|
99 | 99 | } |
|
100 | 100 | } |
|
101 | 101 | } |
|
102 | 102 | |
|
103 | 103 | PYBIND11_MODULE(pysciqlopcore, m) |
|
104 | 104 | { |
|
105 | 105 | pybind11::bind_vector<std::vector<double>>(m, "VectorDouble"); |
|
106 | 106 | |
|
107 | 107 | py::enum_<DataSeriesType>(m, "DataSeriesType") |
|
108 | 108 | .value("SCALAR", DataSeriesType::SCALAR) |
|
109 | 109 | .value("SPECTROGRAM", DataSeriesType::SPECTROGRAM) |
|
110 | 110 | .value("VECTOR", DataSeriesType::VECTOR) |
|
111 | 111 | .value("MULTICOMPONENT", DataSeriesType::MULTICOMPONENT) |
|
112 | 112 | .value("NONE", DataSeriesType::NONE) |
|
113 | 113 | .export_values(); |
|
114 | 114 | |
|
115 | 115 | py::class_<Response>(m, "Response") |
|
116 | 116 | .def("status_code", &Response::status_code); |
|
117 | 117 | |
|
118 | 118 | py::class_<Downloader>(m, "Downloader") |
|
119 | 119 | .def_static("get", Downloader::get) |
|
120 | 120 | .def_static("getAsync", Downloader::getAsync) |
|
121 | 121 | .def_static("downloadFinished", Downloader::downloadFinished); |
|
122 | 122 | |
|
123 | 123 | py::class_<IDataProvider, std::shared_ptr<IDataProvider>>(m, "IDataProvider"); |
|
124 | 124 | |
|
125 | 125 | py::class_<TimeSeries::ITimeSerie, std::shared_ptr<TimeSeries::ITimeSerie>>( |
|
126 | 126 | m, "ITimeSerie") |
|
127 | 127 | .def_property_readonly( |
|
128 | 128 | "size", [](const TimeSeries::ITimeSerie& ts) { return ts.size(); }) |
|
129 | 129 | .def("__len__", |
|
130 | 130 | [](const TimeSeries::ITimeSerie& ts) { return ts.size(); }) |
|
131 | 131 | .def_property_readonly( |
|
132 | 132 | "shape", [](const TimeSeries::ITimeSerie& ts) { return ts.shape(); }) |
|
133 | 133 | .def_property_readonly( |
|
134 | 134 | "t", |
|
135 | 135 | [](TimeSeries::ITimeSerie& ts) -> decltype(ts.axis(0))& { |
|
136 | 136 | return ts.axis(0); |
|
137 | 137 | }, |
|
138 | 138 | py::return_value_policy::reference) |
|
139 | 139 | .def( |
|
140 | 140 | "axis", |
|
141 | 141 | [](TimeSeries::ITimeSerie& ts, unsigned int index) |
|
142 | 142 | -> decltype(ts.axis(0))& { return ts.axis(index); }, |
|
143 | 143 | py::return_value_policy::reference); |
|
144 | 144 | |
|
145 | 145 | py::class_<ScalarTimeSerie, TimeSeries::ITimeSerie, |
|
146 | 146 | std::shared_ptr<ScalarTimeSerie>>(m, "ScalarTimeSerie") |
|
147 | 147 | .def(py::init<>()) |
|
148 | 148 | .def(py::init<std::size_t>()) |
|
149 | 149 | .def(py::init([](py::array_t<double> t, py::array_t<double> values) { |
|
150 | 150 | assert(t.size() == values.size()); |
|
151 | 151 | ScalarTimeSerie::axis_t _t(t.size()); |
|
152 | 152 | ScalarTimeSerie::axis_t _values(t.size()); |
|
153 | 153 | copy_scalar(t, values, _t, _values); |
|
154 | 154 | return ScalarTimeSerie(_t, _values); |
|
155 | 155 | })) |
|
156 | 156 | .def("__getitem__", |
|
157 | 157 | [](ScalarTimeSerie& ts, std::size_t key) { return ts[key]; }) |
|
158 | 158 | .def("__setitem__", [](ScalarTimeSerie& ts, std::size_t key, |
|
159 | 159 | double value) { *(ts.begin() + key) = value; }); |
|
160 | 160 | |
|
161 | 161 | py::class_<VectorTimeSerie::raw_value_type>(m, "vector") |
|
162 | 162 | .def(py::init<>()) |
|
163 | 163 | .def(py::init<double, double, double>()) |
|
164 | 164 | .def("__repr__", __repr__<VectorTimeSerie::raw_value_type>) |
|
165 | 165 | .def_readwrite("x", &VectorTimeSerie::raw_value_type::x) |
|
166 | 166 | .def_readwrite("y", &VectorTimeSerie::raw_value_type::y) |
|
167 | 167 | .def_readwrite("z", &VectorTimeSerie::raw_value_type::z); |
|
168 | 168 | |
|
