Merge pull request #3205 from akohlmey/dump-style-yaml

Add dump style yaml

doc/src/dump.rst

@@ -36,7 +36,7 @@ Syntax
 
 * ID = user-assigned name for the dump
 * group-ID = ID of the group of atoms to be dumped
-* style = *atom* or *atom/gz* or *atom/zstd or *atom/mpiio* or *cfg* or *cfg/gz* or *cfg/zstd* or *cfg/mpiio* or *custom* or *custom/gz* or *custom/zstd* or *custom/mpiio* or *dcd* or *h5md* or *image* or *local* or *local/gz* or *local/zstd* or *molfile* or *movie* or *netcdf* or *netcdf/mpiio* or *vtk* or *xtc* or *xyz* or *xyz/gz* or *xyz/zstd* or *xyz/mpiio*
+* style = *atom* or *atom/gz* or *atom/zstd or *atom/mpiio* or *cfg* or *cfg/gz* or *cfg/zstd* or *cfg/mpiio* or *custom* or *custom/gz* or *custom/zstd* or *custom/mpiio* or *dcd* or *h5md* or *image* or *local* or *local/gz* or *local/zstd* or *molfile* or *movie* or *netcdf* or *netcdf/mpiio* or *vtk* or *xtc* or *xyz* or *xyz/gz* or *xyz/zstd* or *xyz/mpiio* or *yaml*
 * N = dump every this many timesteps
 * file = name of file to write dump info to
 * args = list of arguments for a particular style
@@ -68,8 +68,9 @@ Syntax
        *xyz/gz* args = none
        *xyz/zstd* args = none
        *xyz/mpiio* args = none
+       *yaml* args = same as *custom* args, see below
 
-* *custom* or *custom/gz* or *custom/zstd* or *custom/mpiio* or *netcdf* or *netcdf/mpiio* args = list of atom attributes
+* *custom* or *custom/gz* or *custom/zstd* or *custom/mpiio* or *netcdf* or *netcdf/mpiio* or *yaml* args = list of atom attributes
 
   .. parsed-literal::
 
@@ -386,6 +387,70 @@ from using the (numerical) atom type to an element name (or some
 other label). This will help many visualization programs to guess
 bonds and colors.
 
+Dump style *yaml* has the same command syntax as style *custom* and
+writes YAML format files that can be easily parsed by a variety of data
+processing tools and programming languages.  Each timestep will be
+written as a YAML "document" (i.e. starts with "---" and ends with
+"...").  The style supports writing one file per timestep through the
+"\*" wildcard but not multi-processor outputs with the "%" token in the
+filename.  In addition to per-atom data, :doc:`thermo <thermo>` data can
+be included in the *yaml* style dump file using the :doc:`dump_modify
+thermo yes <dump_modify>`. The data included in the dump file uses the
+"thermo" tag and is otherwise identical to data specified by the
+:doc:`thermo_style <thermo_style>` command.
+
+Below is an example for a YAML format dump created by the following commands.
+
+.. code-block:: LAMMPS
+
+   dump out all yaml 100 dump.yaml id type x y z vx vy vz ix iy iz
+   dump_modify out time yes units yes thermo yes format 1 %5d format "% 10.6e"
+
+The tags "time", "units", and "thermo" are optional and enabled by the
+dump_modify command. The list under the "box" tag has 3 lines for
+orthogonal boxes and 4 lines with triclinic boxes, where the first 3 are
+the box boundaries and the 4th the three tilt factors (xy, xz, yz).  The
+"thermo" data follows the format of the *yaml* thermo style.  The
+"keywords" tag lists the per-atom properties contained in the "data"
+columns, which contain a list with one line per atom.  The keywords may
+be renamed using the dump_modify command same as for the *custom* dump
+style.
+
+.. code-block:: yaml
+
+   ---
+   timestep: 0
+   units: lj
+   time: 0
+   natoms: 4000
+   boundary: [ p, p, p, p, p, p, ]
+   thermo:
+     - keywords: [ Step, Temp, E_pair, E_mol, TotEng, Press, ]
+     - data: [ 0, 0, -27093.472213010766, 0, 0, 0, ]
+   box:
+     - [ 0, 16.795961913825074 ]
+     - [ 0, 16.795961913825074 ]
+     - [ 0, 16.795961913825074 ]
+     - [ 0, 0, 0 ]
+   keywords: [ id, type, x, y, z, vx, vy, vz, ix, iy, iz,  ]
+   data:
+     - [     1 , 1 ,  0.000000e+00 ,  0.000000e+00 ,  0.000000e+00 ,  -1.841579e-01 , -9.710036e-01 , -2.934617e+00 , 0 , 0 , 0, ]
+     - [     2 , 1 ,  8.397981e-01 ,  8.397981e-01 ,  0.000000e+00 ,  -1.799591e+00 ,  2.127197e+00 ,  2.298572e+00 , 0 , 0 , 0, ]
+     - [     3 , 1 ,  8.397981e-01 ,  0.000000e+00 ,  8.397981e-01 ,  -1.807682e+00 , -9.585130e-01 ,  1.605884e+00 , 0 , 0 , 0, ]
+
+     [...]
+   ...
+   ---
+   timestep: 100
+   units: lj
+   time: 0.5
+
+     [...]
+
+   ...
+
+----------
+
 Note that *atom*, *custom*, *dcd*, *xtc*, and *xyz* style dump files
 can be read directly by `VMD <http://www.ks.uiuc.edu/Research/vmd>`_, a
 popular molecular viewing program.
@@ -427,9 +492,9 @@ If a "%" character appears in the filename, then each of P processors
 writes a portion of the dump file, and the "%" character is replaced
 with the processor ID from 0 to P-1.  For example, tmp.dump.% becomes
 tmp.dump.0, tmp.dump.1, ... tmp.dump.P-1, etc.  This creates smaller
-files and can be a fast mode of output on parallel machines that
+files and can be a fast mode of output on parallel machines that support
-support parallel I/O for output. This option is not available for the
+parallel I/O for output. This option is **not** available for the *dcd*,
-*dcd*, *xtc*, and *xyz* styles.
+*xtc*, *xyz*, and *yaml* styles.
 
