Merge pull request #423 from nisse3000/master

Added Xr class to I/O module

pymatgen/io/tests/test_xr.py

@@ -0,0 +1,77 @@
+# coding: utf-8
+# Copyright (c) Pymatgen Development Team.
+# Distributed under the terms of the MIT License.
+
+from __future__ import division, unicode_literals
+
+__author__ = "Nils Edvin Richard Zimmermann"
+__copyright__ = "Copyright 2016, The Materials Project"
+__version__ = "0.1"
+__maintainer__ = "Nils Edvin Richard Zimmermann"
+__email__ = "nils.e.r.zimmermann@gmail.com"
+__date__ = "June 23, 2016"
+
+import unittest2 as unittest
+import os
+
+from pymatgen.io.xr import Xr
+from pymatgen.io.vasp.inputs import Poscar
+from pymatgen.core.structure import Structure
+
+test_dir = os.path.join(os.path.dirname(__file__), "..", "..", "..",
+                        'test_files')
+
+class XrTest(unittest.TestCase):
+
+    def setUp(self):
+
+        filepath = os.path.join(test_dir, 'POSCAR')
+        p = Poscar.from_file(filepath)
+        self.xr = Xr(p.structure)
+
+    def test_str(self):
+        expected_string = """pymatgen   10.4118 6.0672 4.7595
+90.000 90.000 90.000
+24 0
+0 Fe4 P4 O16
+1 Fe 2.2773 4.5504 2.2601
+2 Fe 2.9285 1.5168 4.6399
+3 Fe 7.4832 4.5504 0.1196
+4 Fe 8.1344 1.5168 2.4994
+5 P 0.9851 1.5168 1.9906
+6 P 4.2208 4.5504 4.3704
+7 P 6.1910 1.5168 0.3891
+8 P 9.4267 4.5504 2.7689
+9 O 0.4516 4.5504 3.3656
+10 O 1.0062 1.5168 3.5283
+11 O 1.7253 0.2795 1.3583
+12 O 1.7253 2.7541 1.3583
+13 O 3.4806 3.3131 3.7380
+14 O 3.4806 5.7876 3.7380
+15 O 4.1997 4.5504 1.1486
+16 O 4.7543 1.5168 0.9859
+17 O 5.6575 4.5504 3.7736
+18 O 6.2121 1.5168 3.6109
+19 O 6.9312 0.2795 1.0215
+20 O 6.9312 2.7541 1.0215
+21 O 8.6864 3.3131 3.4012
+22 O 8.6864 5.7876 3.4012
+23 O 9.4055 4.5504 1.2312
+24 O 9.9602 1.5168 1.3939
+10.4118 0.0000 0.0000
+0.0000 6.0672 0.0000
+0.0000 0.0000 4.7595
+10.4118 0.0000 0.0000
+0.0000 6.0672 0.0000
+0.0000 0.0000 4.7595"""
+        self.assertEqual(str(self.xr), expected_string)
+
+    def test_from_file(self):
+        filename = os.path.join(test_dir, "EDI.xr")
+        xr = Xr.from_file(filename)
+        self.assertIsInstance(xr.structure, Structure)
+        xr2 = Xr.from_file(filename, use_cores=False)
+        self.assertIsInstance(xr2.structure, Structure)
+
+if __name__ == "__main__":
+    unittest.main()