169 | 169 | py::class_<VectorTimeSerie, TimeSeries::ITimeSerie, |
|
170 | 170 | std::shared_ptr<VectorTimeSerie>>(m, "VectorTimeSerie") |
|
171 | 171 | .def(py::init<>()) |
|
172 | 172 | .def(py::init<std::size_t>()) |
|
173 | 173 | .def(py::init([](py::array_t<double> t, py::array_t<double> values) { |
|
174 | 174 | assert(t.size() * 3 == values.size()); |
|
175 | 175 | VectorTimeSerie::axis_t _t(t.size()); |
|
176 | 176 | VectorTimeSerie::container_type<VectorTimeSerie::raw_value_type> |
|
177 | 177 | _values(t.size()); |
|
178 | 178 | copy_vector(t, values, _t, _values); |
|
179 | 179 | return VectorTimeSerie(_t, _values); |
|
180 | 180 | })) |
|
181 | 181 | .def( |
|
182 | 182 | "__getitem__", |
|
183 | 183 | [](VectorTimeSerie& ts, std::size_t key) |
|
184 | 184 | -> VectorTimeSerie::raw_value_type& { return ts[key]; }, |
|
185 | 185 | py::return_value_policy::reference) |
|
186 | 186 | .def("__setitem__", [](VectorTimeSerie& ts, std::size_t key, |
|
187 | 187 | VectorTimeSerie::raw_value_type value) { |
|
188 | 188 | *(ts.begin() + key) = value; |
|
189 | 189 | }); |
|
190 | 190 | |
|
191 | 191 | py::class_<MultiComponentTimeSerie::iterator_t>(m, |
|
192 | 192 | "MultiComponentTimeSerieItem") |
|
193 | 193 | .def("__getitem__", [](MultiComponentTimeSerie::iterator_t& self, |
|
194 | 194 | std::size_t key) { return (*self)[key]; }) |
|
195 | 195 | .def("__setitem__", |
|
196 | 196 | [](MultiComponentTimeSerie::iterator_t& self, std::size_t key, |
|
197 | 197 | double value) { (*self)[key] = value; }); |
|
198 | 198 | |
|
199 | 199 | py::class_<MultiComponentTimeSerie, TimeSeries::ITimeSerie, |
|
200 | 200 | std::shared_ptr<MultiComponentTimeSerie>>( |
|
201 | 201 | m, "MultiComponentTimeSerie") |
|
202 | 202 | .def(py::init<>()) |
|
203 | 203 | .def(py::init<const std::vector<std::size_t>>()) |
|
204 | 204 | .def(py::init([](py::array_t<double> t, py::array_t<double> values) { |
|
205 | 205 | assert((t.size() < values.size()) | |
|
206 | 206 | (t.size() == 0)); // TODO check geometry |
|
207 | 207 | MultiComponentTimeSerie::axis_t _t(t.size()); |
|
208 | 208 | MultiComponentTimeSerie::container_type< |
|
209 | 209 | MultiComponentTimeSerie::raw_value_type> |
|
210 | 210 | _values(values.size()); |
|
211 | 211 | copy_multicomp(t, values, _t, _values); |
|
212 | 212 | std::vector<std::size_t> shape; |
|
213 | 213 | shape.push_back(values.shape(0)); |
|
214 | 214 | shape.push_back(values.shape(1)); |
|
215 | 215 | return MultiComponentTimeSerie(_t, _values, shape); |
|
216 | 216 | })) |
|
217 | 217 | .def("__getitem__", |
|
218 | 218 | [](MultiComponentTimeSerie& ts, |
|
219 | 219 | std::size_t key) -> MultiComponentTimeSerie::iterator_t { |
|
220 | 220 | return ts.begin() + key; |
|
221 | 221 | }); |
|
222 | 222 | |
|
223 | 223 | py::class_<SpectrogramTimeSerie::iterator_t>(m, "SpectrogramTimeSerieItem") |
|
224 | 224 | .def("__getitem__", [](SpectrogramTimeSerie::iterator_t& self, |
|
225 | 225 | std::size_t key) { return (*self)[key]; }) |
|
226 | 226 | .def("__setitem__", |
|
227 | 227 | [](SpectrogramTimeSerie::iterator_t& self, std::size_t key, |
|
228 | 228 | double value) { (*self)[key] = value; }); |
|
229 | 229 | |
|
230 | 230 | py::class_<SpectrogramTimeSerie, TimeSeries::ITimeSerie, |
|
231 | 231 | std::shared_ptr<SpectrogramTimeSerie>>(m, "SpectrogramTimeSerie") |
|
232 | 232 | .def(py::init<>()) |
|
233 | 233 | .def(py::init<const std::vector<std::size_t>>()) |
|
234 | 234 | .def(py::init([](py::array_t<double> t, py::array_t<double> y, |
|