 By default, P = the number of processors meaning one file per
 processor, but P can be set to a smaller value via the *nfile* or
@@ -722,8 +787,8 @@ are part of the MPIIO package.  They are only enabled if LAMMPS was
 built with that package.  See the :doc:`Build package <Build_package>`
 doc page for more info.
 
-The *xtc* and *dcd* styles are part of the EXTRA-DUMP package.  They
+The *xtc*, *dcd* and *yaml* styles are part of the EXTRA-DUMP package.
-are only enabled if LAMMPS was built with that package.  See the
+They are only enabled if LAMMPS was built with that package.  See the
 :doc:`Build package <Build_package>` page for more info.
 
 Related commands

doc/src/dump_modify.rst

@@ -712,8 +712,8 @@ run, this option is ignored since the output is already balanced.
 
 ----------
 
-The *thermo* keyword only applies the dump *netcdf* style.  It
+The *thermo* keyword only applies the dump styles *netcdf* and *yaml*.
-triggers writing of :doc:`thermo <thermo>` information to the dump file
+It triggers writing of :doc:`thermo <thermo>` information to the dump file
 alongside per-atom data.  The values included in the dump file are
 identical to the values specified by :doc:`thermo_style <thermo_style>`.
 

examples/yaml/in.yaml

@@ -29,9 +29,9 @@ fix             1 all nve
 thermo_style    yaml
 thermo 10
 
-dump            1 all yaml 25 dump.yaml id type x y z ix iy iz vx vy vz
+#dump            1 all yaml 25 dump.yaml id type x y z ix iy iz vx vy vz
-dump_modify     1 sort id thermo yes units yes time yes format 1 %5d format float "% 12.8e" format int %2d
+#dump_modify     1 sort id thermo yes units yes time yes format 1 %5d format float "% 12.8e" format int %2d
 
-run             100
+run             100 post no
 
-run             100
+run             100 post no

python/lammps/formats.py

@@ -21,7 +21,7 @@ import re, yaml
 try:
   from yaml import CSafeLoader as Loader, CSafeDumper as Dumper
 except ImportError:
-    from yaml import SafeLoader as Loader, SafeDumper as Dumper
+  from yaml import SafeLoader as Loader, SafeDumper as Dumper
 
 class LogFile:
   """Reads LAMMPS log files and extracts the thermo information

src/.gitignore

@@ -602,6 +602,8 @@
 /dump_xyz_mpiio.h
 /dump_xyz_zstd.cpp
 /dump_xyz_zstd.h
+/dump_yaml.cpp
+/dump_yaml.h
 /dynamical_matrix.cpp
 /dynamical_matrix.h
 /ewald.cpp

