File: gemmi.cpp

package info (click to toggle)
gemmi 0.6.5%2Bds-3
  • links: PTS, VCS
  • area: main
  • in suites: trixie
  • size: 5,836 kB
  • sloc: cpp: 54,719; python: 4,743; ansic: 3,972; sh: 384; makefile: 73; f90: 42; javascript: 12
file content (186 lines) | stat: -rw-r--r-- 6,894 bytes parent folder | download
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
// Copyright 2017 Global Phasing Ltd.

#include "gemmi/version.hpp"   // for GEMMI_VERSION
#include "gemmi/math.hpp"      // for hc
#include "gemmi/dirwalk.hpp"   // for CifWalk, CoorFileWalk
#include "gemmi/pdb_id.hpp"    // for expand_if_pdb_code
#include "gemmi/bessel.hpp"    // for bessel_i1_over_i0
#include "gemmi/pirfasta.hpp"  // for read_pir_or_fasta
#include "gemmi/seqtools.hpp"  // for calculate_sequence_weight
#include "gemmi/stats.hpp"     // for Correlation
#include "tao/pegtl/parse_error.hpp" // for parse_error

#include "common.h"
#include <pybind11/stl.h>
#include <pybind11/numpy.h>    // for vectorize

namespace py = pybind11;

namespace {
template<typename T> int get_max_bin(const T& bins) {
  int max_bin = 0;
  for (int i = 0; i < bins.shape(0); ++i) {
    if (bins(i) < 0)
      throw py::value_error("bins argument must have no negative elements");
    max_bin = std::max(max_bin, bins(i));
  }
  if (max_bin > 1000000)
    throw py::value_error("bin numbers must be smaller than million");
  return max_bin;
}
} // anonymous namespace

void add_misc(py::module& m) {
  py::class_<gemmi::CifWalk>(m, "CifWalk")
    .def(py::init<const char*, char>(), py::arg("path"), py::arg("try_pdbid")='\0')
    .def("__iter__", [](gemmi::CifWalk& self) {
        return py::make_iterator(self);
    }, py::keep_alive<0, 1>());
  py::class_<gemmi::CoorFileWalk>(m, "CoorFileWalk")
    .def(py::init<const char*, char>(), py::arg("path"), py::arg("try_pdbid")='\0')
    .def("__iter__", [](gemmi::CoorFileWalk& self) {
        return py::make_iterator(self);
    }, py::keep_alive<0, 1>());
  m.def("is_pdb_code", &gemmi::is_pdb_code);
  m.def("expand_pdb_code_to_path", &gemmi::expand_pdb_code_to_path,
        py::arg("code"), py::arg("filetype"), py::arg("throw_if_unset")=false);
  m.def("expand_if_pdb_code", &gemmi::expand_if_pdb_code,
        py::arg("code"), py::arg("filetype")='M');
  m.attr("hc") = py::float_(gemmi::hc());
  m.def("bessel_i1_over_i0", py::vectorize(gemmi::bessel_i1_over_i0));
  m.def("log_bessel_i0", py::vectorize(gemmi::log_bessel_i0));
  m.def("log_cosh", py::vectorize([](double x) {
        // ln(cosh(x)) = ln(e^x + e^-x) - ln(2) = ln(e^x * (1 + e^-2x)) - ln(2)
        x = std::abs(x);
        return x + std::log1p(std::exp(-2 * x)) - std::log(2);
  }));

  // pirfasta.hpp
  py::class_<gemmi::FastaSeq>(m, "FastaSeq")
    .def_readonly("header", &gemmi::FastaSeq::header)
    .def_readonly("seq", &gemmi::FastaSeq::seq)
    ;
  m.def("read_pir_or_fasta", &gemmi::read_pir_or_fasta);

  // seqtools.hpp
  m.def("calculate_sequence_weight", &gemmi::calculate_sequence_weight,
        py::arg("sequence"), py::arg("unknown")=0.);
  m.def("one_letter_code", &gemmi::one_letter_code);
  m.def("pdbx_one_letter_code", &gemmi::pdbx_one_letter_code);
  m.def("sequence_kind", &gemmi::sequence_kind);

  // stats.hpp
  py::class_<gemmi::Correlation>(m, "Correlation")
    .def_readonly("n", &gemmi::Correlation::n)
    .def("coefficient", &gemmi::Correlation::coefficient)
    .def("mean_ratio", &gemmi::Correlation::mean_ratio)
    ;