235 | 235 | py::array_t<double> values, double min_sampling, |
|
236 | double max_sampling) { | |
|
236 | double max_sampling, bool y_is_log) { | |
|
237 | 237 | if(t.size() >= values.size() and t.size() != 0) |
|
238 | 238 | SCIQLOP_ERROR(decltype(py::self), "Doesn't look like a Spectrogram"); |
|
239 | 239 | if(y.size() != values.shape(1)) |
|
240 | 240 | SCIQLOP_ERROR(decltype(py::self), |
|
241 | 241 | "Y axis size and data shape are incompatible"); |
|
242 | 242 | SpectrogramTimeSerie::axis_t _t(t.size()); |
|
243 | 243 | SpectrogramTimeSerie::axis_t _y(y.size()); |
|
244 | 244 | SpectrogramTimeSerie::container_type< |
|
245 | 245 | SpectrogramTimeSerie::raw_value_type> |
|
246 | 246 | _values(values.size()); |
|
247 | 247 | copy_spectro(t, y, values, _t, _y, _values); |
|
248 | 248 | std::vector<std::size_t> shape; |
|
249 | 249 | shape.push_back(values.shape(0)); |
|
250 | 250 | shape.push_back(values.shape(1)); |
|
251 | 251 | return SpectrogramTimeSerie(std::move(_t), std::move(_y), |
|
252 | 252 | std::move(_values), shape, min_sampling, |
|
253 | max_sampling); | |
|
253 | max_sampling, y_is_log); | |
|
254 | 254 | })) |
|
255 | 255 | .def("__getitem__", |
|
256 | 256 | [](SpectrogramTimeSerie& ts, |
|
257 | 257 | std::size_t key) -> SpectrogramTimeSerie::iterator_t { |
|
258 | 258 | return ts.begin() + key; |
|
259 | 259 | }); |
|
260 | 260 | |
|
261 | 261 | py::class_<Variable2, std::shared_ptr<Variable2>>(m, "Variable2") |
|
262 | 262 | .def(py::init<const QString&>()) |
|
263 | 263 | .def_property("name", &Variable2::name, &Variable2::setName) |
|
264 | 264 | .def_property_readonly("range", &Variable2::range) |
|
265 | 265 | .def_property_readonly("nbPoints", &Variable2::nbPoints) |
|
266 | 266 | .def_property_readonly( |
|
267 | 267 | "data", |
|
268 | 268 | [](Variable2& var) -> std::shared_ptr<TimeSeries::ITimeSerie> { |
|
269 | 269 | return var.data(); |
|
270 | 270 | }) |
|
271 | 271 | .def("set_data", |
|
272 | 272 | [](Variable2& var, std::vector<TimeSeries::ITimeSerie*> ts_list, |
|
273 | 273 | const DateTimeRange& range) { var.setData(ts_list, range); }) |
|
274 | 274 | .def("__len__", &Variable2::nbPoints) |
|
275 | 275 | .def("__repr__", __repr__<Variable2>); |
|
276 | 276 | |
|
277 | 277 | py::class_<DateTimeRange>(m, "SqpRange") |
|
278 | 278 | //.def("fromDateTime", &DateTimeRange::fromDateTime, |
|
279 | 279 | // py::return_value_policy::move) |
|
280 | 280 | .def(py::init([](double start, double stop) { |
|
281 | 281 | return DateTimeRange{start, stop}; |
|
282 | 282 | })) |
|
283 | 283 | .def(py::init( |
|
284 | 284 | [](system_clock::time_point start, system_clock::time_point stop) { |
|
285 | 285 | double start_ = |
|
286 | 286 | 0.001 * |
|
287 | 287 | duration_cast<milliseconds>(start.time_since_epoch()).count(); |
|
288 | 288 | double stop_ = |
|
289 | 289 | 0.001 * |
|
290 | 290 | duration_cast<milliseconds>(stop.time_since_epoch()).count(); |
|
291 | 291 | return DateTimeRange{start_, stop_}; |
|
292 | 292 | })) |
|
293 | 293 | .def_property_readonly("start", |
|
294 | 294 | [](const DateTimeRange& range) { |
|
295 | 295 | return system_clock::from_time_t(range.m_TStart); |
|
296 | 296 | }) |
|
297 | 297 | .def_property_readonly("stop", |
|
298 | 298 | [](const DateTimeRange& range) { |
|
299 | 299 | return system_clock::from_time_t(range.m_TEnd); |
|
300 | 300 | }) |
|
301 | 301 | .def("__repr__", __repr__<DateTimeRange>); |
|
302 | 302 | } |
General Comments 0
You need to be logged in to leave comments.
Login now