src/EXTRA-DUMP/dump_yaml.cpp

@@ -0,0 +1,143 @@
+/* ----------------------------------------------------------------------
+   LAMMPS - Large-scale Atomic/Molecular Massively Parallel Simulator
+   https://www.lammps.org/, 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_yaml.h"
+
+#include "comm.h"
+#include "domain.h"
+#include "error.h"
+#include "output.h"
+#include "thermo.h"
+#include "update.h"
+
+using namespace LAMMPS_NS;
+
+/* ---------------------------------------------------------------------- */
+DumpYAML::DumpYAML(class LAMMPS *_lmp, int narg, char **args) :
+    DumpCustom(_lmp, narg, args), thermo(false)
+{
+  buffer_allow = 0;
+  buffer_flag = 0;
+}
+
+/* ---------------------------------------------------------------------- */
+
+void DumpYAML::init_style()
+{
+  if (binary) error->all(FLERR, "Dump style yaml does not support binary output");
+  if (multiproc) error->all(FLERR, "Dump style yaml does not support multi-processor output");
+
+  DumpCustom::init_style();
+}
+
+/* ---------------------------------------------------------------------- */
+
+void DumpYAML::write()
+{
+  // temporarily enable so write_header() is called
+  // by all MPI ranks to compute thermo data
+  if (thermo) filewriter = 1;
+
+  Dump::write();
+}
+
+/* ---------------------------------------------------------------------- */
+
+void DumpYAML::write_header(bigint ndump)
+{
+  std::string thermo_data;
+  if (thermo) {
+    Thermo *th = output->thermo;
+    thermo_data += "thermo:\n  - keywords: [ ";
+    for (int i = 0; i < th->nfield; ++i) thermo_data += fmt::format("{}, ", th->keyword[i]);
+    thermo_data += "]\n  - data: [ ";
+
+    for (int i = 0; i < th->nfield; ++i) {
+      th->call_vfunc(i);
+      if (th->vtype[i] == Thermo::FLOAT)
+        thermo_data += fmt::format("{}, ", th->dvalue);
+      else if (th->vtype[i] == Thermo::INT)
+        thermo_data += fmt::format("{}, ", th->ivalue);
+      else if (th->vtype[i] == Thermo::BIGINT)
+        thermo_data += fmt::format("{}, ", th->bivalue);
+    }
+    thermo_data += "]\n";
+    MPI_Barrier(world);
+  }
+
+  if (comm->me == 0) {
+    const std::string boundary(boundstr);
+    fmt::print(fp, "---\ntimestep: {}\n", update->ntimestep);
+    if (unit_flag) fmt::print(fp, "units: {}\n", update->unit_style);
+    if (time_flag) fmt::print(fp, "time: {:.16g}\n", compute_time());
+
+    fmt::print(fp, "natoms: {}\n", ndump);
+    fputs("boundary: [ ", fp);
+    for (const auto bflag : boundary) {
+      if (bflag == ' ') continue;
+      fmt::print(fp, "{}, ", bflag);
+    }
+    fputs("]\n", fp);
+
+    if (thermo) fmt::print(fp, thermo_data);
+
+    fmt::print(fp, "box:\n  - [ {}, {} ]\n", boxxlo, boxxhi);
+    fmt::print(fp, "  - [ {}, {} ]\n", boxylo, boxyhi);
+    fmt::print(fp, "  - [ {}, {} ]\n", boxzlo, boxzhi);
+    if (domain->triclinic) fmt::print(fp, "  - [ {}, {}, {} ]\n", boxxy, boxxz, boxyz);
+
+    fmt::print(fp, "keywords: [ ");
+    for (const auto &item : utils::split_words(columns)) fmt::print(fp, "{}, ", item);
+    fputs(" ]\ndata:\n", fp);
+  } else    // reset so that the remainder of the output is not multi-proc
+    filewriter = 0;
+}
+
+/* ---------------------------------------------------------------------- */
+
+void DumpYAML::write_data(int n, double *mybuf)
+{
+  int m = 0;
+  for (int i = 0; i < n; i++) {
+    fputs("  - [ ", fp);
+    for (int j = 0; j < nfield; j++) {
+      if (vtype[j] == Dump::INT)
+        fprintf(fp, vformat[j], static_cast<int>(mybuf[m]));
+      else if (vtype[j] == Dump::DOUBLE)
+        fprintf(fp, vformat[j], mybuf[m]);
+      else if (vtype[j] == Dump::STRING)
+        fprintf(fp, vformat[j], typenames[(int) mybuf[m]]);
+      else if (vtype[j] == Dump::BIGINT)
+        fprintf(fp, vformat[j], static_cast<bigint>(mybuf[m]));
+      m++;
+      fputs(", ", fp);
+    }
+    fputs("]\n", fp);
+  }
+  fputs("...\n", fp);
+}
+
+/* ---------------------------------------------------------------------- */
+
+int DumpYAML::modify_param(int narg, char **arg)
+{
+  int n = DumpCustom::modify_param(narg, arg);
+  if (n > 0) return n;
+
+  if (strcmp(arg[0], "thermo") == 0) {
+    if (narg < 2) error->all(FLERR, "expected 'yes' or 'no' after 'thermo' keyword.");
+    thermo = utils::logical(FLERR, arg[1], false, lmp) == 1;
+    return 2;
+  } else
+    return 0;
+}

