def test_append_normal(self):
""" Tests genopen appending a normal text file """
with closing(genopen(get_fn('test.txt', written=True), 'w')) as f:
f.write(ALPHABET)
with closing(genopen(get_fn('test.txt', written=True), 'a')) as f:
f.write(ALPHABET)
self.assertEqual(open(get_fn('test.txt', written=True)).read(), ALPHABET*2)
def testAppendBzip(self):
""" Tests genopen appending a bzipped file """
with closing(genopen(get_fn('test.txt.bz2', written=True), 'a')) as f:
f.write(ALPHABET)
with closing(genopen(get_fn('test.txt.bz2', written=True), 'a')) as f:
f.write(ALPHABET)
text = bz2.BZ2File(get_fn('test.txt.bz2', written=True)).read()
self.assertEqual(text.decode('ascii'), ALPHABET*2)
def test_append_gzip(self):
""" Tests genopen appending a gzipped file """
with closing(genopen(get_fn('test.txt.gz', written=True), 'a')) as f:
f.write(ALPHABET)
with closing(genopen(get_fn('test.txt.gz', written=True), 'a')) as f:
f.write(ALPHABET)
text = gzip.open(get_fn('test.txt.gz', written=True)).read()
self.assertEqual(text.decode('ascii'), ALPHABET*2)
def testWriteRemoteFile(self):
""" Tests that genopen writing a remote file fails """
url = 'http://q4md-forcefieldtools.org/REDDB/projects/W-73/tripos1.mol2'
self.assertRaises(ValueError, lambda: genopen(url, 'w'))
try:
genopen(url, 'w')
self.assertTrue(False)
except ValueError as e:
self.assertEqual(str(e), 'Cannot write or append a webpage')
def rdparm_slow(self, fname):
"""
Parses the Amber format file. This parser is written in pure Python and
is therefore slower than the C++-optimized version
"""
current_flag = ''
fmtre = re.compile(r'%FORMAT *\((.+)\)')
version = None
# Open up the file and read the data into memory
with closing(genopen(fname, 'r')) as prm:
for line in prm:
if line[0] == '%':
if line[0:8] == '%VERSION':
self.version = line.strip()
continue
elif line[0:5] == '%FLAG':
current_flag = line[6:].strip()
self.formats[current_flag] = ''
self.parm_data[current_flag] = []
self.parm_comments[current_flag] = []
self.flag_list.append(current_flag)
continue
elif line[0:8] == '%COMMENT':
self.parm_comments[current_flag].append(line[9:].strip())
continue
elif line[0:7] == '%FORMAT':
fmt = FortranFormat(fmtre.match(line).groups()[0])
# RESIDUE_ICODE can have a lot of blank data...
if current_flag == 'RESIDUE_ICODE':
fmt.read = fmt._read_nostrip
self.formats[current_flag] = fmt
continue
try:
self.parm_data[current_flag].extend(fmt.read(line))
except KeyError:
if version is not None:
raise
break # Skip out of the loop down to the old-format parser
# convert charges to fraction-electrons
if 'CTITLE' in self.parm_data:
CHARGE_SCALE = CHARMM_ELECTROSTATIC
else:
CHARGE_SCALE = AMBER_ELECTROSTATIC
if self.charge_flag in self.parm_data:
for i, chg in enumerate(self.parm_data[self.charge_flag]):
self.parm_data[self.charge_flag][i] = chg / CHARGE_SCALE
# If we don't have a version, then read in an old-file topology
if self.version is None:
with closing(genopen(self.name, 'r')) as f:
self.rdparm_old(f.readlines())
return
def id_format(filename):
"""
Identifies the file type as either Amber-format file (like prmtop) or an
old-style topology file.
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is an Amber-style format, False otherwise
"""
if isinstance(filename, string_types):
with closing(genopen(filename, 'r')) as f:
lines = [f.readline() for i in range(5)]
elif (hasattr(filename, 'readline') and hasattr(filename, 'seek')
and hasattr(filename, 'tell')):
cur = filename.tell()
lines = [filename.readline() for i in range(5)]
filename.seek(cur)
if lines[0].startswith('%VERSION'):
return True
# Try old-style format
try:
return AmberFormat().rdparm_old(lines, check=True)
except ValueError:
return False
def __init__(self, fname, natom, hasbox, mode='r', title=None):
if mode == 'r':
self._status = 'old'
elif mode == 'w':
self._status = 'new'
# We need to have some way to know whether we need to write the
# coordinates or the box for this particular frame. Each frame must
# be written as coordinates first, then box.
self._writebox = False
else:
raise ValueError("%s mode must be 'r' or 'w'" % type(self).__name__)
self._file = genopen(fname, mode)
self.natom = natom
self.hasbox = hasbox
self._full_lines_per_frame = self.natom * 3 // self.CRDS_PER_LINE
self._nextras = self.natom * 3 - (self._full_lines_per_frame *
self.CRDS_PER_LINE)
self.closed = False
if self._status == 'old':
self._parse()
elif self._status == 'new':
if title is None:
if self.DEFAULT_TITLE is None:
raise NotImplemented('This object must be subclassed')
self._file.write('%s\n' % self.DEFAULT_TITLE)
else:
self._file.write(title.rstrip() + '\n')
def id_format(filename):
""" Identifies the file type as an Amber mdcrd file
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is an Amber mdcrd file. False otherwise
"""
f = genopen(filename, 'r')
lines = [f.readline() for i in range(5)]
f.close()
# Next 4 lines, make sure we have %8.3f format
try:
for i in range(4):
i += 1
for j in range(10):
j8 = j * 8
if lines[i][j8+4] != '.': return False
float(lines[i][j8:j8+8])
if lines[i][j8+7] not in '0123456789':
return False
except (IndexError, ValueError):
return False
# Must be a mdcrd
return True
def write(struct, dest):
""" Writes a CHARMM coordinate file from a structure
Parameters
----------
struct : :class:`parmed.structure.Structure`
The input structure to write the CHARMM coordinate file from
dest : str or file-like object
The file name or file object to write the coordinate file to
"""
if isinstance(dest, string_types):
dest = io.genopen(dest, 'w')
own_handle = True
else:
own_handle = False
dest.write('* GENERATED BY PARMED (HTTPS://GITHUB.COM/PARMED/PARMED)\n')
dest.write('*\n')
dest.write('%10d EXT\n' % len(struct.atoms))
add = 0 if struct.residues[0].number > 0 else 1-struct.residues[0].number
for i, atom in enumerate(struct.atoms):
res = atom.residue
segid = res.segid.strip() or res.chain.strip() or 'SYS'
dest.write('%10d%10d %-8s %-8s%20.10f%20.10f%20.10f %-8s '
'%-8s%20.10f\n' % (i+1, atom.residue.number+add,
atom.residue.name, atom.name, atom.xx, atom.xy, atom.xz,
segid, atom.residue.number, 0))
if own_handle:
dest.close()
def __init__(self, fname, defines=None, includes=None, notfound_fatal=True):
if isinstance(fname, string_types):
self._fileobj = genopen(fname, 'r')
self._ownhandle = True
curpath = path.abspath(path.split(fname)[0])
self.filename = fname
else:
self._fileobj = fname
self._ownhandle = False
curpath = path.abspath(path.curdir)
self.filename = None
if includes is None:
self._includes = [curpath]
else:
self._includes = [curpath] + list(includes)
if defines is None:
self.defines = OrderedDict()
else:
# Convert every define to a string
self.defines = OrderedDict()
for define, value in iteritems(defines):
self.defines[define] = str(value)
self._notfound_fatal = notfound_fatal
# Now to keep track of other basic logic stuff
self.included_files = []
self._ifstack = []
self._elsestack = []
self._satisfiedstack = []
self._num_ignoring_if = 0
self._includefile = None
def load_parameters(self, fname):
""" Load a set of parameters from a single parameter file
Parameters
----------
fname : str or file-like
Parameter file to parse
"""
if isinstance(fname, string_types):
f = genopen(fname, 'r')
own_handle = True
else:
f = fname
own_handle = False
self.titles.append(f.readline().strip())
try:
for line in f:
if not line.strip():
return self._parse_frcmod(f, line)
elif line.strip() in ('MASS', 'BOND', 'ANGLE', 'ANGL', 'DIHE',
'DIHED', 'DIHEDRAL', 'IMPR', 'IMPROP',
'IMPROPER', 'NONB', 'NONBON', 'NONBOND',
'NONBONDED'):
return self._parse_frcmod(f, line)
else:
return self._parse_parm_dat(f, line)
finally:
if own_handle:
f.close()
def id_format(filename):
""" Identifies the file type as an XML file
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is an XML format, False otherwise
"""
with closing(genopen(filename, 'r')) as f:
for line in f:
line = line.strip()
if not line: continue
rematch = _xmlre.match(line)
if not rematch:
return False
stuff = rematch.groups()[0]
if stuff[0] in '?!':
continue
kind = stuff.split()[0]
if kind in ('System', 'State', 'ForceField', 'Integrator'):
return True
return False
def id_format(filename):
"""
Identifies the file type as either Amber-format file (like prmtop) or an
old-style topology file.
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is an Amber-style format, False otherwise
"""
with closing(genopen(filename, 'r')) as f:
lines = [f.readline() for i in range(5)]
if lines[0].startswith('%VERSION'):
return True
# Try old-style format
try:
return AmberFormat().rdparm_old(lines, check=True)
except ValueError:
return False
def parse(filename):
""" Parses an Amber OFF library
Parameters
----------
filename : str or file-like iterable
The file name or file object to parse. If it is an iterable, it will
be exhausted
Returns
-------
residues : OrderedDict {str : :class:`ResidueTemplate`}
Dictionary pairing residue names with their :class:`ResidueTemplate`
objects
Raises
------
ValueError if the first line does not match the file format. This line
will be consumed
IOError if filename is the name of a file that does not exist
RuntimeError if EOF is reached prematurely or other formatting issues
found
"""
if isinstance(filename, string_types):
fileobj = genopen(filename, 'r')
own_handle = True
else:
fileobj = filename
own_handle = False
# Now parse the library file
line = fileobj.readline()
if not AmberOFFLibrary._headerre.match(line):
raise ValueError('Unrecognized OFF file format')
# Build the return value
residues = OrderedDict()
# Pull a list of all the residues we expect to find
line = fileobj.readline()
rematch = AmberOFFLibrary._resre.match(line)
while rematch and line:
name = rematch.groups()[0]
residues[name] = None
line = fileobj.readline()
rematch = AmberOFFLibrary._resre.match(line)
if not line:
raise RuntimeError('Unexpected EOF in Amber OFF library')
# Now make sure we have the next expected line
while line:
rematch = AmberOFFLibrary._sec1re.match(line)
if not rematch:
raise RuntimeError('Expected atoms table not found')
name = rematch.groups()[0]
residues[name] = AmberOFFLibrary._parse_residue(fileobj, name)
line = fileobj.readline()
if own_handle: fileobj.close()
return residues
def rdparm(self, fname, slow=False):
""" Parses the Amber format file """
self.name = fname
self.version = None # reset all top info each time rdparm is called
self.formats = {}
self.parm_data = {}
self.parm_comments = {}
self.flag_list = []
# See if we have the optimized parser available
try:
from parmed.amber import _rdparm
except ImportError:
return self.rdparm_slow(fname)
# The optimized parser only works on local, uncompressed files
# TODO: Add gzip and bzip2 support to the optimized reader
if (hasattr(fname, 'read') or slow
or fname.startswith('http://') or fname.startswith('https://')
or fname.startswith('ftp://')
or fname.endswith('.bz2') or fname.endswith('.gz')):
return self.rdparm_slow(fname)
# We have the optimized version and a local file
try:
ret = _rdparm.rdparm(fname)
except TypeError:
# This is raised if VERSION is not found
with closing(genopen(fname, 'r')) as f:
return self.rdparm_old(f.readlines())
else:
# Unpack returned contents
parm_data, parm_comments, formats, unkflg, flag_list, version = ret
# Now assign them to instance attributes and process where necessary
self.parm_data = parm_data
self.parm_comments = parm_comments
for key in formats:
self.formats[key] = FortranFormat(formats[key])
self.flag_list = flag_list
self.version = version
# Now we have to process all of those sections that the optimized
# parser couldn't figure out
for flag in unkflg:
rawdata = self.parm_data[flag]
self.parm_data[flag] = []
for line in rawdata:
self.parm_data[flag].extend(self.formats[flag].read(line))
if 'CTITLE' in self.parm_data:
CHARGE_SCALE = CHARMM_ELECTROSTATIC
else:
CHARGE_SCALE = AMBER_ELECTROSTATIC
try:
for i, chg in enumerate(self.parm_data[self.charge_flag]):
self.parm_data[self.charge_flag][i] = chg / CHARGE_SCALE
except KeyError:
pass
def id_format(filename):
""" Identifies the file type as an Amber restart/inpcrd file
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is an Amber restart/inpcrd file. False otherwise
"""
f = genopen(filename, 'r')
lines = [f.readline() for i in range(5)]
f.close()
# Look for natom
words = lines[1].split()
if len(words) > 2 or len(words) < 1:
return False
try:
natom = int(words[0])
float(words[1])
except ValueError:
return False
except IndexError:
pass
# Next 3 lines, make sure we have %12.7f format. This only works if we
# have at least 6 atoms. Any fewer than that means the restart file is
# shorter than that.
try:
if natom <= 0:
return False
i = 0
for line in lines[2:]:
i += 1
if i > natom: break
for j in range(3):
j12 = j * 12
if line[j12+4] != '.': return False
float(line[j12:j12+12])
if line[j12+11] not in '0123456789':
return False
i += 1
if i > natom: break
for j in range(3):
j12 = j * 12 + 36
if line[j12+4] != '.': return False
float(line[j12:j12+12])
if line[j12+11] not in '0123456789':
return False
except (IndexError, ValueError):
return False
# Must be a restart...
return True
def __init__(self, fname, mode='r'):
if mode not in ('r', 'w'):
raise ValueError('Cannot open CharmmFile with mode "%s"' % mode)
if mode == 'r':
self.status = 'OLD'
else:
self.status = 'NEW'
self._handle = genopen(fname, mode)
self.closed = False
self.line_number = 0
self.comment = ''
def __init__(self, fname, seq=None):
super(XyzFile, self).__init__()
if isinstance(fname, string_types):
fxyz = genopen(fname, 'r')
own_handle_xyz = True
else:
fxyz = fname
own_handle_xyz = False
if seq is not None:
seqstruct = load_file(seq)
# Now parse the file
try:
natom = int(fxyz.readline().split()[0])
except (ValueError, IndexError):
raise TinkerError('Bad XYZ file format; first line')
if seq is not None and natom != len(seqstruct.atoms):
raise ValueError('Sequence file %s # of atoms does not match the # '
'of atoms in the XYZ file' % seq)
words = fxyz.readline().split()
if len(words) == 6 and not XyzFile._check_atom_record(words):
self.box = [float(w) for w in words]
words = fxyz.readline().split()
atom = Atom(atomic_number=AtomicNum[element_by_name(words[1])],
name=words[1], type=words[5])
atom.xx, atom.xy, atom.xz = [float(w) for w in words[2:5]]
residue = Residue('SYS')
residue.number = 1
residue._idx = 0
if seq is not None:
residue = seqstruct.residues[0]
self.add_atom(atom, residue.name, residue.number, residue.chain,
residue.insertion_code, residue.segid)
bond_ids = [[int(w) for w in words[6:]]]
for i, line in enumerate(fxyz):
words = line.split()
atom = Atom(atomic_number=AtomicNum[element_by_name(words[1])],
name=words[1], type=words[5])
atom.xx, atom.xy, atom.xz = [float(w) for w in words[2:5]]
if seq is not None:
residue = seqstruct.atoms[i+1].residue
self.add_atom(atom, residue.name, residue.number, residue.chain,
residue.insertion_code, residue.segid)
bond_ids.append([int(w) for w in words[6:]])
# All of the bonds are stored now -- go ahead and make them now
for atom, bonds in zip(self.atoms, bond_ids):
i = atom.idx + 1
for idx in bonds:
if idx > i:
self.bonds.append(Bond(atom, self.atoms[idx-1]))
if own_handle_xyz:
fxyz.close()
def write_parm(self, name):
"""
Writes the current data in parm_data into a new topology file with
the given name
Parameters
----------
name : str
Name of the file to write the topology file to
"""
# now that we know we will write the new prmtop file, open the new file
if isinstance(name, string_types):
new_prm = genopen(name, 'w')
own_handle = True
else:
new_prm = name
own_handle = False
try:
# get current time to put into new prmtop file if we had a %VERSION
self.set_version()
# convert charges back to amber charges...
if 'CTITLE' in self.parm_data:
CHARGE_SCALE = CHARMM_ELECTROSTATIC
else:
CHARGE_SCALE = AMBER_ELECTROSTATIC
if self.charge_flag in self.parm_data.keys():
for i in range(len(self.parm_data[self.charge_flag])):
self.parm_data[self.charge_flag][i] *= CHARGE_SCALE
# write version to top of prmtop file
new_prm.write('%s\n' % self.version)
# write data to prmtop file, inserting blank line if it's an empty field
for flag in self.flag_list:
new_prm.write('%%FLAG %s\n' % flag)
# Insert any comments before the %FORMAT specifier
for comment in self.parm_comments[flag]:
new_prm.write('%%COMMENT %s\n' % comment)
new_prm.write('%%FORMAT(%s)\n' % self.formats[flag])
if len(self.parm_data[flag]) == 0: # empty field...
new_prm.write('\n')
continue
self.formats[flag].write(self.parm_data[flag], new_prm)
finally:
if own_handle:
new_prm.close()
if self.charge_flag in self.parm_data.keys():
# Convert charges back to electron-units
for i in range(len(self.parm_data[self.charge_flag])):
self.parm_data[self.charge_flag][i] /= CHARGE_SCALE
def id_format(filename):
""" Identifies the file type as a CHARMM restart file
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is a CHARMM restart file
"""
with closing(io.genopen(filename)) as f:
line = f.readline()
return line.startswith('REST')
def _parse(self, fname):
with closing(io.genopen(fname, 'r')) as crdfile:
readingHeader = True
while readingHeader:
line = crdfile.readline()
if not len(line):
raise CharmmError('Premature end of file')
line = line.strip()
words = line.split()
if len(line) != 0:
if words[0] == 'ENERGIES' or words[0] == '!ENERGIES':
readingHeader = False
else:
self.header.append(line.strip())
else:
self.header.append(line.strip())
for row in range(len(self.header)):
if len(self.header[row].strip()) != 0:
line = self.header[row].strip().split()
if line[0][0:5] == 'NATOM' or line[0][0:6] == '!NATOM':
try:
line = self.header[row+1].strip().split()
self.natom = int(line[0])
self.npriv = int(line[1]) # num. previous steps
self.nstep = int(line[2]) # num. steps in file
self.nsavc = int(line[3]) # coord save frequency
self.nsavv = int(line[4]) # velocities "
self.jhstrt = int(line[5]) # Num total steps?
break
except (ValueError, IndexError):
raise CharmmError('Problem parsing CHARMM restart')
self.scan(crdfile, '!XOLD')
self._get_formatted_crds(crdfile, self.coordsold)
self.coordsold = np.array(self.coordsold).reshape((-1,self.natom,3))
self.scan(crdfile, '!VX')
self._get_formatted_crds(crdfile, self.vels)
self.vels = np.array(self.vels).reshape((-1, self.natom, 3))
# Convert velocities to angstroms/ps
self.vels *= ONE_TIMESCALE
self.scan(crdfile, '!X')
self._get_formatted_crds(crdfile, self.coords)
self.coords = np.array(self.coords).reshape((-1, self.natom, 3))
def id_format(filename):
""" Identifies the file type as a CHARMM coordinate file
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is a CHARMM coordinate file
"""
with closing(io.genopen(filename)) as f:
line = f.readline()
while line and len(line.strip()) == 0: # Skip whitespace
line = f.readline()
intitle = True
while intitle:
line = f.readline()
if len(line.strip()) == 0:
intitle = False
elif line[0] != '*':
intitle = False
else:
intitle = True
while line and len(line.strip()) == 0: # Skip whitespace
line = f.readline()
try:
natom = int(line.split()[0])
for row in range(min(natom, 3)):
line = f.readline().split()
int(line[0])
int(line[1])
float(line[4])
float(line[5])
float(line[6])
int(line[8])
float(line[9])
except (IndexError, ValueError):
return False
return True
def id_format(filename):
""" Identifies the file type as a CHARMM PSF file
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is a CHARMM or Xplor-style PSF file
"""
f = genopen(filename, 'r')
line = f.readline()
f.close()
return line.strip().startswith('PSF')
def id_format(filename):
""" Sees if an open file is an OFF library file.
Parameters
----------
filename : str
The name of the file to see if it is an OFF file format
Returns
-------
is_fmt : bool
True if it is recognized as OFF, False otherwise
"""
with closing(genopen(filename, 'r')) as f:
if AmberOFFLibrary._headerre.match(f.readline()):
return True
return False
def _parse(self, fname):
with closing(io.genopen(fname, 'r')) as crdfile:
line = crdfile.readline().strip()
while len(line) == 0: # Skip whitespace, as a precaution
line = crdfile.readline().strip()
intitle = True
while intitle:
self.title.append(line)
line = crdfile.readline().strip()
if len(line) == 0:
intitle = False
elif line[0] != '*':
intitle = False
else:
intitle = True
while len(line) == 0: # Skip whitespace
line = crdfile.readline().strip()
try:
self.natom = int(line.split()[0])
for row in range(self.natom):
line = crdfile.readline().split()
self.atomno.append(int(line[0]))
self.resno.append(int(line[1]))
self.resname.append(line[2])
self.attype.append(line[3])
self.coords.append(float(line[4]))
self.coords.append(float(line[5]))
self.coords.append(float(line[6]))
self.segid.append(line[7])
self.resid.append(int(line[8]))
self.weighting.append(float(line[9]))
if 3*self.natom != len(self.coords):
raise CharmmError("Error parsing CHARMM .crd file: %d "
"atoms requires %d coords (not %d)" %
(self.natom, 3*self.natom,
len(self.coords))
)
except (ValueError, IndexError):
raise CharmmError('Error parsing CHARMM coordinate file')
self.coords = np.array(self.coords).reshape((-1, self.natom, 3))
def parse(filename):
"""
Parses XML file and returns deserialized object. The return value
depends on the serialized object, summarized below
- System : returns simtk.openmm.System
- State : returns simtk.openmm.State
- Integrator : returns simtk.openmm.Integrator subclass
- ForceField : returns simtk.openmm.app.ForceField
Parameters
----------
filename : str or file-like
The file name or file object containing the XML-serialized object
Returns
-------
obj : System, State, Integrator, or ForceField
The deserialized object
Notes
-----
OpenMM requires the entire contents of this file read into memory. As a
result, this function may require a significant amount of memory.
