1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
|
// Copyright 2020 Global Phasing Ltd.
#include "common.h"
#include "array.h"
#include "make_iterator.h"
#include <nanobind/stl/array.h>
#include <nanobind/stl/complex.h>
#include <nanobind/stl/string.h>
#include "gemmi/recgrid.hpp"
#include "gemmi/fourier.hpp" // for get_size_for_hkl, get_f_phi_on_grid
#include "gemmi/sprintf.hpp"
using namespace gemmi;
namespace {
template<typename T, typename Func>
auto make_new_column(const AsuData<T>& asu_data, Func f) {
if (!asu_data.unit_cell().is_crystal())
throw std::runtime_error("AsuData: unknown unit cell parameters");
auto numpy_arr = make_numpy_array<float>({asu_data.size()});
float* arr = numpy_arr.data();
for (size_t i = 0; i < asu_data.size(); ++i)
arr[i] = static_cast<float>(f(asu_data.unit_cell(), asu_data.get_hkl(i)));
return numpy_arr;
}
template<typename T> void add_to_asu_data(T&) {}
template<> void add_to_asu_data(nb::class_<AsuData<std::complex<float>>>& cl) {
using AsuData = AsuData<std::complex<float>>;
cl.def("get_size_for_hkl", &get_size_for_hkl<AsuData>,
nb::arg("min_size")=std::array<int,3>{{0,0,0}}, nb::arg("sample_rate")=0.);
cl.def("data_fits_into", &data_fits_into<AsuData>, nb::arg("size"));
cl.def("get_f_phi_on_grid", get_f_phi_on_grid<float, AsuData>,
nb::arg("size"), nb::arg("half_l")=false, nb::arg("order")=AxisOrder::XYZ);
cl.def("transform_f_phi_to_map", &transform_f_phi_to_map2<float, AsuData>,
nb::arg("min_size")=std::array<int,3>{{0,0,0}},
nb::arg("sample_rate")=0.,
nb::arg("exact_size")=std::array<int,3>{{0,0,0}},
nb::arg("order")=AxisOrder::XYZ);
cl.def("calculate_correlation", [](const AsuData& self, const AsuData& other) {
return calculate_hkl_complex_correlation(self.v, other.v);
});
}
template<> void add_to_asu_data(nb::class_<AsuData<float>>& cl) {
using AsuData = AsuData<float>;
cl.def("calculate_correlation", [](const AsuData& self, const AsuData& other) {
return calculate_hkl_value_correlation(self.v, other.v);
});
}
template<> void add_to_asu_data(nb::class_<AsuData<ValueSigma<float>>>& cl) {
cl.def("discard_by_sigma_ratio", &discard_by_sigma_ratio<float>);
}
template<typename T> struct array_for {
using type = cpu_c_array<T>;
};
template<> struct array_for<ValueSigma<float>> {
using type = nb::ndarray<float, nb::shape<-1, 2>, nb::device::cpu, nb::c_contig>;
};
template<typename T>
void add_asudata(nb::module_& m, const std::string& prefix) {
nb::class_<HklValue<T>>(m, (prefix + "HklValue").c_str())
.def_ro("hkl", &HklValue<T>::hkl)
.def_rw("value", &HklValue<T>::value)
.def("__repr__", [prefix](const HklValue<T>& self) {
return nb::str("<gemmi.{}HklValue ({},{},{}) {}>")
.format(prefix, self.hkl[0], self.hkl[1], self.hkl[2], self.value);
});
using AsuData = AsuData<T>;
nb::class_<AsuData> asu_data(m, (prefix + "AsuData").c_str());
asu_data
.def("__init__", [](AsuData* p, const UnitCell& unit_cell, const SpaceGroup* sg,
const cpu_miller_array& hkl, const typename array_for<T>::type& values) {
new(p) AsuData;
auto h = hkl.view();
auto v = values.view();
if (h.shape(0) != v.shape(0))
throw std::domain_error("error: arrays have different lengths");
p->spacegroup_ = sg;
p->unit_cell_ = unit_cell;
p->unit_cell_.set_cell_images_from_spacegroup(p->spacegroup_);
p->v.reserve(h.shape(0));
for (size_t i = 0; i < h.shape(0); ++i) {
Miller hkl {h(i, 0), h(i, 1), h(i, 2)};
if constexpr (std::is_same<T, ValueSigma<float>>::value)
p->v.push_back({hkl, T{v(i, 0), v(i, 1)}});
else
p->v.push_back({hkl, v(i)});
}
}, nb::arg("cell"), nb::arg("sg")/*.none(false)*/,
nb::arg("miller_array"), nb::arg("value_array"))
.def("__iter__", [](AsuData& self) {
return usual_iterator(self, self.v);
}, nb::keep_alive<0, 1>())
.def("__len__", [](const AsuData& self) { return self.v.size(); })
.def("__getitem__", [](AsuData& self, int index) -> HklValue<T>& {
return self.v.at(normalize_index(index, self.v));
}, nb::arg("index"), nb::rv_policy::reference_internal)
.def_rw("spacegroup", &AsuData::spacegroup_)
.def_rw("unit_cell", &AsuData::unit_cell_)
.def_prop_ro("miller_array", [](AsuData& self) {
// cf. vector_member_array
int64_t stride = int64_t(sizeof(HklValue<T>) / sizeof(int));
return nb::ndarray<nb::numpy, int, nb::shape<-1,3>>(
&self.v[0].hkl[0], {self.v.size(), 3}, nb::handle(), {stride, 1});
}, nb::rv_policy::reference_internal)