src/EXTRA-DUMP/dump_yaml.h

@@ -0,0 +1,89 @@
+/* -*- c++ -*- ----------------------------------------------------------
+   LAMMPS - Large-scale Atomic/Molecular Massively Parallel Simulator
+   https://www.lammps.org/, 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
+// clang-format off
+DumpStyle(yaml,DumpYAML);
+// clang-format on
+#else
+
+#ifndef LMP_DUMP_YAML_H
+#define LMP_DUMP_YAML_H
+
+#include "dump_custom.h"
+
+namespace LAMMPS_NS {
+
+class DumpYAML : public DumpCustom {
+ public:
+  DumpYAML(class LAMMPS *, int, char **);
+
+protected:
+  bool thermo;
+
+  void init_style() override;
+  void write() override;
+  void write_header(bigint) override;
+  void write_data(int, double *) override;
+
+  int modify_param(int, char **) override;
+};
+
+}    // namespace LAMMPS_NS
+
+#endif
+#endif
+
+/* ERROR/WARNING messages:
+
+E: Cannot open dump file %s
+
+The output file for the dump command cannot be opened.  Check that the
+path and name are correct.
+
+E: Too much per-proc info for dump
+
+Number of local atoms times number of columns must fit in a 32-bit
+integer for dump.
+
+E: Dump_modify format line is too short
+
+UNDOCUMENTED
+
+E: Could not find dump custom compute ID
+
+Self-explanatory.
+
+E: Could not find dump custom fix ID
+
+Self-explanatory.
+
+E: Dump custom and fix not computed at compatible times
+
+The fix must produce per-atom quantities on timesteps that dump custom
+needs them.
+
+E: Could not find dump custom variable name
+
+Self-explanatory.
+
+E: Region ID for dump custom does not exist
+
+Self-explanatory.
+
+U: Dump_modify format string is too short
+
+There are more fields to be dumped in a line of output than your
+format string specifies.
+
+*/

src/output.cpp

@@ -192,7 +192,7 @@ void Output::setup(int memflag)
   // decide whether to write snapshot and/or calculate next step for dump
 
   if (ndump && update->restrict_output == 0) {
-    next_time_dump_any = MAXBIGINT;
+    next_dump_any = next_time_dump_any = MAXBIGINT;
 
     for (int idump = 0; idump < ndump; idump++) {
 
@@ -256,8 +256,7 @@ void Output::setup(int memflag)
 
       if (mode_dump[idump] && (dump[idump]->clearstep || var_dump[idump]))
         next_time_dump_any = MIN(next_time_dump_any,next_dump[idump]);
-      if (idump) next_dump_any = MIN(next_dump_any,next_dump[idump]);
+      next_dump_any = MIN(next_dump_any,next_dump[idump]);
-      else next_dump_any = next_dump[0];
     }
 