"""
if isinstance(filename, string_types):
with closing(genopen(filename, 'r')) as f:
contents = f.read()
else:
contents = f.read()
# ForceField is not handled by XmlSerializer
if '<ForceField' in contents:
obj = StringIO()
obj.write(contents)
obj.seek(0)
return app.ForceField(obj)
obj = mm.XmlSerializer.deserialize(contents)
if isinstance(obj, (mm.System, mm.Integrator)):
return obj
elif isinstance(obj, mm.State):
return _OpenMMStateContents(obj)
return
def id_format(filename):
""" Identifies the file as a GROMACS GRO file
Parameters
----------
filename : str
Name of the file to check if it is a Gromacs GRO file
Returns
-------
is_fmt : bool
If it is identified as a Gromacs GRO file, return True. False
otherwise
"""
with closing(genopen(filename)) as f:
f.readline() # Title line
try:
int(f.readline().strip()) # number of atoms
except ValueError:
return False
line = f.readline()
try:
int(line[:5])
if not line[5:10].strip(): return False
if not line[10:15].strip(): return False
int(line[15:20])
pdeci = [i for i, x in enumerate(line) if x == '.']
ndeci = pdeci[1] - pdeci[0] - 5
for i in range(1, 4):
wbeg = (pdeci[0]-4)+(5+ndeci)*(i-1)
wend = (pdeci[0]-4)+(5+ndeci)*i
float(line[wbeg:wend])
i = 4
wbeg = (pdeci[0]-4)+(5+ndeci)*(i-1)
wend = (pdeci[0]-4)+(5+ndeci)*i
if line[wbeg:wend].strip():
for i in range(4, 7):
wbeg = (pdeci[0]-4)+(5+ndeci)*(i-1)
wend = (pdeci[0]-4)+(5+ndeci)*i
float(line[wbeg:wend])
except ValueError:
return False
return True
def id_format(filename):
""" Identify the file as a Mol2 (or Mol3) file format or not
Parameters
----------
filename : str
Name of the file to test whether or not it is a mol2 file
Returns
-------
is_fmt : bool
True if it is a mol2 (or mol3) file, False otherwise
"""
f = genopen(filename, 'r')
try:
for line in f:
if line.startswith('#'): continue
if not line.strip(): continue
return line.startswith('@<TRIPOS>')
return False
finally:
f.close()
def id_format(filename):
""" Identify the file as a Tinker XYZ file
Parameters
----------
filename : str
Name of the file to test whether or not it is a mol2 file
Returns
-------
is_fmt : bool
True if it is a xyz file, False otherwise
"""
f = genopen(filename, 'r')
words = f.readline().split() # natom and title
if not words:
return False
try:
natom = int(words[0])
except (ValueError, IndexError):
return False
else:
if natom <= 0:
return False
words = f.readline().split()
# Either a box line or a line with the first atom
if len(words) == 6:
try:
[float(w) for w in words]
except ValueError:
if XyzFile._check_atom_record(words):
return True
return False
else:
# Next line needs to be an atom record
words = f.readline().split()
return XyzFile._check_atom_record(words)
def write(lib, dest):
""" Writes a dictionary of ResidueTemplate units to a file in OFF format
Parameters
----------
lib : dict {str : :class:`ResidueTemplate`}
Items can be either :class:`ResidueTemplate` or
:class:`ResidueTemplateContainer` instances
dest : str or file-like
Either a file name or a file-like object to write the file to
"""
own_handle = False
if not hasattr(dest, 'write'):
dest = genopen(dest, 'w')
own_handle = True
# Write the residues in alphabetical order
names = sorted(lib.keys())
dest.write('!!index array str\n')
for name in names:
dest.write(' "%s"\n' % name)
for name in names:
AmberOFFLibrary._write_residue(dest, lib[name])
if own_handle: dest.close()
def test_append_remote_file(self):
""" Tests that genopen appending a remote file fails """
url = 'http://q4md-forcefieldtools.org/REDDB/projects/W-73/tripos1.mol2'
self.assertRaises(ValueError, lambda: genopen(url, 'a'))
self.assertRaises(ValueError, lambda: genopen(url, 'w'))
def test_read_gzipped_URL(self):
""" Tests genopen reading a gzipped remote file """
url = 'https://github.com/ParmEd/ParmEd/raw/master/test/files/4lzt.pdb.gz'
with closing(genopen(url, 'r')) as f:
self.assertEqual(f.read(), genopen(get_fn('4lzt.pdb.gz')).read())
def load_file(filename, *args, **kwargs):
"""
Identifies the file format of the specified file and returns its parsed
contents.
Parameters
----------
filename : str
The name of the file to try to parse. If the filename starts with
http:// or https:// or ftp://, it is treated like a URL and the file will be
loaded directly from its remote location on the web
structure : object, optional
For some classes, such as the Mol2 file class, the default return object
is not a Structure, but can be made to return a Structure if the
``structure=True`` keyword argument is passed. To facilitate writing
easy code, the ``structure`` keyword is always processed and only passed
on to the correct file parser if that parser accepts the structure
keyword. There is no default, as each parser has its own default.
natom : int, optional
This is needed for some coordinate file classes, but not others. This is
treated the same as ``structure``, above. It is the # of atoms expected
hasbox : bool, optional
Same as ``structure``, but indicates whether the coordinate file has
unit cell dimensions
skip_bonds : bool, optional
Same as ``structure``, but indicates whether or not bond searching will
be skipped if the topology file format does not contain bond information
(like PDB, GRO, and PQR files).
*args : other positional arguments
Some formats accept positional arguments. These will be passed along
**kwargs : other options
Some formats can only be instantiated with other options besides just a
file name.
Returns
-------
object
The returned object is the result of the parsing function of the class
associated with the file format being parsed
Notes
-----
Compressed files are supported and detected by filename extension. This
applies both to local and remote files. The following names are supported:
- ``.gz`` : gzip compressed file
- ``.bz2`` : bzip2 compressed file
SDF file is loaded via `rdkit` package.
Examples
--------
Load a Mol2 file
>>> load_file('tripos1.mol2')
<ResidueTemplate DAN: 31 atoms; 33 bonds; head=None; tail=None>
Load a Mol2 file as a Structure
>>> load_file('tripos1.mol2', structure=True)
<Structure 31 atoms; 1 residues; 33 bonds; NOT parametrized>
Load an Amber topology file
>>> load_file('trx.prmtop', xyz='trx.inpcrd')
<AmberParm 1654 atoms; 108 residues; 1670 bonds; parametrized>
Load a CHARMM PSF file
>>> load_file('ala_ala_ala.psf')
<CharmmPsfFile 33 atoms; 3 residues; 32 bonds; NOT parametrized>
Load a PDB and CIF file
>>> load_file('4lzt.pdb')
<Structure 1164 atoms; 274 residues; 0 bonds; PBC (triclinic); NOT parametrized>
>>> load_file('4LZT.cif')
<Structure 1164 atoms; 274 residues; 0 bonds; PBC (triclinic); NOT parametrized>
Load a Gromacs topology file -- only works with Gromacs installed
>>> load_file('1aki.ff99sbildn.top')
<GromacsTopologyFile 40560 atoms [9650 EPs]; 9779 residues; 30934 bonds; parametrized>
Load a SDF file -- only works with rdkit installed
>>> load_file('mol.sdf', structure=True)
<Structure 34 atoms; 1 residues; 66 bonds; NOT parametrized>
Raises
------
IOError
If ``filename`` does not exist
parmed.exceptions.FormatNotFound
If no suitable file format can be identified, a TypeError is raised
TypeError
If the identified format requires additional arguments that are not
provided as keyword arguments in addition to the file name
"""
global PARSER_REGISTRY, PARSER_ARGUMENTS
# Check that the file actually exists and that we can read it
if filename.startswith('http://') or filename.startswith('https://')\
or filename.startswith('ftp://'):
# This raises IOError if it does not exist; assert silences linters
with closing(genopen(filename)) as f:
assert f
elif not os.path.exists(filename):
raise IOError('%s does not exist' % filename)
elif not os.access(filename, os.R_OK):
raise IOError('%s does not have read permissions set' % filename)
for name, cls in iteritems(PARSER_REGISTRY):
if not hasattr(cls, 'id_format'):
continue
try:
if cls.id_format(filename):
break
except UnicodeDecodeError:
continue
else:
# We found no file format
raise FormatNotFound('Could not identify file format')
# We found a file format that is compatible. Parse it!
other_args = PARSER_ARGUMENTS[name]
for arg in other_args:
if not arg in kwargs:
raise TypeError(
'%s constructor expects %s keyword argument' % name, arg)
# Pass on the following keywords IFF the target function accepts a target
# keyword. Otherwise, get rid of it: structure, natom, hasbox, skip_bonds
if hasattr(cls, 'parse'):
_prune_argument(cls.parse, kwargs, 'structure')
_prune_argument(cls.parse, kwargs, 'natom')
_prune_argument(cls.parse, kwargs, 'hasbox')
_prune_argument(cls.parse, kwargs, 'skip_bonds')
return cls.parse(filename, *args, **kwargs)
elif hasattr(cls, 'open_old'):
_prune_argument(cls.open_old, kwargs, 'structure')
_prune_argument(cls.open_old, kwargs, 'natom')
_prune_argument(cls.open_old, kwargs, 'hasbox')
_prune_argument(cls.open_old, kwargs, 'skip_bonds')
return cls.open_old(filename, *args, **kwargs)
elif hasattr(cls, 'open'):
_prune_argument(cls.open, kwargs, 'structure')
_prune_argument(cls.open, kwargs, 'natom')
_prune_argument(cls.open, kwargs, 'hasbox')
_prune_argument(cls.open, kwargs, 'skip_bonds')
return cls.open(filename, *args, **kwargs)
_prune_argument(cls.__init__, kwargs, 'structure')
_prune_argument(cls.__init__, kwargs, 'natom')
_prune_argument(cls.__init__, kwargs, 'hasbox')
_prune_argument(cls.__init__, kwargs, 'skip_bonds')
return cls(filename, *args, **kwargs)
def test_read_bzipped(self):
""" Tests genopen reading a bzipped file """
with closing(genopen(get_fn('4lzt.pdb.bz2'), 'r')) as f:
text = bz2.BZ2File(get_fn('4lzt.pdb.bz2'), 'r').read()
self.assertEqual(text.decode('ascii'), f.read())
def test_read_gzipped(self):
""" Tests genopen reading a gzipped file """
with closing(genopen(get_fn('4lzt.pdb.gz'))) as f:
text = gzip.open(get_fn('4lzt.pdb.gz'), 'r').read()
self.assertEqual(f.read(), text.decode('ascii'))
def test_read_normal(self):
""" Tests genopen reading a normal text file """
with closing(genopen(get_fn('4lzt.pdb'))) as f:
self.assertEqual(f.read(), open(get_fn('4lzt.pdb')).read())
with closing(genopen(get_fn('4lzt.pdb'), 'r')) as f:
self.assertEqual(f.read(), open(get_fn('4lzt.pdb')).read())
def write(self,
dest,
provenance=None,
write_unused=True,
separate_ljforce=False,
improper_dihedrals_ordering='default'):
""" Write the parameter set to an XML file for use with OpenMM
Parameters
----------
dest : str or file-like
The name of the file or the file-like object (with a ``write``
attribute) to which the XML file will be written
provenance : dict, optional
If present, the XML file will be tagged with the available fields.
