Commit fc4b82fe authored by Giovanni La Mura's avatar Giovanni La Mura
Browse files

Merge branch 'c_fortran_binary_io' into 'master' 

Wrap FORTRAN binary I/O in C++

See merge request giacomo.mulas/np_tmcode!4
parents 884f6c11 8e7de8d3

.gitignore

+1 −0
Original line number Diff line number Diff line 
build/include

build/cluster/*

build/sphere/*

build/trapping/*

doc/src/config.dox

+2 −1
Original line number Diff line number Diff line
@@ -342,7 +342,7 @@ OPTIMIZE_OUTPUT_SLICE = NO 
#

# Note see also the list of default file extension mappings.



EXTENSION_MAPPING      =

EXTENSION_MAPPING      = f=FortranFixed f90=FortranFree



# If the MARKDOWN_SUPPORT tag is enabled then doxygen pre-processes all comments

# according to the Markdown format, which allows for more readable
@@ -984,6 +984,7 @@ INPUT_FILE_ENCODING = 
# *.f18, *.f, *.for, *.vhd, *.vhdl, *.ucf, *.qsf and *.ice.



FILE_PATTERNS          = *.cpp \

		         *.f \

		         *.h \

			 *.md

src/sphere/List.h→src/include/List.h

+2 −2
Original line number Diff line number Diff line
@@ -154,8 +154,8 @@ template<class T> class List { 
  T* to_array() {

    T *array = new T[size];

    current = last;

    for (int i = size; i > 0; i--) {

      array[i - 1] = current->value;

    for (int i = size - 1; i > -1; i--) {

      array[i] = current->value;

      current = current->p_prev;

    }

    return array;

src/include/file_io.h

0 → 100644
+43 −0
Original line number Diff line number Diff line 
/*! \file file_io.h

 *

 * C++ wrapper of FORTRAN I/O operations with files.

 */



/*! \brief Open a file for subsequent access.

 *

 * \param uid: `int*` Pointer to the unit ID to be associated to the file.

 * \param name: `char*` C-string for file name (max. length of 63).

 * \param sta: `char*` C-string for file status (max. length of 7).

 * \param mode: `char*` C-string for access mode (max. length of 11).

 */

extern "C" void open_file_(int* uid, const char* name, const char* sta, const char* mode);

/*! \brief Close a previously opened file.

 *

 * \param uid: `int*` Pointer to the unit ID of the file.

 */

extern "C" void close_file_(int* uid);

/*! \brief Read an integer value from a file.

 *

 * \param uid: `int*` Pointer to the unit ID of the file.

 * \param value: `int*` Pointer of the variable to be updated.

 */

extern "C" void read_int_(int* uid, int* value);

/*! \brief Write a complex value to a file.

 *

 * \param uid: `int*` Pointer to the unit ID of the file.

 * \param real: `double*` Pointer to the real part of the value.

 * \param imag: `double*` Pointer to the imaginary part of the value.

 */

extern "C" void write_complex_(int* uid, double* real, double* imag);

/*! \brief Write a double precision float value to a file.

 *

 * \param uid: `int*` Pointer to the unit ID of the file.

 * \param value: `double*` Pointer to the variable to be written.

 */

extern "C" void write_double_(int* uid, double* value);

/*! \brief Write an integer value to a file.

 *

 * \param uid: `int*` Pointer to the unit ID of the file.

 * \param value: `int*` Pointer to the variable to be written.

 */

extern "C" void write_int_(int* uid, int* value);

src/sphere/edfb.cpp

+29 −23
Original line number Diff line number Diff line
@@ -7,7 +7,8 @@ 
#include <cstring>

#include <iostream>

#include <fstream>

#include "List.h"

#include "../include/file_io.h"

#include "../include/List.h"



using namespace std;
@@ -45,7 +46,6 @@ int main(int argc, char **argv) { 
  double *xiv, *wns, *wls, *pus, *evs, *vss;

  string vns[5];



  int max_nsh = 0; // A helper variable to set the size of dc0m

  int ici;



  // Input file reading section
@@ -344,7 +344,7 @@ int main(int argc, char **argv) { 
      }

      vss = xi_vector.to_array();

      xiv = xi_vector.to_array();

      pu = xip + wp;

      pu = xip * wp;

      wn = pu / 3.0e8;

      fprintf(output, "          XIP          WN           WL           PU           EV\n");

      fprintf(output, "     %13.4lE", xip);
@@ -352,7 +352,7 @@ int main(int argc, char **argv) { 
      fprintf(output, "%13.4lE", pigt / wn);

      fprintf(output, "%13.4lE", pu);

      fprintf(output, "%13.4lE\n", pu * evc);

      fprintf(output, "  SCALE FACTORS XI\n", pu * evc);

      fprintf(output, "  SCALE FACTORS XI\n");

      for (int jxi6612 = 1; jxi6612 <= nxi; jxi6612++)

	fprintf(output, "%5d%13.4lE\n", jxi6612, xiv[jxi6612 - 1]);