.def_prop_ro("value_array", [](AsuData& self) {
if constexpr (std::is_same<T, ValueSigma<float>>::value) {
auto stride = sizeof(ValueSigma<HklValue<float>>) / sizeof(float);
return nb::ndarray<nb::numpy, float, nb::shape<-1, 2>>(
&(self.v.data()->value.value),
{self.v.size(), 2},
nb::handle(),
{(int64_t)stride, 1});
} else {
return vector_member_array(self.v, &HklValue<T>::value);
}
}, nb::rv_policy::reference_internal)
.def("make_1_d2_array", [](const AsuData& asu_data) {
return make_new_column(asu_data, [](const UnitCell& cell, Miller hkl) {
return cell.calculate_1_d2(hkl);
});
})
.def("make_d_array", [](const AsuData& asu_data) {
return make_new_column(asu_data, [](const UnitCell& cell, Miller hkl) {
return cell.calculate_d(hkl);
});
})
.def("count_equal_values", [](const AsuData& self, const AsuData& other) {
return count_equal_values(self.v, other.v);
})
.def("ensure_sorted", &AsuData::ensure_sorted)
.def("ensure_asu", &AsuData::ensure_asu, nb::arg("tnt_asu")=false)
.def("copy", [](const AsuData& self) {
return new AsuData(self);
})
.def("__repr__", [prefix](const AsuData& self) {
return cat("<gemmi.", prefix, "AsuData with ", self.v.size(), " values>");
});
add_to_asu_data(asu_data);
}
template<typename TA, typename TG=TA>
void add_asudata_and_recgrid(nb::module_& m,
const std::string& prefix_asu,
const std::string& rgrid_name) {
using RecGr = ReciprocalGrid<TG>;
// needed before add_asudata for ComplexAsuData.get_f_phi_on_grid
nb::class_<RecGr, GridBase<TG>> recgrid(m, rgrid_name.c_str());
add_asudata<TA>(m, prefix_asu);
recgrid
.def_ro("half_l", &RecGr::half_l)
.def(nb::init<>())
.def("__init__", [](RecGr* grid, int nx, int ny, int nz) {
new(grid) RecGr();
grid->set_size_without_checking(nx, ny, nz);
grid->axis_order = AxisOrder::XYZ;
}, nb::arg("nx"), nb::arg("ny"), nb::arg("nz"))
.def("__init__", [](RecGr* grid,
const nb::ndarray<TG, nb::ndim<3>, nb::device::cpu>& arr,
const UnitCell *cell, const SpaceGroup* sg) {
auto r = arr.view();
new(grid) RecGr();
grid->set_size_without_checking((int)r.shape(0), (int)r.shape(1), (int)r.shape(2));
grid->axis_order = AxisOrder::XYZ;
for (size_t k = 0; k < r.shape(2); ++k)
for (size_t j = 0; j < r.shape(1); ++j)
for (size_t i = 0; i < r.shape(0); ++i)
grid->data[grid->index_q(i, j, k)] = r(i, j, k);
if (cell)
grid->unit_cell = *cell;
if (sg)
grid->spacegroup = sg;
}, nb::arg().noconvert(), nb::arg("cell")=nb::none(), nb::arg("spacegroup")=nb::none())
.def("get_value", &RecGr::get_value)
.def("get_value_or_zero", &RecGr::get_value_or_zero)
.def("set_value", &RecGr::set_value)
.def("to_hkl", &RecGr::to_hkl)
.def("calculate_1_d2", &RecGr::calculate_1_d2)
.def("calculate_d", &RecGr::calculate_d)
.def("get_value_by_hkl", [](RecGr &self, const cpu_miller_array& hkl, double unblur,
bool mott_bethe, TA mott_bethe_000) {
auto h = hkl.view();
auto vals = make_numpy_array<TA>({h.shape(0)});
TA* ptr = vals.data();
for (size_t i = 0; i < h.shape(0); ++i) {
Miller mi{h(i, 0), h(i, 1), h(i, 2)};
if (mott_bethe && mi[0] == 0 && mi[1] == 0 && mi[2] == 0)
ptr[i] = mott_bethe_000;
else
ptr[i] = self.get_value_by_hkl(mi, unblur, mott_bethe);
}
return vals;
}, nb::arg("hkl"), nb::arg("unblur")=0, nb::arg("mott_bethe")=false,
nb::arg("mott_bethe_000")=0)
.def("prepare_asu_data", &RecGr::template prepare_asu_data<TA>,
nb::arg("dmin")=0., nb::arg("unblur")=0.,
nb::arg("with_000")=false, nb::arg("with_sys_abs")=false,
nb::arg("mott_bethe")=false)
.def("__repr__", [=](const RecGr& self) {
return cat("<gemmi.", rgrid_name, '(', self.nu, ", ", self.nv, ", ", self.nw, ")>");
});
}
} // anonymous namespace
void add_recgrid(nb::module_& m) {
using VS = ValueSigma<float>;
nb::class_<VS>(m, "ValueSigma")
.def_rw("value", &VS::value)
.def_rw("sigma", &VS::sigma)
.def("__repr__", [](const VS& self) {
char buf[64];
snprintf_z(buf, 64, "<gemmi.ValueSigma(%g, %g)>", self.value, self.sigma);
return std::string(buf);
});
nb::class_<ComplexCorrelation>(m, "ComplexCorrelation")
.def_ro("n", &ComplexCorrelation::n)
.def("coefficient", &ComplexCorrelation::coefficient)
.def("mean_ratio", &ComplexCorrelation::mean_ratio)
;
add_asudata_and_recgrid<int, int8_t>(m, "Int", "ReciprocalInt8Grid");
add_asudata_and_recgrid<float>(m, "Float", "ReciprocalFloatGrid");
add_asudata_and_recgrid<std::complex<float>>(m, "Complex", "ReciprocalComplexGrid");
add_asudata<VS>(m, "ValueSigma");
}
|