Keys of the dictionary become XML element tags, the values of the
dictionary must be instances of any of:
- str / unicode (Py2) or str (Py3) - one XML element with this
content is written
- list - one XML element per each item of the list is written, all
these XML elements use the same tag (key in provenance dict)
- dict - one of the keys of this dict must be the same as the key of
of the provenance dict under which this dict is nested. The value
under this key becomes the content of the XML element. Remaining keys
and their values are used to construct attributes of the XML element.
Note that OrderedDict's should be used to ensure appropriate order
of the XML elements and their attributes.
Default is no provenance.
Example (unordered):
provenance = {'Reference' : ['Nature', 'Cell'],
'Source' : {'Source': 'leaprc.ff14SB', sourcePackage :
'AmberTools', sourcePackageVersion : '15'},
'User' : 'Mark'}
write_unused : bool
If False: a) residue templates using unavailable atom types will not
be written, b) atom types that are not used in any of the residue
templates remaining and parameters including those atom types will
not be written. A ParameterWarning is issued if any such residues are
found in a).
separate_ljforce : bool
If True will use a separate LennardJonesForce to create a
CostumNonbondedForce to compute L-J interactions. It will set sigma
to 1 and epsilon to 0 in the NonbondedForce so that the
NonbondedForce only calculates the electrostatic contribution. It
should be set to True when converting a CHARMM force field file that
doesn't have pair-specific L-J modifications (NBFIX in CHARMM) so
that the ffxml conversion is compatible with the main charmm36.xml file.
Note:
----
When pair-specific L-J modifications are present (NBFIX in CHARMM), this
behavior is always present and this flag is ignored.
Notes
-----
The generated XML file will have the XML tag ``DateGenerated`` added to
the provenance information set to the current date. Therefore, you
should not provide this information in ``provenance`` (it will be
removed if it is provided).
"""
if isinstance(dest, string_types):
dest = genopen(dest, 'w')
own_handle = True
else:
own_handle = False
if not write_unused:
skip_residues = self._find_unused_residues()
skip_types = self._find_unused_types(skip_residues)
if skip_residues:
warnings.warn(
'Some residue templates using unavailable AtomTypes '
'were found. They will not be written to the ffxml '
'as write_unused is set to False', ParameterWarning)
else:
skip_residues = set()
skip_types = set()
if self.atom_types:
try:
self.typeify_templates()
except KeyError:
warnings.warn(
'Some residue templates are using unavailable '
'AtomTypes', ParameterWarning)
try:
dest.write('<ForceField>\n')
self._write_omm_provenance(dest, provenance)
self._write_omm_atom_types(dest, skip_types)
self._write_omm_residues(dest, skip_residues)
self._write_omm_bonds(dest, skip_types)
self._write_omm_angles(dest, skip_types)
self._write_omm_urey_bradley(dest, skip_types)
self._write_omm_dihedrals(dest, skip_types,
improper_dihedrals_ordering)
self._write_omm_impropers(dest, skip_types)
# self._write_omm_rb_torsions(dest, skip_types)
self._write_omm_cmaps(dest, skip_types)
self._write_omm_scripts(dest, skip_types)
self._write_omm_nonbonded(dest, skip_types, separate_ljforce)
self._write_omm_LennardJonesForce(dest, skip_types,
separate_ljforce)
finally:
dest.write('</ForceField>\n')
if own_handle:
dest.close()
def parse(filename, structure=False):
""" Parses a mol2 file (or mol3) file
Parameters
----------
filename : str or file-like
Name of the file to parse or file-like object to parse from
structure : bool, optional
If True, the return value is a :class:`Structure` instance. If
False, it is either a :class:`ResidueTemplate` or
:class:`ResidueTemplateContainter` instance, depending on whether
there is one or more than one residue defined in it. Default is
False
Returns
-------
molecule : :class:`Structure`, :class:`ResidueTemplate`, or
:class:`ResidueTemplateContainer`
The molecule defined by this mol2 file
Raises
------
Mol2Error
If the file format is not recognized or non-numeric values are
present where integers or floating point numbers are expected. Also
raises Mol2Error if you try to parse a mol2 file that has multiple
@<MOLECULE> entries with ``structure=True``.
"""
if isinstance(filename, string_types):
f = genopen(filename, 'r')
own_handle = True
else:
f = filename
own_handle = False
rescont = ResidueTemplateContainer()
struct = Structure()
restemp = ResidueTemplate()
mol_info = []
multires_structure = False
try:
section = None
last_residue = None
headtail = 'head'
molecule_number = 0
for line in f:
if line.startswith('#'): continue
if not line.strip() and section is None: continue
if line.startswith('@<TRIPOS>'):
section = line[9:].strip()
if section == 'MOLECULE' and (restemp.atoms or rescont):
if structure:
raise Mol2Error('Cannot convert MOL2 with multiple '
'@<MOLECULE>s to a Structure')
# Set the residue name from the MOL2 title if the
# molecule had only 1 residue and it was given a name in
# the title
if not multires_structure and mol_info[0]:
restemp.name = mol_info[0]
multires_structure = False
rescont.append(restemp)
restemp = ResidueTemplate()
struct = Structure()
last_residue = None
molecule_number += 1
mol_info = []
continue
if section is None:
raise Mol2Error('Bad mol2 file format')
if section == 'MOLECULE':
# Section formatted as follows:
# mol_name
# num_atoms [num_bonds [num_substr [num_feat [num_sets]]]]
# mol_type
# charge_type
# [status_bits]
# [mol_comment]
# TODO: Do something with the name.
if len(mol_info) == 0:
mol_info.append(line.strip())
elif len(mol_info) == 1:
mol_info.append([int(x) for x in line.split()])
elif len(mol_info) == 2:
mol_info.append(line.strip())
elif len(mol_info) == 3:
mol_info.append(line.strip())
# Ignore the rest
continue
if section == 'ATOM':
# Section formatted as follows:
# atom_id -- serial number of atom
# atom_name -- name of the atom
# x -- X-coordinate of the atom
# y -- Y-coordinate of the atom
# z -- Z-coordinate of the atom
# atom_type -- type of the atom
# subst_id -- Residue serial number
# subst_name -- Residue name
# charge -- partial atomic charge
# status_bit -- ignored
words = line.split()
id = int(words[0])
name = words[1]
x = float(words[2])
y = float(words[3])
z = float(words[4])
typ = words[5]
try:
resid = int(words[6])
except IndexError:
resid = 0
try:
resname = words[7]
except IndexError:
resname = 'UNK'
if 'NO_CHARGES' not in mol_info:
try:
charge = float(words[8])
except IndexError:
charge = 0
else:
charge = 0
if last_residue is None:
last_residue = (resid, resname)
restemp.name = resname
atom = Atom(name=name, type=typ, number=id, charge=charge)
atom.xx, atom.xy, atom.xz = x, y, z
struct.add_atom(atom, resname, resid)
if last_residue != (resid, resname):
rescont.append(restemp)
restemp = ResidueTemplate()
restemp.name = resname
last_residue = (resid, resname)
multires_structure = True
try:
restemp.add_atom(copy.copy(atom))
except ValueError:
# Allow mol2 files being parsed as a Structure to have
# duplicate atom names
if not structure:
raise
continue
if section == 'BOND':
# Section formatted as follows:
# bond_id -- serial number of bond (ignored)
# origin_atom_id -- serial number of first atom in bond
# target_atom_id -- serial number of other atom in bond
# bond_type -- string describing bond type (ignored)
# status_bits -- ignored
words = line.split()
int(words[0]) # Bond serial number... redundant and ignored
a1 = int(words[1])
a2 = int(words[2])
atom1 = struct.atoms.find_original_index(a1)
atom2 = struct.atoms.find_original_index(a2)
struct.bonds.append(Bond(atom1, atom2))
# Now add it to our residue container
# See if it's a head/tail connection
if atom1.residue is not atom2.residue:
if atom1.residue.idx == len(rescont):
res1 = restemp
elif atom1.residue.idx < len(rescont):
res1 = rescont[atom1.residue.idx]
assert atom.residue.idx <= len(rescont), 'Bad bond!'
if atom2.residue.idx == len(rescont):
res2 = restemp
elif atom2.residue.idx < len(rescont):
res2 = rescont[atom2.residue.idx]
assert atom.residue.idx <= len(rescont), 'Bad bond!'