pymatgen/io/xr.py

@@ -0,0 +1,164 @@
+# coding: utf-8
+# Copyright (c) Pymatgen Development Team.
+# Distributed under the terms of the MIT License.
+
+from __future__ import division, unicode_literals
+
+"""
+This module provides input and output mechanisms
+for the xr file format, which is a modified CSSR
+file format and, for example, used in GULP.
+In particular, the module makes it easy
+to remove shell positions from relaxations
+that employed core-shell models.
+"""
+
+
+__author__ = "Nils Edvin Richard Zimmermann"
+__copyright__ = "Copyright 2016, The Materials Project"
+__version__ = "0.1"
+__maintainer__ = "Nils Edvin Richard Zimmermann"
+__email__ = "nils.e.r.zimmermann@gmail.com"
+__date__ = "June 23, 2016"
+
+import re
+
+from six.moves import map
+import numpy as np
+
+from monty.io import zopen
+from math import fabs
+from pymatgen.core.lattice import Lattice
+from pymatgen.core.structure import Structure
+
+class Xr(object):
+    """
+    Basic object for working with xr files.
+
+    Args:
+        structure (Structure/IStructure): Structure object to create the
+                Xr object.
+    """
+
+    def __init__(self, structure):
+        if not structure.is_ordered:
+            raise ValueError("Xr file can only be constructed from ordered "
+                             "structure")
+        self.structure = structure
+
+    def __str__(self):
+        output = ["pymatgen   {:.4f} {:.4f} {:.4f}"
+                  .format(*self.structure.lattice.abc),
+                  "{:.3f} {:.3f} {:.3f}"
+                  .format(*self.structure.lattice.angles),
+                  "{} 0".format(len(self.structure)),
+                  "0 {}".format(self.structure.formula)]
+        # There are actually 10 more fields per site
+        # in a typical xr file from GULP, for example.
+        for i, site in enumerate(self.structure.sites):
+            output.append("{} {} {:.4f} {:.4f} {:.4f}"
+                          .format(i + 1, site.specie, site.x, site.y, site.z))
+        mat = self.structure.lattice.matrix
+        for i in range(2):
+            for j in range(3):
+                output.append("{:.4f} {:.4f} {:.4f}".format(
+                        mat[j][0], mat[j][1], mat[j][2]))
+        return "\n".join(output)
+
+    def write_file(self, filename):
+        """
+        Write out an xr file.
+
+        Args:
+            filename (str): name of the file to write to.
+        """
+        with zopen(filename, 'wt') as f:
+            f.write(str(self) + "\n")
+
+    @staticmethod
+    def from_string(string, use_cores=True, thresh=1.e-4):
+        """
+        Creates an Xr object from a string representation.
+
+        Args:
+            string (str): string representation of an Xr object.
+            use_cores (bool): use core positions and discard shell
+                    positions if set to True (default).  Otherwise,
+                    use shell positions and discard core positions.
+            thresh (float): relative threshold for consistency check
+                    between cell parameters (lengths and angles) from
+                    header information and cell vectors, respectively.
+
+        Returns:
+            xr (Xr): Xr object corresponding to the input
+                    string representation.
+        """
+        lines = string.split("\n")
+        toks = lines[0].split()
+        lengths = [float(toks[i]) for i in range(1, len(toks))]
+        toks = lines[1].split()
+        angles = [float(i) for i in toks[0:3]]
+        toks = lines[2].split()
+        nsites = int(toks[0])
+        mat = np.zeros((3,3), dtype=float)
+        for i in range(3):
+            toks = lines[4+nsites+i].split()
+            toks2 = lines[4+nsites+i+3].split()
+            for j, item in enumerate(toks):
+                if item != toks2[j]:
+                    raise RuntimeError("expected both matrices"
+                            " to be the same in xr file")
+            mat[i] = np.array([float(w) for w in toks])
+        lat = Lattice(mat)
+        if fabs(lat.a-lengths[0])/fabs(lat.a) > thresh or \
+                fabs(lat.b-lengths[1])/fabs(lat.b) > thresh or \
+                fabs(lat.c-lengths[2])/fabs(lat.c) > thresh or \
+                fabs(lat.alpha-angles[0])/fabs(lat.alpha) > thresh or \
+                fabs(lat.beta-angles[1])/fabs(lat.beta) > thresh or \
+                fabs(lat.gamma-angles[2])/fabs(lat.gamma) > thresh:
+            raise RuntimeError("cell parameters in header ("+str(lengths)+\
+                    ", "+str(angles)+") are not consistent with Cartesian"+\
+                    " lattice vectors ("+str(lat.abc)+", "+\
+                    str(lat.angles)+")")
+        # Ignore line w/ index 3.
+        sp = []
+        coords = []
+        for j in range(nsites):
+            m = re.match("\d+\s+(\w+)\s+([0-9\-\.]+)\s+([0-9\-\.]+)\s+" +
+                         "([0-9\-\.]+)", lines[4+j].strip())
+            if m:
+                tmp_sp = m.group(1)
+                if use_cores and tmp_sp[len(tmp_sp)-2:] == "_s":
+                    continue
+                if not use_cores and tmp_sp[len(tmp_sp)-2:] == "_c":
+                    continue
+                if tmp_sp[len(tmp_sp)-2] == "_":
+                    sp.append(tmp_sp[0:len(tmp_sp)-2])
+                else:
+                    sp.append(tmp_sp)
+                coords.append([float(m.group(i)) for i in range(2, 5)])
+        return Xr(Structure(lat, sp, coords, coords_are_cartesian=True))
+
+    @staticmethod
+    def from_file(filename, use_cores=True, thresh=1.e-4):
+        """
+        Reads an xr-formatted file to create an Xr object.
+
+        Args:
+            filename (str): name of file to read from.
+            use_cores (bool): use core positions and discard shell
+                    positions if set to True (default).  Otherwise,
+                    use shell positions and discard core positions.
+            thresh (float): relative threshold for consistency check
+                    between cell parameters (lengths and angles) from
+                    header information and cell vectors, respectively.
+
+        Returns:
+            xr (Xr): Xr object corresponding to the input
+                    file.
+        """
+        with zopen(filename, "rt") as f:
+            return Xr.from_string(
+                    f.read(), use_cores=use_cores,
+                    thresh=thresh)
+