  // utilities inspired by numpy.bincount()
  m.def("binmean", [](py::array_t<int> bins, py::array_t<double> values) {
      auto bins_ = bins.unchecked<1>();
      auto values_ = values.unchecked<1>();
      auto len = bins_.shape(0);
      if (len != values_.shape(0))
        throw std::domain_error("arrays have different lengths");
      int ret_size = get_max_bin(bins_) + 1;
      py::array_t<double> ret(ret_size);
      double* retp = (double*) ret.request().ptr;
      for (int i = 0; i != ret_size; ++i)
        retp[i] = 0.;
      std::vector<int> counts(ret_size);
      for (int i = 0; i != len; ++i)
        if (!std::isnan(values_(i))) {
          int n = bins_(i);
          counts[n]++;
          retp[n] += values_(i);
        }
      for (int i = 0; i != ret_size; ++i)
        retp[i] /= counts[i];
      return ret;
  }, py::arg("nbins"), py::arg("values"));

  m.def("binrfactor", [](py::array_t<int> bins, py::array_t<double> obs,
                         py::array_t<double> calc, bool riso) {
      auto bins_ = bins.unchecked<1>();
      auto obs_ = obs.unchecked<1>();
      auto calc_ = calc.unchecked<1>();
      auto len = bins_.shape(0);
      if (len != obs_.shape(0) || len != calc_.shape(0))
        throw std::domain_error("arrays have different lengths");
      int ret_size = get_max_bin(bins_) + 1;
      py::array_t<double> ret(ret_size);
      double* retp = (double*) ret.request().ptr;
      for (int i = 0; i != ret_size; ++i)
        retp[i] = 0.;
      std::vector<double> sum_fobs(ret_size);
      for (int i = 0; i != len; ++i)
        if (!std::isnan(obs_(i)) && !std::isnan(calc_(i))) {
          int n = bins_(i);
          retp[n] += std::fabs(obs_(i) - calc_(i));
          sum_fobs[n] += riso ? (obs_(i) + calc_(i)) : obs_(i);
        }
      for (int i = 0; i != ret_size; ++i)
        retp[i] /= (riso ? 0.5 * sum_fobs[i] : sum_fobs[i]);
      return ret;
  }, py::arg("nbins"), py::arg("obs"), py::arg("calc"), py::arg("riso")=false);

  m.def("bincorr", [](py::array_t<int> bins, py::array_t<double> obs,
                                             py::array_t<double> calc) {
      auto bins_ = bins.unchecked<1>();
      auto obs_ = obs.unchecked<1>();
      auto calc_ = calc.unchecked<1>();
      auto len = bins_.shape(0);
      if (len != obs_.shape(0) || len != calc_.shape(0))
        throw std::domain_error("arrays have different lengths");
      int ret_size = get_max_bin(bins_) + 1;
      std::vector<gemmi::Correlation> cor(ret_size);
      for (int i = 0; i != len; ++i)
        if (!std::isnan(obs_(i)) && !std::isnan(calc_(i)))
          cor[bins_(i)].add_point(obs_(i), calc_(i));
      return cor;
  }, py::arg("nbins"), py::arg("obs"), py::arg("calc"));
}

PYBIND11_MODULE(gemmi, mg) {
  mg.doc() = "Python bindings to GEMMI - a library used in macromolecular\n"
             "crystallography and related fields";
  mg.attr("__version__") = GEMMI_VERSION;

  py::register_exception_translator([](std::exception_ptr p) {
    try {
      if (p)
        std::rethrow_exception(p);
    } catch (const std::system_error &e) {
      const int errornum = e.code().value();
      PyErr_SetObject(PyExc_IOError, py::make_tuple(errornum, e.what()).ptr());
    } catch (const tao::pegtl::parse_error &e) {
      PyErr_SetString(PyExc_ValueError, e.what());
    }
  });

  py::module cif = mg.def_submodule("cif", "CIF file format");
  add_cif(cif);
  add_symmetry(mg);
  add_unitcell(mg);
  add_elem(mg);
  add_meta(mg);
  add_mol(mg);
  add_small(mg);
  add_misc(mg);
  add_grid(mg);
  add_recgrid(mg);
  add_ccp4(mg);
  add_sf(mg);
  add_cif_read(cif);
  add_mtz(mg);
  add_hkl(mg);
  add_chemcomp(mg);
  add_monlib(mg);
  add_topo(mg);
  add_alignment(mg);
  add_search(mg);
  add_read_structure(mg);
  add_scaling(mg);
  add_custom(mg);
}