assert res1 is not res2, 'BAD identical residues'
idx1 = atom1.idx - atom1.residue[0].idx
idx2 = atom2.idx - atom2.residue[0].idx
if atom1.residue.idx < atom2.residue.idx:
res1.tail = res1[idx1]
res2.head = res2[idx2]
else:
res1.head = res1[idx1]
res2.tail = res2[idx2]
elif not multires_structure:
if not structure:
restemp.add_bond(a1-1, a2-1)
else:
# Same residue, add the bond
offset = atom1.residue[0].idx
if atom1.residue.idx == len(rescont):
res = restemp
else:
res = rescont[atom1.residue.idx]
res.add_bond(atom1.idx-offset, atom2.idx-offset)
continue
if section == 'CRYSIN':
# Section formatted as follows:
# a -- length of first unit cell vector
# b -- length of second unit cell vector
# c -- length of third unit cell vector
# alpha -- angle b/w b and c
# beta -- angle b/w a and c
# gamma -- angle b/w a and b
# space group -- number of space group (ignored)
# space group setting -- ignored
words = line.split()
box = [float(w) for w in words[:6]]
if len(box) != 6:
raise ValueError('%d box dimensions found; needed 6' %
len(box))
struct.box = copy.copy(box)
rescont.box = copy.copy(box)
continue
if section == 'SUBSTRUCTURE':
# Section formatted as follows:
# subst_id -- residue number
# subst_name -- residue name
# root_atom -- first atom of residue
# subst_type -- ignored (usually 'RESIDUE')
# dict_type -- type of substructure (ignored)
# chain -- chain ID of residue
# sub_type -- type of the chain
# inter_bonds -- # of inter-substructure bonds
# status -- ignored
# comment -- ignored
words = line.split()
if not words: continue
id = int(words[0])
resname = words[1]
try:
chain = words[5]
except IndexError:
chain = ''
# Set the chain ID
for res in struct.residues:
if res.number == id and res.name == resname:
res.chain = chain
continue
# MOL3 sections
if section == 'HEADTAIL':
atname, residx = line.split()
residx = int(residx)
if residx in (0, 1) or residx - 1 == len(rescont):
res = restemp
elif residx - 1 < len(rescont):
res = rescont[residx-1]
else:
raise Mol2Error('Residue out of range in head/tail')
for atom in res:
if atom.name == atname:
if headtail == 'head':
res.head = atom
headtail = 'tail'
else:
res.tail = atom
headtail = 'head'
break
else:
if headtail == 'head':
headtail = 'tail'
else:
headtail = 'head'
continue
if section == 'RESIDUECONNECT':
words = line.split()
residx = int(words[0])
if residx - 1 == len(rescont):
res = restemp
elif residx - 1 < len(rescont):
res = rescont[residx-1]
else:
raise Mol2Error('Residue out of range in '
'residueconnect')
for a in words[3:]:
if a == '0': continue
for atom in res:
if atom.name == a:
res.connections.append(atom)
break
else:
raise Mol2Error('Residue connection atom %s not '
'found in residue %d' % (a, residx))
if structure:
return struct
elif len(rescont) > 0:
if not multires_structure and mol_info[0]:
restemp.name = mol_info[0]
rescont.append(restemp)
return rescont
else:
return restemp
except ValueError as e:
raise Mol2Error('String conversion trouble: %s' % e)
finally:
if own_handle: f.close()
def write(struct, dest, mol3=False, split=False):
""" Writes a mol2 file from a structure or residue template
Parameters
----------
struct : :class:`Structure` or :class:`ResidueTemplate` or
:class:`ResidueTemplateContainer`
The input structure to write the mol2 file from
dest : str or file-like obj
Name of the file to write or open file handle to write to
mol3 : bool, optional
If True and ``struct`` is a ResidueTemplate or container, write
HEAD/TAIL sections. Default is False
split : bool, optional
If True and ``struct`` is a ResidueTemplateContainer or a Structure
with multiple residues, each residue is printed in a separate
@<MOLECULE> section that appear sequentially in the output file
"""
own_handle = False
if not hasattr(dest, 'write'):
own_handle = True
dest = genopen(dest, 'w')
if split:
# Write sequentially if it is a multi-residue container or Structure
if isinstance(struct, ResidueTemplateContainer):
try:
for res in struct:
Mol2File.write(res, dest, mol3)
finally:
if own_handle: dest.close()
return
elif isinstance(struct, Structure) and len(struct.residues) > 1:
try:
for res in ResidueTemplateContainer.from_structure(struct):
Mol2File.write(res, dest, mol3)
finally:
if own_handle: dest.close()
return
try:
if isinstance(struct, ResidueTemplateContainer):
natom = sum([len(c) for c in struct])
# To find the number of bonds, we need to total number of bonds
# + the number of bonds that would be formed by "stitching"
# together residues via their head and tail
bonds = []
charges = []
bases = [1 for res in struct]
for i, res in enumerate(struct):
if i < len(struct) - 1:
bases[i+1] = bases[i] + len(res)
for i, res in enumerate(struct):
for bond in res.bonds:
bonds.append((bond.atom1.idx+bases[i],
bond.atom2.idx+bases[i]))
if i < len(struct)-1 and (res.tail is not None and
struct[i+1].head is not None):
bonds.append((res.tail.idx+bases[i],
struct[i+1].head.idx+bases[i+1]))
charges.extend([a.charge for a in res])
residues = struct
name = struct.name or struct[0].name
else:
natom = len(struct.atoms)
bonds = [(b.atom1.idx+1, b.atom2.idx+1) for b in struct.bonds]
if isinstance(struct, ResidueTemplate):
residues = [struct]
name = struct.name
else:
residues = struct.residues
name = struct.residues[0].name
charges = [a.charge for a in struct.atoms]
dest.write('@<TRIPOS>MOLECULE\n')
dest.write('%s\n' % name)
dest.write('%d %d %d 0 1\n' % (natom, len(bonds), len(residues)))
if len(residues) == 1:
dest.write('SMALL\n')
else:
for residue in residues:
if AminoAcidResidue.has(residue.name):
dest.write('PROTEIN\n')
break
if (RNAResidue.has(residue.name) or
DNAResidue.has(residue.name)):
dest.write('NUCLEIC\n')
break
else:
dest.write('BIOPOLYMER\n')
if not any(charges):
dest.write('NO_CHARGES\n')
printchg = False
else:
dest.write('USER_CHARGES\n')
printchg = True
# See if we want to print box info
if hasattr(struct, 'box') and struct.box is not None:
box = struct.box
dest.write('@<TRIPOS>CRYSIN\n')
dest.write('%10.4f %10.4f %10.4f %10.4f %10.4f %10.4f 1 1\n' %
(box[0], box[1], box[2], box[3], box[4], box[5]))
# Now do ATOM section
dest.write('@<TRIPOS>ATOM\n')
j = 1
for i, res in enumerate(residues):
for atom in res:
try:
x = atom.xx
except AttributeError:
x = 0
try:
y = atom.xy
except AttributeError:
y = 0
try:
z = atom.xz
except AttributeError:
z = 0
dest.write('%8d %-8s %10.4f %10.4f %10.4f %-8s %6d %-8s' % (
j, atom.name, x, y, z,
atom.type.strip() or atom.name, i+1, res.name))
if printchg:
dest.write(' %10.6f\n' % atom.charge)
else:
dest.write('\n')
j += 1
dest.write('@<TRIPOS>BOND\n')
for i, bond in enumerate(bonds):
dest.write('%8d %8d %8d 1\n' % (i+1, bond[0], bond[1]))
dest.write('@<TRIPOS>SUBSTRUCTURE\n')
first_atom = 0
for i, res in enumerate(residues):
if not hasattr(res, 'chain') or not res.chain:
chain = '****'
else:
chain = res.chain
intresbonds = 0
if isinstance(res, ResidueTemplate):
if i != len(residues)-1 and (res.tail is not None and
residues[i+1].head is not None):
intresbonds += 1
if i != 0 and (res.head is not None and residues[i-1].tail
is not None):
intresbonds += 1
else:
for atom in res:
for a2 in atom.bond_partners:
if a2.residue is not res:
intresbonds += 1
dest.write('%8d %-8s %8d RESIDUE %4d %-4s ROOT %6d\n' % (i+1,
res.name, first_atom+1, 0, chain[:4], intresbonds))
first_atom += len(res)
if mol3:
dest.write('@<TRIPOS>HEADTAIL\n')
for i, res in enumerate(residues):
if isinstance(res, ResidueTemplate):
if res.head is not None:
dest.write('%s %d\n' % (res.head.name, i+1))
else:
dest.write('0 0\n')
if res.tail is not None:
dest.write('%s %d\n' % (res.tail.name, i+1))
else:
dest.write('0 0\n')
else:
head = tail = None
for atom in res:
for a2 in atom.bond_partners:
if a2.residue.idx == res.idx - 1:
head = atom
if a2.residue.idx == res.idx + 1:
tail = atom
if head is not None:
dest.write('%s %d\n' % (head.name, i+1))
else:
dest.write('0 0\n')
if tail is not None:
dest.write('%s %d\n' % (tail.name, i+1))
else:
dest.write('0 0\n')
dest.write('@<TRIPOS>RESIDUECONNECT\n')
for i, res in enumerate(residues):
if isinstance(res, ResidueTemplate):
con = [res.head, res.tail, None, None, None, None]
for i, a in enumerate(res.connections):
con[i+2] = a
else:
con = [None, None, None, None, None, None]
ncon = 2
for atom in res:
for a2 in atom.bond_partners:
if a2.residue.idx == res.idx - 1:
con[0] = atom
elif a2.residue.idx == res.idx + 1:
con[1] = atom
elif a2.residue.idx != res.idx:
con[ncon] = atom
ncon += 1
dest.write('%d' % (i+1))
for a in con:
if a is not None:
dest.write(' %s' % a.name)
else:
dest.write(' 0')
dest.write('\n')
finally:
if own_handle: dest.close()
def rdparm_slow(self, fname):
"""
Parses the Amber format file. This parser is written in pure Python and
is therefore slower than the C++-optimized version
"""
current_flag = ''
fmtre = re.compile(r'%FORMAT *\((.+)\)')
version = None
if isinstance(fname, string_types):
prm = genopen(fname, 'r')
own_handle = True
elif hasattr(fname, 'read'):
prm = fname
own_handle = False
else:
raise TypeError('%s must be a file name or file-like object' %
fname)
# Open up the file and read the data into memory
for line in prm:
if line[0] == '%':
if line[0:8] == '%VERSION':
self.version = line.strip()
continue
elif line[0:5] == '%FLAG':
current_flag = line[6:].strip()
self.formats[current_flag] = ''
self.parm_data[current_flag] = []
self.parm_comments[current_flag] = []
self.flag_list.append(current_flag)
continue
elif line[0:8] == '%COMMENT':
self.parm_comments[current_flag].append(line[9:].strip())
continue
elif line[0:7] == '%FORMAT':
fmt = FortranFormat(fmtre.match(line).groups()[0])
# RESIDUE_ICODE can have a lot of blank data...
if current_flag == 'RESIDUE_ICODE':
fmt.read = fmt._read_nostrip
self.formats[current_flag] = fmt
continue
try:
self.parm_data[current_flag].extend(fmt.read(line))
except KeyError:
if version is not None:
raise
break # Skip out of the loop down to the old-format parser
# convert charges to fraction-electrons
if 'CTITLE' in self.parm_data:
CHARGE_SCALE = CHARMM_ELECTROSTATIC
else:
CHARGE_SCALE = AMBER_ELECTROSTATIC
if self.charge_flag in self.parm_data:
for i, chg in enumerate(self.parm_data[self.charge_flag]):
self.parm_data[self.charge_flag][i] = chg / CHARGE_SCALE
# If we don't have a version, then read in an old-file topology
if self.version is None:
prm.seek(0)
return self.rdparm_old(prm.readlines())
if own_handle:
prm.close()
return
def id_format(filename):
"""
Identifies the file type as either an Amber-style frcmod or parm.dat
file.
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is an Amber-style parameter file. False otherwise.
"""
with closing(genopen(filename, 'r')) as f:
f.readline()
line = f.readline()
if not line.strip(): # Must be an frcmod file
while line and not line.strip():
line = f.readline()
if not line:
return False
if line.rstrip() not in ('MASS', 'BOND', 'ANGLE', 'ANGL', 'DIHE', 'DIHED',
'DIHEDRAL', 'IMPR', 'IMPROP', 'IMPROPER', 'NONB', 'NONBON',
'NONBOND', 'NONBONDED'):
return False
if line.rstrip() in ('MASS', 'BOND', 'ANGLE', 'ANGL', 'DIHE', 'DIHED', 'DIHEDRAL',
'IMPR', 'IMPROP', 'IMPROPER', 'NONB', 'NONBON', 'NONBOND',
'NONBONDED'):
return True # frcmod file
# This must be an atom definition in the parm.dat file
words = line.split()
if len(words) < 2:
return False
# The first word is the atom type (must be <= 2 characters, and not
# an integer)
if len(words[0]) > 2: return False
try:
float(words[0])
return False
except ValueError:
pass
try:
float(words[1])
except ValueError:
return False
# Polarizability might not be present...
try:
float(words[2])
except (IndexError, ValueError):