      //INXI branch ends here.
@@ -361,7 +361,10 @@ int main(int argc, char **argv) { 
  last_read_line++;

  iog = new int[nsph];

  for (int i = 0; i < nsph; i++) {

    sscanf(file_lines[last_read_line].c_str(), " %d", (iog + i));

    string read_format = "";

    for (int j = 0; j < i; j++) read_format += " %*d";

    read_format += " %d";

    sscanf(file_lines[last_read_line].c_str(), read_format.c_str(), (iog + i));

  }

  nshl = new int[nsph];

  ros = new double[nsph];
@@ -376,7 +379,6 @@ int main(int argc, char **argv) { 
    ros[i113 - 1] = ros_val * pow(10.0, ros_val_exp);

    nsh = nshl[i113 -1];

    if (i113 == 1) nsh += ies;

    if ((nsh + 1) / 2 + ies > max_nsh) max_nsh = (nsh + 1) / 2 + ies;

    rcf[i113 - 1] = new double[nsh];

    for (int ns = 0; ns < nsh; ns++) {

      double ns_rcf;
@@ -388,28 +390,33 @@ int main(int argc, char **argv) { 
  // The FORTRAN code writes an auxiliary file in binary format. This should

  // be avoided or possibly replaced with the use of standard file formats for

  // scientific use (e.g. FITS).

  ofstream tedf_file;

  tedf_file.open("c_TEDF", ofstream::binary);

  tedf_file.write(reinterpret_cast<char *>(&nsph), sizeof(nsph));

  int uid = 27;

  string bin_file_name = "c_TEDF";

  string status = "UNKNOWN";

  string mode = "UNFORMATTED";

  open_file_(&uid, bin_file_name.c_str(), status.c_str(), mode.c_str());

  write_int_(&uid, &nsph);

  for (int iogi = 0; iogi < nsph; iogi++)

    tedf_file.write(reinterpret_cast<char *>(iog + iogi), sizeof(iog[iogi]));

  tedf_file.write(reinterpret_cast<char *>(&exdc), sizeof(exdc));

  tedf_file.write(reinterpret_cast<char *>(&wp), sizeof(wp));

  tedf_file.write(reinterpret_cast<char *>(&xip), sizeof(xip));

  tedf_file.write(reinterpret_cast<char *>(&idfc), sizeof(idfc));

  tedf_file.write(reinterpret_cast<char *>(&nxi), sizeof(nxi));

    write_int_(&uid, (iog + iogi));

  write_double_(&uid, &exdc);

  write_double_(&uid, &wp);

  write_double_(&uid, &xip);

  write_int_(&uid, &idfc);

  write_int_(&uid, &nxi);

  for (int xivi = 0; xivi < nxi; xivi++)

    write_double_(&uid, (xiv + xivi));

  for (int i115 = 1; i115 <= nsph; i115++) {

    if (iog[i115 - 1] < i115) continue;

    tedf_file.write(reinterpret_cast<char *>(nshl + i115 - 1), sizeof(nshl[i115 - 1]));

    tedf_file.write(reinterpret_cast<char *>(ros + i115 - 1), sizeof(ros[i115 - 1]));

    write_int_(&uid, (nshl + i115 -1));

    write_double_(&uid, (ros + i115 - 1));

    nsh = nshl[i115 - 1];

    if (i115 == 1) nsh += ies;

    for (int ins = 0; ins < nsh; ins++)

      tedf_file.write(reinterpret_cast<char *>(rcf[i115 - 1] + ins), sizeof(rcf[i115 - 1][ins]));

      write_double_(&uid, (rcf[i115 - 1] + ins));

  }

  // Remake the dc0m matrix.

  dc0m = new complex<double>**[max_nsh];

  for (int dim1 = 0; dim1 < max_nsh; dim1++) {

  dc0m = new complex<double>**[nsph];

  for (int dim1 = 0; dim1 < nsph; dim1++) {

    dc0m[dim1] = new complex<double>*[nsph];

    for (int dim2 = 0; dim2 < nxi; dim2++) {

      dc0m[dim1][dim2] = new complex<double>[nxi];
@@ -432,12 +439,11 @@ int main(int argc, char **argv) { 
	// The FORTRAN code writes the complex numbers as a 16-byte long binary stream.

	// Here we assume that the 16 bytes are equally split in 8 bytes to represent the

	// real part and 8 bytes to represent the imaginary one.

	tedf_file.write(reinterpret_cast<char *>(&dc0_real), sizeof(dc0_real));

	tedf_file.write(reinterpret_cast<char *>(&dc0_img), sizeof(dc0_img));

	write_complex_(&uid, &dc0_real, &dc0_img);

      }

    }

  }

  tedf_file.close();

  close_file_(&uid);

  if (idfc != 0) {

    fprintf(output, "  DIELECTRIC CONSTANTS\n");

    for (int i473 = 1; i473 <= nsph; i473++) {