Make CRD EXT format selectable in CRDWriter (#3635)

* Fixes issue #3605
* Inclusion of an boolean kwarg extension=False|True for the CRDWriter.write() function to
   force use of the EXTENDED output format
* add tests for EXT writing
* update CHANGELOG
* update AUTHORS

package/AUTHORS

@@ -191,7 +191,8 @@ Chronological list of authors
   - Sukeerti T
   - Robot Jelly
   - Mark Verma
-
+  - Marcelo D. Poleto
+
 
 External code
 -------------

package/CHANGELOG

@@ -51,6 +51,8 @@ Enhancements
     monatomic ion charges or edge cases with nitrogen, sulfur, phosphorus and
     conjugated systems should now have correctly assigned bond orders and
     charges.
+  * CRD extended format can now be explicitly requested with the `extended` 
+    keyword (Issue #3605)
 
 Changes
   * ITPParser no longer adds an angle for water molecules that have the SETTLE

package/MDAnalysis/coordinates/CRD.py

@@ -125,6 +125,9 @@ class CRDWriter(base.WriterBase):
 
     .. versionchanged:: 0.11.0
        Frames now 0-based instead of 1-based
+    .. versionchanged:: 2.2.0
+       CRD extended format can now be explicitly requested with the
+       `extended` keyword
     """
     format = 'CRD'
     units = {'time': None, 'length': 'Angstrom'}
@@ -151,11 +154,23 @@ def __init__(self, filename, **kwargs):
         ----------
         filename : str or :class:`~MDAnalysis.lib.util.NamedStream`
              name of the output file or a stream
+
+        extended : bool (optional)
+             By default, noextended CRD format is used [``False``].
+             However, extended CRD format can be forced by
+             specifying `extended` ``=True``. Note that the extended format
+             is *always* used if the number of atoms exceeds 99,999, regardless
+             of the setting of `extended`.
+
+             .. versionadded:: 2.2.0
         """
 
         self.filename = util.filename(filename, ext='crd')
         self.crd = None
 
+        # account for explicit crd format, if requested
+        self.extended = kwargs.pop("extended", False)
+
     def write(self, selection, frame=None):
         """Write selection at current trajectory frame to file.
 
@@ -182,14 +197,15 @@ def write(self, selection, frame=None):
             except AttributeError:
                 frame = 0  # should catch cases when we are analyzing a single PDB (?)
 
+
         atoms = selection.atoms  # make sure to use atoms (Issue 46)
         coor = atoms.positions  # can write from selection == Universe (Issue 49)
 
         n_atoms = len(atoms)
         # Detect which format string we're using to output (EXT or not)
         # *len refers to how to truncate various things,
         # depending on output format!
-        if n_atoms > 99999:
+        if self.extended or n_atoms > 99999:
             at_fmt = self.fmt['ATOM_EXT']
             serial_len = 10
             resid_len = 8
@@ -238,7 +254,7 @@ def write(self, selection, frame=None):
             crd.write("*\n")
 
             # Write NUMATOMS
-            if n_atoms > 99999:
+            if self.extended or n_atoms > 99999:
                 crd.write(self.fmt['NUMATOMS_EXT'].format(n_atoms))
             else:
                 crd.write(self.fmt['NUMATOMS'].format(n_atoms))

testsuite/MDAnalysisTests/coordinates/test_crd.py

@@ -24,10 +24,11 @@
 
 import pytest
 from numpy.testing import (
-    assert_equal,
+    assert_equal,assert_allclose
 )
 
 import MDAnalysis as mda
+import os
 
 from MDAnalysisTests.datafiles import CRD
 from MDAnalysisTests import make_Universe
@@ -40,7 +41,7 @@ def u(self):
 
     @pytest.fixture()
     def outfile(self, tmpdir):
-        return str(tmpdir) + '/out.crd'
+        return os.path.join(str(tmpdir), 'test.crd')
 
     def test_write_atoms(self, u, outfile):
         # Test that written file when read gives same coordinates
@@ -66,11 +67,35 @@ def CRD_iter(fn):
         for ref, other in zip(CRD_iter(CRD), CRD_iter(outfile)):
             assert ref == other
 
-    def test_write_EXT(self):
-        # TODO: Write tests that use EXT output format
-        # Must have *lots* of atoms, maybe fake the system
-        # to make tests faster
-        pass
+    def test_write_EXT(self, u, outfile):
+        # Use the `extended` keyword to force the EXT format
+        u.atoms.write(outfile, extended=True)
+
+        with open(outfile, 'r') as inf:
+            format_line = inf.readlines()[2]
+        assert 'EXT' in format_line, "EXT format expected"
+
+    def test_write_EXT_read(self, u, outfile):
+        # Read EXT format and check atom positions
+        u.atoms.write(outfile, extended=True)
+
+        u2 = mda.Universe(outfile)
+
+        sel1 = u.select_atoms('all')
+        sel2 = u2.select_atoms('all')
+
+        cog1 = sel1.center_of_geometry()
+        cog2 = sel2.center_of_geometry()
+
+        assert_equal(len(u.atoms), len(u2.atoms)), 'Equal number of '\
+                'atoms expected in both CRD formats'
+        assert_equal(len(u.atoms.residues),
+            len(u2.atoms.residues)), 'Equal number of residues expected in'\
+                        'both CRD formats'
+        assert_equal(len(u.atoms.segments), 
+            len(u2.atoms.segments)), 'Equal number of segments expected in'\
+                        'both CRD formats'
+        assert_allclose(cog1, cog2, rtol=1e-6, atol=0), 'Same centroid expected for both CRD formats'
 
 
 class TestCRDWriterMissingAttrs(object):