# UGLY: Check the mass and make sure it matches the element's
# mass to within 1 amu. Do our best to guess the element. If
# it's a two-letter element, the element might be a 2-letter
# element (like Br), or it might be a 1-letter element specified
# by either the first or second letter. Check all possibilities.
# Special-case instances like CA, which are just as likely (more
# likely?) to be a carbon atom as it is to be a calcium (i.e.,
# check for carbon atoms in anything that starts with C). If at
# any point, what *should* be the mass doesn't match the mass of
# the guessed element, tag this as *not* a parameter file
if len(words[0]) == 2:
if words[0][0].isalpha():
if words[0][1].isalpha():
key = words[0][0].upper() + words[0][1].lower()
if key in Mass and abs(Mass[key] - float(words[1])) > 1:
if key[0] == 'C' and abs(Mass['C'] - float(words[1])) > 1:
return False
elif key[0] != 'C':
return False
elif key not in Mass:
if key[0] in Mass and abs(Mass[key[0]] - float(words[1])) > 1:
return False
else:
return False
else:
key = words[0][0].upper()
if key in Mass and abs(Mass[key] - float(words[1])) > 1:
return False
elif key not in Mass:
return False
else:
key = words[0][1].upper()
if key in Mass:
if abs(Mass[key] - float(words[1])) > 1:
return False
else:
return False
else:
key = words[0][0].upper()
if key in Mass:
if abs(Mass[key] - float(words[1])) > 1:
return False
else:
return False
if len(words) > 3:
# Heuristic, but anything that comes after the polarizability is
# a comment, and I have yet to see a leading comment that is a
# number
try:
float(words[3])
return False
except ValueError:
return True
else:
return True
def write(self, dest, title='Created by ParmEd', style='frcmod'):
""" Writes a parm.dat file with the current parameters
Parameters
----------
dest : str or file-like
The file name or file-like object to write the parameters to
title : str, optional
The title of the frcmod to write. Default is 'Created by ParmEd'
style : str, optional
If 'frcmod', the parameters are written in frcmod-format. If 'parm',
the parameters are written in parm.dat-format. Default is 'frcmod'
"""
if isinstance(dest, string_types):
outfile = genopen(dest, 'w')
own_handle = True
else:
outfile = dest
own_handle = False
if style not in ('frcmod', 'parm'):
raise ValueError('style must be either frcmod or parm, not %s' % style)
outfile.write(title.rstrip('\r\n'))
outfile.write('\n')
# Write the atom mass
outfile.write('MASS\n')
for atom, typ in iteritems(self.atom_types):
outfile.write('%s%6.3f\n' % (atom.ljust(6), typ.mass))
outfile.write('\n')
# Write the bonds
outfile.write('BOND\n')
done = set()
for (a1, a2), typ in iteritems(self.bond_types):
if id(typ) in done: continue
done.add(id(typ))
outfile.write('%s-%s %8.3f %6.3f\n' % (a1.ljust(2), a2.ljust(2), typ.k, typ.req))
outfile.write('\n')
# Write the angles
outfile.write('ANGLE\n')
done = set()
for (a1, a2, a3), typ in iteritems(self.angle_types):
if id(typ) in done: continue
done.add(id(typ))
outfile.write('%s-%s-%s %8.3f %6.3f\n' % (a1.ljust(2), a2.ljust(2), a3.ljust(2),
typ.k, typ.theteq))
outfile.write('\n')
# Write the dihedrals
outfile.write('DIHE\n')
done = set()
for (a1, a2, a3, a4), typ in iteritems(self.dihedral_types):
if id(typ) in done: continue
done.add(id(typ))
if isinstance(typ, DihedralType) or len(typ) == 1:
if not isinstance(typ, DihedralType):
typ = typ[0]
outfile.write('%s-%s-%s-%s %4i %14.8f %8.3f %5.1f SCEE=%s SCNB=%s\n' %
(a1.ljust(2), a2.ljust(2), a3.ljust(2), a4.ljust(2), 1, typ.phi_k,
typ.phase, typ.per, typ.scee, typ.scnb))
else:
for dtyp in typ[:-1]:
outfile.write('%s-%s-%s-%s %4i %14.8f %8.3f %5.1f SCEE=%s SCNB=%s\n' %
(a1.ljust(2), a2.ljust(2), a3.ljust(2), a4.ljust(2), 1,
dtyp.phi_k, dtyp.phase, -dtyp.per, dtyp.scee, dtyp.scnb))
dtyp = typ[-1]
outfile.write('%s-%s-%s-%s %4i %14.8f %8.3f %5.1f SCEE=%s SCNB=%s\n' %
(a1.ljust(2), a2.ljust(2), a3.ljust(2), a4.ljust(2), 1, dtyp.phi_k,
dtyp.phase, dtyp.per, dtyp.scee, dtyp.scnb))
outfile.write('\n')
# Write the impropers
outfile.write('IMPROPER\n')
written_impropers = dict()
for (a1, a2, a3, a4), typ in iteritems(self.improper_periodic_types):
# Make sure wild-cards come at the beginning
if a2 == 'X':
assert a4 == 'X', 'Malformed generic improper!'
a1, a2, a3, a4 = a2, a4, a3, a1
elif a4 == 'X':
a1, a2, a3, a4 = a4, a1, a3, a2
a1, a2, a4 = sorted([a1, a2, a4])
if (a1, a2, a3, a4) in written_impropers:
if written_impropers[(a1, a2, a3, a4)] != typ:
raise ValueError('Multiple impropers with the same atom set not allowed')
continue
outfile.write('%s-%s-%s-%s %14.8f %8.3f %5.1f\n' %
(a1.ljust(2), a2.ljust(2), a3.ljust(2), a4.ljust(2), typ.phi_k,
typ.phase, typ.per))
written_impropers[(a1, a2, a3, a4)] = typ
outfile.write('\n')
# Write the LJ terms
outfile.write('NONB\n')
for atom, typ in iteritems(self.atom_types):
outfile.write('%s %12.8f %12.8f\n' % (atom.ljust(2), typ.rmin, typ.epsilon))
outfile.write('\n')
# Write the NBFIX terms
if self.nbfix_types:
outfile.write('LJEDIT\n')
for (a1, a2), (eps, rmin) in iteritems(self.nbfix_types):
outfile.write('%s %s %12.8f %12.8f %12.8f %12.8f\n' %
(a1.ljust(2), a2.ljust(2), eps, rmin/2, eps, rmin/2))
if own_handle:
outfile.close()
def from_leaprc(cls, fname, search_oldff=False):
""" Load a parameter set from a leaprc file
Parameters
----------
fname : str or file-like
Name of the file or open file-object from which a leaprc-style file
will be read
search_oldff : bool, optional, default=False
If True, search the oldff directories in the main Amber leap
folders. Default is False
Notes
-----
This does not read all parts of a leaprc file -- only those pertinent to
defining force field information. For instance, the following sections
and commands are processed:
- addAtomTypes
- loadAmberParams
- loadOFF
- loadMol2
- loadMol3
"""
params = cls()
if isinstance(fname, string_types):
f = genopen(fname, 'r')
own_handle = True
else:
f = fname
own_handle = False
# To make parsing easier, and because leaprc files are usually quite
# short, I'll read the whole file into memory
def joinlines(lines):
newlines = []
composite = []
for line in lines:
if line.endswith('\\\n'):
composite.append(line[:-2])
continue
else:
composite.append(line)
newlines.append(''.join(composite))
composite = []
if composite:
newlines.append(''.join(composite))
return newlines
lines = joinlines(map(lambda line: line if '#' not in line else line[:line.index('#')], f))
text = ''.join(lines)
if own_handle: f.close()
lowertext = text.lower() # commands are case-insensitive
# Now process the parameter files
def process_fname(fname):
if fname[0] in ('"', "'"):
fname = fname[1:-1]
fname = fname.replace('_BSTOKEN_', r'\ ').replace(r'\ ', ' ')
return fname
for line in lines:
line = line.replace(r'\ ', '_BSTOKEN_')
if _loadparamsre.findall(line):
fname = process_fname(_loadparamsre.findall(line)[0])
params.load_parameters(_find_amber_file(fname, search_oldff))
elif _loadoffre.findall(line):
fname = process_fname(_loadoffre.findall(line)[0])
params.residues.update(AmberOFFLibrary.parse(_find_amber_file(fname, search_oldff)))
elif _loadmol2re.findall(line):
(resname, fname), = _loadmol2re.findall(line)
residue = Mol2File.parse(_find_amber_file(fname, search_oldff))
if isinstance(residue, ResidueTemplateContainer):
warnings.warn('Multi-residue mol2 files not supported by tleap. Loading anyway '
'using names in mol2', AmberWarning)
for res in residue:
params.residues[res.name] = res
else:
params.residues[resname] = residue
# Now process the addAtomTypes
try:
idx = lowertext.index('addatomtypes')
except ValueError:
# Does not exist in this file
atom_types_str = ''
else:
i = idx + len('addatomtypes')
while i < len(text) and text[i] != '{':
if text[i] not in '\r\n\t ':
raise ParameterError('Unsupported addAtomTypes syntax in leaprc file')
i += 1
if i == len(text):
raise ParameterError('Unsupported addAtomTypes syntax in leaprc file')
# We are at our first brace
chars = []
nopen = 1
i += 1
while i < len(text):
char = text[i]
if char == '{':
nopen += 1
elif char == '}':
nopen -= 1
if nopen == 0: break
elif char == '\n':
char = ' '
chars.append(char)
i += 1
atom_types_str = ''.join(chars).strip()
for _, name, symb, hyb in _atomtypere.findall(atom_types_str):
if symb not in AtomicNum:
raise ParameterError('%s is not a recognized element' % symb)
if name in params.atom_types:
params.atom_types[name].atomic_number = AtomicNum[symb]
return params
def parse(filename):
""" Read a PQR file and return a populated `Structure` class
Parameters
----------
filename : str or file-like
Name of the PQR file to read, or a file-like object that can iterate
over the lines of a PQR. Compressed file names can be specified and
are determined by file-name extension (e.g., file.pqr.gz,
file.pqr.bz2)
Returns
-------
structure : :class:`Structure`
The Structure object initialized with all of the information from
the PDB file. No bonds or other topological features are added by
default.
"""
if isinstance(filename, string_types):
own_handle = True
fileobj = genopen(filename, 'r')
else:
own_handle = False
fileobj = filename
struct = Structure()
# Add metadata fields
modelno = 1 # For PDB files with multiple MODELs
atomno = 0
coordinates = []
all_coordinates = []
# Support hexadecimal numbering like that printed by VMD
try:
for line in fileobj:
words = line.split()
if words[0] in ('ATOM', 'HETATM'):
atomno += 1
if len(words) == 10:
_, num, nam, res, resn, x, y, z, chg, rad = words
chn = ''
elif len(words) >= 11:
_, num, nam, res, chn, resn, x, y, z, chg, rad = (
words[i] for i in range(11))
# If the radius is not a float (but rather a letter,
# like the element or something), then the chain might
# be missing. In this case, shift all tokens "back" one
# and empty the chn string
try:
float(rad)
except ValueError:
resn, x, y, z, chg, rad = chn, resn, x, y, z, chg
else:
raise ValueError('Illegal PQR record format: expected '
'10 or 11 tokens on the atom line')
x, y, z = float(x), float(y), float(z)
chg, rad = float(chg), float(rad)
resn, num = int(resn), int(num)
elem = element_by_name(nam) # Yuck
atomic_number = AtomicNum[elem]
mass = Mass[elem]
if nam in ('EP', 'LP'): # lone pair
atom = ExtraPoint(atomic_number=atomic_number,
name=nam,
charge=chg,
mass=mass,
number=num,
solvent_radius=rad)
else:
atom = Atom(atomic_number=atomic_number,
name=nam,
charge=chg,
mass=mass,
number=num,
solvent_radius=rad)
atom.xx, atom.xy, atom.xz = float(x), float(y), float(z)
if modelno == 1:
struct.add_atom(atom, res, resn, chn)
else:
try:
orig_atom = struct.atoms[atomno - 1]
except IndexError:
raise PDBError('Extra atom in MODEL %d' % modelno)
if (orig_atom.residue.name != res.strip()
or orig_atom.name != nam.strip()):
raise PDBError(
'Atom %d differs in MODEL %d [%s %s '
'vs. %s %s]' %
(atomno, modelno, orig_atom.residue.name,
orig_atom.name, res, nam))
coordinates.extend([atom.xx, atom.xy, atom.xz])
elif words[0] == 'ENDMDL':
# End the current model
if len(struct.atoms) == 0:
raise PDBError('MODEL ended before any atoms read in')
modelno += 1
if len(struct.atoms) * 3 != len(coordinates):
raise PDBError(
'Inconsistent atom numbers in some PDB models')
all_coordinates.append(coordinates)
atomno = 0
coordinates = []
elif words[0] == 'MODEL':
if modelno == 1 and len(struct.atoms) == 0: continue
if len(coordinates) > 0:
if len(struct.atoms) * 3 != len(coordinates):
raise PDBError('Inconsistent atom numbers in '
'some PDB models')
warnings.warn('MODEL not explicitly ended', PDBWarning)
all_coordinates.append(coordinates)
coordinates = []
modelno += 1
atomno = 0
elif words[0] == 'CRYST1':
a, b, c = (float(w) for w in words[1:4])
try:
A, B, C = (float(w) for w in words[4:7])
except ValueError:
A = B = C = 90.0
struct.box = [a, b, c, A, B, C]
finally:
if own_handle: fileobj.close()
struct.unchange()
if coordinates:
if len(coordinates) != 3 * len(struct.atoms):
raise PDBError('bad number of atoms in some PQR models')
all_coordinates.append(coordinates)
struct._coordinates = np.array(all_coordinates).reshape(
(-1, len(struct.atoms), 3))
return struct
def write(struct, dest, precision=3, nobox=False):
""" Write a Gromacs Topology File from a Structure
Parameters
----------
struct : :class:`Structure`
The structure to write to a Gromacs GRO file (must have coordinates)
dest : str or file-like
The name of a file or a file object to write the Gromacs topology to
precision : int, optional
The number of decimal places to print in the coordinates. Default 3
nobox : bool, optional
If the system does not have a periodic box defined, and this option
is True, no box will be written. If False, the periodic box will be
defined to enclose the solute with 0.5 nm clearance on all sides. If
periodic box dimensions *are* defined, this variable has no effect.