test_files/EDI.xr

@@ -0,0 +1,35 @@
+somestring 6.8774 6.8774 6.3558
+90.0 90.0 90.0
+25 1 EDI structure
+1 bulk
+ 1 Si  1.84246 -0.00000 0.75253 0 0 0 0 0 0 0 0  4.000  0.00
+ 2 Si  6.85980  1.86010 5.58528 0 0 0 0 0 0 0 0  4.000  0.00
+ 3 Si  4.99970 -0.00000 0.75253 0 0 0 0 0 0 0 0  4.000  0.00
+ 4 Si  6.85980  5.01733 5.58528 0 0 0 0 0 0 0 0  4.000  0.00
+ 5 Si  6.85980  6.87743 3.16890 0 0 0 0 0 0 0 0  4.000  0.00
+ 6 O_c 1.29725  1.31488 6.34680 0 0 0 0 0 0 0 0  0.869 74.92
+ 7 O_s 1.21428  1.23192 6.34680 0 0 0 0 0 0 0 0 -2.869 74.92
+ 8 O_c 5.54491  1.31488 6.34680 0 0 0 0 0 0 0 0  0.869 74.92
+ 9 O_s 5.62788  1.23192 6.34680 0 0 0 0 0 0 0 0 -2.869 74.92
+10 O_c 5.54491  5.56255 6.34680 0 0 0 0 0 0 0 0  0.869 74.92
+11 O_s 5.62788  5.64551 6.34680 0 0 0 0 0 0 0 0 -2.869 74.92
+12 O_c 1.29725  5.56255 6.34680 0 0 0 0 0 0 0 0  0.869 74.92
+13 O_s 1.21428  5.64551 6.34680 0 0 0 0 0 0 0 0 -2.869 74.92
+14 O_c 1.31620  6.87743 2.27055 0 0 0 0 0 0 0 0  0.869 74.92
+15 O_s 1.22765  6.87743 2.20123 0 0 0 0 0 0 0 0 -2.869 74.92
+16 O_c 6.85980  1.33384 4.06726 0 0 0 0 0 0 0 0  0.869 74.92
+17 O_s 6.85980  1.24529 4.13658 0 0 0 0 0 0 0 0 -2.869 74.92
+18 O_c 5.52596  6.87743 2.27055 0 0 0 0 0 0 0 0  0.869 74.92
+19 O_s 5.61451  6.87743 2.20123 0 0 0 0 0 0 0 0 -2.869 74.92
+20 O_c 6.85980  5.54359 4.06726 0 0 0 0 0 0 0 0  0.869 74.92
+21 O_s 6.85980  5.63214 4.13658 0 0 0 0 0 0 0 0 -2.869 74.92
+22 O_c 3.42108 -0.00000 0.75105 0 0 0 0 0 0 0 0  0.869 74.92
+23 O_s 3.42108 -0.00000 0.75165 0 0 0 0 0 0 0 0 -2.869 74.92
+24 O_c 6.85980  3.43871 5.58676 0 0 0 0 0 0 0 0  0.869 74.92
+25 O_s 6.85980  3.43871 5.58616 0 0 0 0 0 0 0 0 -2.869 74.92
+6.877431  0.000000  0.000000
+0.000000  6.877431 -0.000000
+0.000000 -0.000000  6.355796
+6.877431  0.000000  0.000000
+0.000000  6.877431 -0.000000
+0.000000 -0.000000  6.355796