   // if no dumps, set next_dump_any to last+1 so will not influence next
@@ -356,9 +355,9 @@ void Output::setup(int memflag)
    //     what other command may have added it
 
    if (next_dump_any == ntimestep) {
+     next_dump_any = next_time_dump_any = MAXBIGINT;
 
      for (int idump = 0; idump < ndump; idump++) {
-       next_time_dump_any = MAXBIGINT;
 
        if (next_dump[idump] == ntimestep) {
          if (last_dump[idump] == ntimestep) continue;
@@ -381,8 +380,7 @@ void Output::setup(int memflag)
 
        if (mode_dump[idump] && (dump[idump]->clearstep || var_dump[idump]))
          next_time_dump_any = MIN(next_time_dump_any,next_dump[idump]);
-       if (idump) next_dump_any = MIN(next_dump_any,next_dump[idump]);
+       next_dump_any = MIN(next_dump_any,next_dump[idump]);
-       else next_dump_any = next_dump[0];
      }
    }
 

src/thermo.h

@@ -23,6 +23,7 @@ class Thermo : protected Pointers {
   friend class MinCG;              // accesses compute_pe
   friend class DumpNetCDF;         // accesses thermo properties
   friend class DumpNetCDFMPIIO;    // accesses thermo properties
+  friend class DumpYAML;           // accesses thermo properties
 
  public:
   char *style;

unittest/python/python-formats.py

@@ -2,6 +2,12 @@ import os
 import unittest
 from lammps.formats import LogFile, AvgChunkFile
 
+import yaml
+try:
+    from yaml import CSafeLoader as Loader, CSafeDumper as Dumper
+except ImportError:
+    from yaml import SafeLoader, SafeDumper
+
 EXAMPLES_DIR=os.path.abspath(os.path.join(__file__, '..', '..', '..', 'examples'))
 
 DEFAULT_STYLE_EXAMPLE_LOG="melt/log.8Apr21.melt.g++.1"
@@ -109,5 +115,54 @@ class AvgChunkFiles(unittest.TestCase):
         self.assertEqual(len(chunk['coord'][0]), 1)
 
 
+from lammps import lammps
+has_dump_yaml = False
+try:
+    machine=None
+    if 'LAMMPS_MACHINE_NAME' in os.environ:
+        machine=os.environ['LAMMPS_MACHINE_NAME']
+    lmp=lammps(name=machine)
+    has_dump_yaml = lmp.has_style("atom","full") and lmp.has_style("dump", "yaml")
+    lmp.close()
+except:
+    pass
+
+@unittest.skipIf(not has_dump_yaml, "Either atom_style full or dump_style yaml are not available")
+class PythonDump(unittest.TestCase):
+    def setUp(self):
+        machine = None
+        if 'LAMMPS_MACHINE_NAME' in os.environ:
+            machine=os.environ['LAMMPS_MACHINE_NAME']
+        self.lmp = lammps(name=machine,  cmdargs=['-nocite', '-log','none', '-echo','screen'])
+
+    def tearDown(self):
+        del self.lmp
+
+    def testDumpYaml(self):
+        dumpfile = os.path.join(os.path.abspath('.'), 'dump.yaml')
+        self.lmp.command('shell cd ' + os.environ['TEST_INPUT_DIR'])
+        self.lmp.command("newton on on")
+        self.lmp.file("in.fourmol")
+        self.lmp.command("dump 1 all yaml 2 " + dumpfile + " id type mol q x y z vx vy vz")
+        self.lmp.command("dump_modify 1 time yes sort id units yes")
+        self.lmp.command("run 4 post no")
+        with open(dumpfile) as d:
+            traj = tuple(yaml.load_all(d, Loader=Loader))
+        self.assertEqual(len(traj), 3)
+        self.assertEqual(traj[0]['timestep'], 0)
+        self.assertEqual(traj[0]['time'], 0)
+        self.assertEqual(traj[0]['natoms'], 29)
+        self.assertEqual(traj[0]['units'], 'real')
+        self.assertEqual(len(traj[0]['boundary']), 6)
+        self.assertEqual(traj[0]['boundary'][0], 'p')
+        self.assertEqual(traj[1]['timestep'], 2)
+        self.assertEqual(traj[1]['time'], 0.2)
+        self.assertEqual(traj[2]['timestep'], 4)
+        self.assertEqual(traj[2]['time'], 0.4)
+        self.assertEqual(traj[0]['keywords'],['id', 'type', 'mol', 'q', 'x', 'y', 'z',
+                                              'vx', 'vy', 'vz'])
+        self.assertEqual(traj[0]['data'][0],[1, 3, 1, -0.47, -0.279937, 2.47266, -0.172009,
+                                             0.000778678, 0.000589703, -0.000221795])
+
 if __name__ == "__main__":
     unittest.main()