"""
own_handle = False
if isinstance(dest, string_types):
dest = genopen(dest, 'w')
own_handle = True
elif not hasattr(dest, 'write'):
raise TypeError('dest must be a file name or file-like object')
dest.write('GROningen MAchine for Chemical Simulation\n')
dest.write('%5d\n' % len(struct.atoms))
has_vels = all(hasattr(a, 'vx') for a in struct.atoms)
varwidth = 5 + precision
crdfmt = '%%%d.%df' % (varwidth, precision)
velfmt = '%%%d.%df' % (varwidth, precision+1)
boxfmt = '%%%d.%df ' % (max(varwidth, 10), max(precision, 5))
for atom in struct.atoms:
resid = (atom.residue.idx + 1) % 100000
atid = (atom.idx + 1) % 100000
dest.write('%5d%-5s%5s%5d' % (resid, atom.residue.name[:5],
atom.name[:5], atid))
dest.write((crdfmt % (0.1*atom.xx))[:varwidth])
dest.write((crdfmt % (0.1*atom.xy))[:varwidth])
dest.write((crdfmt % (0.1*atom.xz))[:varwidth])
if has_vels:
dest.write((velfmt % (0.1*atom.vx))[:varwidth])
dest.write((velfmt % (0.1*atom.vy))[:varwidth])
dest.write((velfmt % (0.1*atom.vz))[:varwidth])
dest.write('\n')
# Box, in the weird format...
if struct.box is not None:
a, b, c = reduce_box_vectors(*box_lengths_and_angles_to_vectors(
*struct.box))
if all([abs(x-90) < TINY for x in struct.box[3:]]):
dest.write(boxfmt*3 % (0.1*a[0], 0.1*b[1], 0.1*c[2]))
else:
dest.write(boxfmt*9 % (0.1*a[0], 0.1*b[1], 0.1*c[2], 0.1*a[1],
0.1*a[2], 0.1*b[0], 0.1*b[2], 0.1*c[0], 0.1*c[1]))
dest.write('\n')
elif not nobox and struct.atoms:
# Find the extent of the molecule in all dimensions, and buffer it
# by 5 A
crds = struct.coordinates
diff = 0.1*(crds.max(axis=1) - crds.min(axis=1)) + 0.5
dest.write(boxfmt*3 % (diff[0], diff[1], diff[2]))
dest.write('\n')
if own_handle:
dest.close()
def id_format(filename):
""" Identifies the file type as a PDB file
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is a PQR file
"""
with closing(genopen(filename, 'r')) as f:
for line in f:
words = line.split()
if not words:
continue
elif words[0] in ('CRYST1', 'END', 'END', 'HEADER', 'NUMMDL',
'MASTER', 'AUTHOR', 'CAVEAT', 'COMPND',
'EXPDTA', 'MDLTYP', 'KEYWDS', 'OBSLTE',
'SOURCE', 'SPLIT', 'SPRSDE', 'TITLE ',
'ANISOU', 'CISPEP', 'CONECT', 'DBREF ',
'HELIX ', 'HET', 'LINK', 'MODRES', 'REVDAT',
'SEQADV', 'SHEET', 'SSBOND', 'FORMUL',
'HETNAM', 'HETSYN', 'SEQRES', 'SITE',
'ENDMDL', 'MODEL', 'JRNL', 'REMARK', 'TER',
'USER'):
continue
elif line[:5] in ('ORIGX', 'SCALE', 'MTRIX'):
if line[5] not in '123':
return False
elif words[0] in ('ATOM', 'HETATM'):
# Format is:
# rec atnum atname resname [chain] resnum x y z chg radius
# Where the chain ID is optional. rec must be ATOM or HETATM
if len(words) < 10:
return False
elif PDBFile.id_format(filename):
return False # It is a PDB file
if len(words) == 10:
offset = 0
elif len(words) >= 11:
offset = 1
try:
float(words[10])
except ValueError:
offset = 0
if not words[1].isdigit(): return False
if words[2].isdigit(): return False
if words[3].isdigit(): return False
if not words[4 + offset].isdigit(): return False
try:
float(words[5 + offset])
float(words[6 + offset])
float(words[7 + offset])
float(words[8 + offset])
float(words[9 + offset])
except ValueError:
return False
return True
else:
return False
return False
def test_read_bad_URL(self):
""" Tests proper exception handling of non-existent URL """
self.assertRaises(IOError, lambda: genopen('http://asdkfjasdf.lib'))
def write(self, dest, provenance=None, write_unused=True, separate_ljforce=False,
improper_dihedrals_ordering='default', charmm_imp=False):
""" Write the parameter set to an XML file for use with OpenMM
Parameters
----------
dest : str or file-like
The name of the file or the file-like object (with a ``write``
attribute) to which the XML file will be written
provenance : dict, optional
If present, the XML file will be tagged with the available fields.
Keys of the dictionary become XML etree.Element tags, the values of the
dictionary must be instances of any of:
- str / unicode (Py2) or str (Py3) - one XML element with this
content is written
- list - one XML element per each item of the list is written, all
these XML elements use the same tag (key in provenance dict)
- dict - one of the keys of this dict must be the same as the key of
of the provenance dict under which this dict is nested. The value
under this key becomes the content of the XML element. Remaining keys
and their values are used to construct attributes of the XML element.
Note that OrderedDict's should be used to ensure appropriate order
of the XML elements and their attributes.
Default is no provenance.
Example (unordered):
provenance = {'Reference' : ['Nature', 'Cell'],
'Source' : {'Source': 'leaprc.ff14SB', sourcePackage :
'AmberTools', sourcePackageVersion : '15'},
'User' : 'Mark'}
write_unused : bool
If False: a) residue templates using unavailable atom types will not
be written, b) atom types that are not used in any of the residue
templates remaining and parameters including those atom types will
not be written. A ParameterWarning is issued if any such residues are
found in a).
separate_ljforce : bool
If True will use a separate LennardJonesForce to create a
CostumNonbondedForce to compute L-J interactions. It will set sigma
to 1 and epsilon to 0 in the NonbondedForce so that the
NonbondedForce only calculates the electrostatic contribution. It
should be set to True when converting a CHARMM force field file that
doesn't have pair-specific L-J modifications (NBFIX in CHARMM) so
that the ffxml conversion is compatible with the main charmm36.xml file.
Note:
----
When pair-specific L-J modifications are present (NBFIX in CHARMM), this
behavior is always present and this flag is ignored.
improper_dihedrals_ordering : str
The ordering to use when assigning improper torsions in OpenMM. Default is 'default',
other option is 'amber'
charmm_imp: bool
If True, will check for existence of IMPR in each residue and patch template,
and write out the explicit improper definition without wildcards in the ffxml file.
Notes
-----
The generated XML file will have the XML tag ``DateGenerated`` added to
the provenance information set to the current date. Therefore, you
should not provide this information in ``provenance`` (it will be
removed if it is provided).
"""
if not write_unused:
skip_residues = self._find_unused_residues()
skip_types = self._find_unused_types(skip_residues)
if skip_residues:
warnings.warn('Some residue templates using unavailable AtomTypes '
'were found. They will not be written to the ffxml '
'as write_unused is set to False', ParameterWarning)
else:
skip_residues = set()
skip_types = set()
if self.atom_types:
try:
self.typeify_templates()
except KeyError:
warnings.warn('Some residue templates are using unavailable '
'AtomTypes', ParameterWarning)
[valid_residues_for_patch, valid_patches_for_residue] = self._determine_valid_patch_combinations(skip_residues)
LOGGER.debug('Valid patch combinations:')
for patch_name in self.patches:
LOGGER.debug('%8s : %s', patch_name, valid_residues_for_patch[patch_name])
if charmm_imp:
self._find_explicit_impropers()
self._compress_impropers()
root = etree.Element('ForceField')
self._write_omm_provenance(root, provenance)
self._write_omm_atom_types(root, skip_types)
self._write_omm_residues(root, skip_residues, valid_patches_for_residue)
self._write_omm_patches(root, valid_residues_for_patch)
self._write_omm_bonds(root, skip_types)
self._write_omm_angles(root, skip_types)
self._write_omm_urey_bradley(root, skip_types)
self._write_omm_dihedrals(root, skip_types, improper_dihedrals_ordering)
self._write_omm_impropers(root, skip_types)
#self._write_omm_rb_torsions(root, skip_types)
self._write_omm_cmaps(root, skip_types)
self._write_omm_scripts(root, skip_types)
self._write_omm_nonbonded(root, skip_types, separate_ljforce)
self._write_omm_LennardJonesForce(root, skip_types, separate_ljforce)
tree = etree.ElementTree(root)
xml = etree.tostring(tree, encoding=DEFAULT_ENCODING, pretty_print=True).decode('utf-8')
if isinstance(dest, string_types):
with closing(genopen(dest, 'w')) as f:
f.write(xml)
else:
dest.write(xml)
def test_write_normal(self):
""" Tests genopen writing a normal text file """
with closing(genopen(get_fn('tmp.txt', written=True), 'w')) as f:
f.write(ALPHABET)
self.assertEqual(
open(get_fn('tmp.txt', written=True), 'r').read(), ALPHABET)
def write(struct,
dest,
renumber=True,
coordinates=None,
standard_resnames=False):
""" Write a PDB file from a Structure instance
Parameters
----------
struct : :class:`Structure`
The structure from which to write the PDB file
dest : str or file-like
Either a file name or a file-like object containing a `write`
method to which to write the PDB file. If it is a filename that
ends with .gz or .bz2, a compressed version will be written using
either gzip or bzip2, respectively.
