Add dump cfg/zstd

/* ----------------------------------------------------------------------

   LAMMPS - Large-scale Atomic/Molecular Massively Parallel Simulator

   http://lammps.sandia.gov, Sandia National Laboratories

   Steve Plimpton, sjplimp@sandia.gov

   Copyright (2003) Sandia Corporation.  Under the terms of Contract

   DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government retains

   certain rights in this software.  This software is distributed under

   the GNU General Public License.

   See the README file in the top-level LAMMPS directory.

------------------------------------------------------------------------- */

#include "dump_cfg_zstd.h"

#include "atom.h"

#include "domain.h"

#include "error.h"

#include "update.h"

#include "force.h"

#include <cstring>

#include <fmt/format.h>

using namespace LAMMPS_NS;

#define UNWRAPEXPAND 10.0

DumpCFGZstd::DumpCFGZstd(LAMMPS *lmp, int narg, char **arg) :

  DumpCFG(lmp, narg, arg)

{

  if (!compressed)

    error->all(FLERR,"Dump cfg/zstd only writes compressed files");

}

/* ---------------------------------------------------------------------- */

DumpCFGZstd::~DumpCFGZstd()

{

}

/* ----------------------------------------------------------------------

   generic opening of a dump file

   ASCII or binary or zstdipped

   some derived classes override this function

------------------------------------------------------------------------- */

void DumpCFGZstd::openfile()

{

  // single file, already opened, so just return

  if (singlefile_opened) return;

  if (multifile == 0) singlefile_opened = 1;

  // if one file per timestep, replace '*' with current timestep

  char *filecurrent = filename;

  if (multiproc) filecurrent = multiname;

  if (multifile) {

    char *filestar = filecurrent;

    filecurrent = new char[strlen(filestar) + 16];

    char *ptr = strchr(filestar,'*');

    *ptr = '\0';

    if (padflag == 0)

      sprintf(filecurrent,"%s" BIGINT_FORMAT "%s",

              filestar,update->ntimestep,ptr+1);

    else {

      char bif[8],pad[16];

      strcpy(bif,BIGINT_FORMAT);

      sprintf(pad,"%%s%%0%d%s%%s",padflag,&bif[1]);

      sprintf(filecurrent,pad,filestar,update->ntimestep,ptr+1);

    }

    *ptr = '*';

    if (maxfiles > 0) {

      if (numfiles < maxfiles) {

        nameslist[numfiles] = new char[strlen(filecurrent)+1];

        strcpy(nameslist[numfiles],filecurrent);

        ++numfiles;

      } else {

        remove(nameslist[fileidx]);

        delete[] nameslist[fileidx];

        nameslist[fileidx] = new char[strlen(filecurrent)+1];

        strcpy(nameslist[fileidx],filecurrent);

        fileidx = (fileidx + 1) % maxfiles;

      }

    }

  }

  // each proc with filewriter = 1 opens a file

  if (filewriter) {

    if (append_flag) {

      error->one(FLERR, "dump cfg/zstd currently doesn't support append");

    }

    try {

      writer.open(filecurrent);

    } catch (FileWriterException & e) {

      error->one(FLERR, e.what());

    }

  }

  // delete string with timestep replaced

  if (multifile) delete [] filecurrent;

}

/* ---------------------------------------------------------------------- */

void DumpCFGZstd::write_header(bigint n)

{

  // set scale factor used by AtomEye for CFG viz

  // default = 1.0

  // for peridynamics, set to pre-computed PD scale factor

  //   so PD particles mimic C atoms

  // for unwrapped coords, set to UNWRAPEXPAND (10.0)

  //   so molecules are not split across periodic box boundaries

  double scale = 1.0;

  if (atom->peri_flag) scale = atom->pdscale;

  else if (unwrapflag == 1) scale = UNWRAPEXPAND;

  std::string header = fmt::format("Number of particles = {}\n", n);

  header += fmt::format("A = {0:g} Angstrom (basic length-scale)\n", scale);

  header += fmt::format("H0(1,1) = {0:g} A\n",domain->xprd);

  header += fmt::format("H0(1,2) = 0 A \n");

  header += fmt::format("H0(1,3) = 0 A \n");

  header += fmt::format("H0(2,1) = {0:g} A \n",domain->xy);

  header += fmt::format("H0(2,2) = {0:g} A\n",domain->yprd);

  header += fmt::format("H0(2,3) = 0 A \n");

  header += fmt::format("H0(3,1) = {0:g} A \n",domain->xz);

  header += fmt::format("H0(3,2) = {0:g} A \n",domain->yz);

  header += fmt::format("H0(3,3) = {0:g} A\n",domain->zprd);

  header += fmt::format(".NO_VELOCITY.\n");

  header += fmt::format("entry_count = {}\n",nfield-2);

  for (int i = 0; i < nfield-5; i++)

    header += fmt::format("auxiliary[{}] = {}\n",i,auxname[i]);

  writer.write(header.c_str(), header.length());

}

/* ---------------------------------------------------------------------- */

void DumpCFGZstd::write_data(int n, double *mybuf)

{

  writer.write(mybuf, n);

}

/* ---------------------------------------------------------------------- */

void DumpCFGZstd::write()

{

  DumpCFG::write();

  if (filewriter) {

    if (multifile) {

      writer.close();

    } else {

      if (flush_flag && writer.isopen()) {

        writer.flush();

      }

    }

  }

}

/* ---------------------------------------------------------------------- */

int DumpCFGZstd::modify_param(int narg, char **arg)

{

  int consumed = DumpCFG::modify_param(narg, arg);

  if(consumed == 0) {

    try {

      if (strcmp(arg[0],"checksum") == 0) {

        if (narg < 2) error->all(FLERR,"Illegal dump_modify command");

        if (strcmp(arg[1],"yes") == 0) writer.setChecksum(true);

        else if (strcmp(arg[1],"no") == 0) writer.setChecksum(false);

        else error->all(FLERR,"Illegal dump_modify command");

        return 2;

      } else if (strcmp(arg[0],"compression_level") == 0) {

        if (narg < 2) error->all(FLERR,"Illegal dump_modify command");

        int compression_level = force->inumeric(FLERR,arg[1]);

        writer.setCompressionLevel(compression_level);

        return 2;

      }

    } catch (FileWriterException & e) {

      error->one(FLERR, e.what());

    }

  }

  return consumed;

}

/* -*- c++ -*- ----------------------------------------------------------

   LAMMPS - Large-scale Atomic/Molecular Massively Parallel Simulator

   http://lammps.sandia.gov, Sandia National Laboratories

   Steve Plimpton, sjplimp@sandia.gov

   Copyright (2003) Sandia Corporation.  Under the terms of Contract

   DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government retains

   certain rights in this software.  This software is distributed under

   the GNU General Public License.

   See the README file in the top-level LAMMPS directory.

------------------------------------------------------------------------- */

#ifdef DUMP_CLASS

DumpStyle(cfg/zstd,DumpCFGZstd)

#else

#ifndef LMP_DUMP_CFG_ZSTD_H

#define LMP_DUMP_CFG_ZSTD_H

#include "dump_cfg.h"

#include "zstd_file_writer.h"

namespace LAMMPS_NS {

class DumpCFGZstd : public DumpCFG {

 public:

  DumpCFGZstd(class LAMMPS *, int, char **);

  virtual ~DumpCFGZstd();

 protected:

  ZstdFileWriter writer;

  virtual void openfile();

  virtual void write_header(bigint);

  virtual void write_data(int, double *);

  virtual void write();

  virtual int modify_param(int, char **);

};

}

#endif

#endif

/* ERROR/WARNING messages:

E: Dump cfg/zstd only writes compressed files

The dump cfg/zstd output file name must have a .zstd suffix.

E: Cannot open dump file

Self-explanatory.

*/

        add_test(NAME DumpCustomZstd COMMAND test_dump_custom_zstd WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})

        set_tests_properties(DumpCustomZstd PROPERTIES ENVIRONMENT "LAMMPS_POTENTIALS=${LAMMPS_POTENTIALS_DIR}")

        set_tests_properties(DumpCustomZstd PROPERTIES ENVIRONMENT "ZSTD_BINARY=${ZSTD_BINARY}")

        add_executable(test_dump_cfg_zstd test_dump_cfg_zstd.cpp)

        target_link_libraries(test_dump_cfg_zstd PRIVATE lammps GTest::GMock GTest::GTest)

        add_test(NAME DumpCfgZstd COMMAND test_dump_cfg_zstd WORKING_DIRECTORY ${CMAKE_CURRENT_BINARY_DIR})

        set_tests_properties(DumpCfgZstd PROPERTIES ENVIRONMENT "LAMMPS_POTENTIALS=${LAMMPS_POTENTIALS_DIR}")

        set_tests_properties(DumpCfgZstd PROPERTIES ENVIRONMENT "ZSTD_BINARY=${ZSTD_BINARY}")

    endif()

endif()

/* ----------------------------------------------------------------------

   LAMMPS - Large-scale Atomic/Molecular Massively Parallel Simulator

   http://lammps.sandia.gov, Sandia National Laboratories

   Steve Plimpton, sjplimp@sandia.gov

   Copyright (2003) Sandia Corporation.  Under the terms of Contract

   DE-AC04-94AL85000 with Sandia Corporation, the U.S. Government retains

   certain rights in this software.  This software is distributed under

   the GNU General Public License.

   See the README file in the top-level LAMMPS directory.

------------------------------------------------------------------------- */

#include "gtest/gtest.h"

#include "gmock/gmock.h"

#include "fmt/format.h"

#include "utils.h"

#include "../testing/core.h"

#include "../testing/systems/melt.h"

#include "../testing/utils.h"

char * ZSTD_BINARY = nullptr;

using ::testing::Eq;

class DumpCfgZstdTest : public MeltTest {

    std::string dump_style = "cfg";

public:

    void generate_text_and_compressed_dump(std::string text_file, std::string compressed_file, std::string compression_style,

                                           std::string fields, std::string dump_modify_options, int ntimesteps) {

        if (!verbose) ::testing::internal::CaptureStdout();

        command(fmt::format("dump id0 all {} 1 {} {}", dump_style, text_file, fields));

        command(fmt::format("dump id1 all {} 1 {} {}", compression_style, compressed_file, fields));

        if (!dump_modify_options.empty()) {

            command(fmt::format("dump_modify id0 {}", dump_modify_options));

            command(fmt::format("dump_modify id1 {}", dump_modify_options));

        }

        command(fmt::format("run {}", ntimesteps));

        if (!verbose) ::testing::internal::GetCapturedStdout();

    }

    std::string convert_compressed_to_text(std::string compressed_file) {

        if (!verbose) ::testing::internal::CaptureStdout();

        std::string converted_file = compressed_file.substr(0, compressed_file.find_last_of('.'));

        std::string cmdline = fmt::format("{} -d -c {} > {}", ZSTD_BINARY, compressed_file, converted_file);

        system(cmdline.c_str());

        if (!verbose) ::testing::internal::GetCapturedStdout();

        return converted_file;

    }

};

TEST_F(DumpCfgZstdTest, compressed_run0)

{

    if(!ZSTD_BINARY) GTEST_SKIP();

    auto text_files = "dump_cfg_zstd_text_run*.melt.cfg";

    auto compressed_files = "dump_cfg_zstd_compressed_run*.melt.cfg.zst";

    auto text_file = "dump_cfg_zstd_text_run0.melt.cfg";

    auto compressed_file = "dump_cfg_zstd_compressed_run0.melt.cfg.zst";

    auto fields = "mass type xs ys zs id proc procp1 x y z ix iy iz vx vy vz fx fy fz";

    generate_text_and_compressed_dump(text_files, compressed_files, "cfg/gz", fields, "", 0);

    TearDown();

    ASSERT_FILE_EXISTS(text_file);

    ASSERT_FILE_EXISTS(compressed_file);

    auto converted_file = convert_compressed_to_text(compressed_file);

    ASSERT_FILE_EXISTS(converted_file);

    ASSERT_FILE_EQUAL(text_file, converted_file);

    delete_file(text_file);

    delete_file(compressed_file);

    delete_file(converted_file);

}

TEST_F(DumpCfgZstdTest, compressed_unwrap_run0)

{

    if(!ZSTD_BINARY) GTEST_SKIP();

    auto text_files = "dump_cfg_unwrap_zstd_text_run*.melt.cfg";

    auto compressed_files = "dump_cfg_unwrap_zstd_compressed_run*.melt.cfg.zst";

    auto text_file = "dump_cfg_unwrap_zstd_text_run0.melt.cfg";

    auto compressed_file = "dump_cfg_unwrap_zstd_compressed_run0.melt.cfg.zst";

    auto fields = "mass type xsu ysu zsu id proc procp1 x y z ix iy iz vx vy vz fx fy fz";

    generate_text_and_compressed_dump(text_files, compressed_files, "cfg/zstd", fields, "", 0);

    TearDown();

    ASSERT_FILE_EXISTS(text_file);

    ASSERT_FILE_EXISTS(compressed_file);

    auto converted_file = convert_compressed_to_text(compressed_file);

    ASSERT_FILE_EXISTS(converted_file);

    ASSERT_FILE_EQUAL(text_file, converted_file);

    delete_file(text_file);

    delete_file(compressed_file);

    delete_file(converted_file);

}

int main(int argc, char **argv)

{

    MPI_Init(&argc, &argv);

    ::testing::InitGoogleMock(&argc, argv);

    // handle arguments passed via environment variable

    if (const char *var = getenv("TEST_ARGS")) {

        std::vector<std::string> env = utils::split_words(var);

        for (auto arg : env) {

            if (arg == "-v") {

                verbose = true;

            }

        }

    }

    ZSTD_BINARY = getenv("ZSTD_BINARY");

    if ((argc > 1) && (strcmp(argv[1], "-v") == 0)) verbose = true;

    int rv = RUN_ALL_TESTS();

    MPI_Finalize();

    return rv;

}