renumber : bool, optional
If True, renumber the atoms and residues sequentially as they are
stored in the structure. If False, use the original numbering if
it was assigned previously. Default is True
coordinates : array-like of float, optional
If provided, these coordinates will be written to the PDB file
instead of the coordinates stored in the structure. These
coordinates should line up with the atom order in the structure
(not necessarily the order of the "original" PDB file if they
differ)
standard_resnames : bool, optional
If True, common aliases for various amino and nucleic acid residues
will be converted into the PDB-standard values. Default is False
"""
own_handle = False
if not hasattr(dest, 'write'):
dest = genopen(dest, 'w')
own_handle = True
atomrec = ('ATOM %5d %-3s %-3s %1s %3d %7.3f %7.3f %7.3f %8.4f '
'%8.4f\n')
if struct.box is not None:
dest.write('CRYST1 %8.3f %8.3f %8.3f %6.2f %6.2f %6.2f\n' %
(struct.box[0], struct.box[1], struct.box[2],
struct.box[3], struct.box[4], struct.box[5]))
if coordinates is not None:
coords = np.array(coordinates, copy=False, subok=True)
try:
coords = coords.reshape((-1, len(struct.atoms), 3))
except ValueError:
raise TypeError("Coordinates has unexpected shape")
else:
coords = struct.get_coordinates('all')
# Create a function to process each atom and return which one we want
# to print, based on our alternate location choice
if standard_resnames:
standardize = lambda x: _standardize_resname(x)
else:
standardize = lambda x: x
last_number = 0
last_rnumber = 0
for model, coord in enumerate(coords):
if coords.shape[0] > 1:
dest.write('MODEL %5d\n' % (model + 1))
for res in struct.residues:
if renumber:
atoms = res.atoms
else:
atoms = sorted(res.atoms, key=lambda atom: atom.number)
for atom in atoms:
# Figure out the serial numbers we want to print
if renumber:
anum = (atom.idx + 1)
rnum = (res.idx + 1)
else:
anum = (atom.number or last_number + 1)
rnum = (atom.residue.number or last_rnumber + 1)
last_number = anum
last_rnumber = rnum
# Do any necessary name munging to respect the PDB spec
if (len(atom.name) < 4
and len(Element[atom.atomic_number]) != 2):
aname = ' %-3s' % atom.name
else:
aname = atom.name
xyz = coord[atom.idx]
dest.write(atomrec % (anum, aname, standardize(
res.name), res.chain, rnum, xyz[0], xyz[1], xyz[2],
atom.charge, atom.solvent_radius))
if coords.shape[0] > 1:
dest.write('ENDMDL\n')
dest.write("%-80s\n" % "END")
if own_handle:
dest.close()
def test_write_gzipped(self):
""" Tests genopen writing a gzipped file """
with closing(genopen(get_fn('test.gz', written=True), 'w')) as f:
f.write(ALPHABET)
text = gzip.open(get_fn('test.gz', written=True), 'r').read()
self.assertEqual(text.decode('ascii'), ALPHABET)
def __init__(self, fname, seq=None):
super(XyzFile, self).__init__()
if isinstance(fname, string_types):
fxyz = genopen(fname, 'r')
own_handle_xyz = True
else:
fxyz = fname
own_handle_xyz = False
if seq is not None:
seqstruct = load_file(seq)
# Now parse the file
try:
natom = int(fxyz.readline().split()[0])
except (ValueError, IndexError):
raise TinkerError('Bad XYZ file format; first line')
if seq is not None and natom != len(seqstruct.atoms):
raise ValueError(
'Sequence file %s # of atoms does not match the # '
'of atoms in the XYZ file' % seq)
words = fxyz.readline().split()
if len(words) == 6 and not XyzFile._check_atom_record(words):
self.box = [float(w) for w in words]
words = fxyz.readline().split()
residue = Residue('SYS')
residue.number = 1
residue._idx = 0
if seq is not None:
residue = seqstruct.residues[0]
atomic_number = _guess_atomic_number(words[1], residue)
else:
atomic_number = AtomicNum[element_by_name(words[1])]
atom = Atom(atomic_number=atomic_number, name=words[1], type=words[5])
atom.xx, atom.xy, atom.xz = [float(w) for w in words[2:5]]
self.add_atom(atom, residue.name, residue.number, residue.chain,
residue.insertion_code, residue.segid)
bond_ids = [[int(w) for w in words[6:]]]
for i, line in enumerate(fxyz):
words = line.split()
if seq is not None:
residue = seqstruct.atoms[i + 1].residue
atomic_number = _guess_atomic_number(words[1], residue)
else:
atomic_number = AtomicNum[element_by_name(words[1])]
atom = Atom(atomic_number=atomic_number,
name=words[1],
type=words[5])
atom.xx, atom.xy, atom.xz = [float(w) for w in words[2:5]]
self.add_atom(atom, residue.name, residue.number, residue.chain,
residue.insertion_code, residue.segid)
bond_ids.append([int(w) for w in words[6:]])
# All of the bonds are stored now -- go ahead and make them now
for atom, bonds in zip(self.atoms, bond_ids):
i = atom.idx + 1
for idx in bonds:
if idx > i:
self.bonds.append(Bond(atom, self.atoms[idx - 1]))
if seq is None:
# Try to improve atomic number prediction for monoatomic species
# (like ions) if no sequence as loaded
for atom in self.atoms:
if len(atom.bonds) == 0: # not bonded to anybody else
atom.atomic_number = _guess_atomic_number(atom.name)
if own_handle_xyz:
fxyz.close()
def write(self, dest, provenance=None, write_unused=True):
""" Write the parameter set to an XML file for use with OpenMM
Parameters
----------
dest : str or file-like
The name of the file or the file-like object (with a ``write``
attribute) to which the XML file will be written
provenance : dict, optional
If present, the XML file will be tagged with the available fields.
Keys of the dictionary become XML element tags, the values of the
dictionary must be instances of any of:
- str / unicode (Py2) or str (Py3) - one XML element with this
content is written
- list - one XML element per each item of the list is written, all
these XML elements use the same tag (key in provenance dict)
- dict - one of the keys of this dict must be the same as the key of
of the provenance dict under which this dict is nested. The value
under this key becomes the content of the XML element. Remaining keys
and their values are used to construct attributes of the XML element.
Note that OrderedDict's should be used to ensure appropriate order
of the XML elements and their attributes.
Default is no provenance.
Example (unordered):
provenance = {'Reference' : ['Nature', 'Cell'],
'Source' : {'Source': 'leaprc.ff14SB', sourcePackage :
'AmberTools', sourcePackageVersion : '15'},
'User' : 'Mark'}
write_unused : bool
If False, atom types that are not used in any of the residue templates
and parameters including those atom types will not be written
Notes
-----
The generated XML file will have the XML tag ``DateGenerated`` added to
the provenance information set to the current date. Therefore, you
should not provide this information in ``provenance`` (it will be
removed if it is provided).
"""
if isinstance(dest, string_types):
dest = genopen(dest, 'w')
own_handle = True
else:
own_handle = False
typeified = False
if self.atom_types:
try:
self.typeify_templates()
typeified = True
except KeyError:
warnings.warn(
'Some residue templates are using unavailable '
'AtomTypes', ParameterWarning)
if not write_unused:
if not typeified:
warnings.warn(
'Typification of the templates was not successful. '
'Proceeding with write_unused=False is not advised',
ParameterWarning)
skip_types = self._find_unused_types()
else:
skip_types = set()
try:
dest.write('<ForceField>\n')
self._write_omm_provenance(dest, provenance)
self._write_omm_atom_types(dest, skip_types)
self._write_omm_residues(dest)
self._write_omm_bonds(dest, skip_types)
self._write_omm_angles(dest, skip_types)
self._write_omm_urey_bradley(dest, skip_types)
self._write_omm_dihedrals(dest, skip_types)
self._write_omm_impropers(dest, skip_types)
# self._write_omm_rb_torsions(dest, skip_types)
self._write_omm_cmaps(dest, skip_types)
self._write_omm_scripts(dest, skip_types)
self._write_omm_nonbonded(dest, skip_types)
finally:
dest.write('</ForceField>\n')
if own_handle:
dest.close()
def __init__(self, psf_name=None):
"""
Opens and parses a PSF file, then instantiates a CharmmPsfFile
instance from the data.
"""
global _resre
Structure.__init__(self)
# Bail out if we don't have a filename
if psf_name is None:
return
conv = CharmmPsfFile._convert
# Open the PSF and read the first line. It must start with "PSF"
with closing(genopen(psf_name, 'r')) as psf:
self.name = psf_name
line = psf.readline()
if not line.startswith('PSF'):
raise CharmmError('Unrecognized PSF file. First line is %s' %
line.strip())
# Store the flags
psf_flags = line.split()[1:]
# Now get all of the sections and store them in a dict
psf.readline()
# Now get all of the sections
psfsections = _ZeroDict()
while True:
try:
sec, ptr, data = CharmmPsfFile._parse_psf_section(psf)
except _FileEOF:
break
psfsections[sec] = (ptr, data)
# store the title
self.title = psfsections['NTITLE'][1]
# Next is the number of atoms
natom = conv(psfsections['NATOM'][0], int, 'natom')
# Parse all of the atoms
for i in range(natom):
words = psfsections['NATOM'][1][i].split()
atid = int(words[0])
if atid != i + 1:
raise CharmmError('Nonsequential atoms detected!')
segid = words[1]
rematch = _resre.match(words[2])
if not rematch:
raise CharmmError('Could not interpret residue number %s'
% # pragma: no cover
words[2])
resid, inscode = rematch.groups()
resid = conv(resid, int, 'residue number')
resname = words[3]
name = words[4]
attype = words[5]
# Try to convert the atom type to an integer a la CHARMM
try:
attype = int(attype)
except ValueError:
pass
charge = conv(words[6], float, 'partial charge')
mass = conv(words[7], float, 'atomic mass')
props = words[8:]
atom = Atom(name=name, type=attype, charge=charge, mass=mass)
atom.props = props
self.add_atom(atom,
resname,
resid,
chain=segid,
inscode=inscode,
segid=segid)
# Now get the number of bonds
nbond = conv(psfsections['NBOND'][0], int, 'number of bonds')
if len(psfsections['NBOND'][1]) != nbond * 2:
raise CharmmError(
'Got %d indexes for %d bonds' % # pragma: no cover
(len(psfsections['NBOND'][1]), nbond))
it = iter(psfsections['NBOND'][1])
for i, j in zip(it, it):
self.bonds.append(Bond(self.atoms[i - 1], self.atoms[j - 1]))
# Now get the number of angles and the angle list
ntheta = conv(psfsections['NTHETA'][0], int, 'number of angles')
if len(psfsections['NTHETA'][1]) != ntheta * 3:
raise CharmmError(
'Got %d indexes for %d angles' % # pragma: no cover
(len(psfsections['NTHETA'][1]), ntheta))
it = iter(psfsections['NTHETA'][1])
for i, j, k in zip(it, it, it):
self.angles.append(
Angle(self.atoms[i - 1], self.atoms[j - 1],
self.atoms[k - 1]))
self.angles[-1].funct = 5 # urey-bradley
# Now get the number of torsions and the torsion list
nphi = conv(psfsections['NPHI'][0], int, 'number of torsions')
if len(psfsections['NPHI'][1]) != nphi * 4:
raise CharmmError(
'Got %d indexes for %d torsions' % # pragma: no cover
(len(psfsections['NPHI']), nphi))
it = iter(psfsections['NPHI'][1])
for i, j, k, l in zip(it, it, it, it):
self.dihedrals.append(
Dihedral(self.atoms[i - 1], self.atoms[j - 1],
self.atoms[k - 1], self.atoms[l - 1]))
self.dihedrals.split = False
# Now get the number of improper torsions
nimphi = conv(psfsections['NIMPHI'][0], int, 'number of impropers')
if len(psfsections['NIMPHI'][1]) != nimphi * 4:
raise CharmmError(
'Got %d indexes for %d impropers' % # pragma: no cover
(len(psfsections['NIMPHI'][1]), nimphi))
it = iter(psfsections['NIMPHI'][1])
for i, j, k, l in zip(it, it, it, it):
self.impropers.append(
Improper(self.atoms[i - 1], self.atoms[j - 1],
self.atoms[k - 1], self.atoms[l - 1]))
# Now handle the donors (what is this used for??)
ndon = conv(psfsections['NDON'][0], int, 'number of donors')
if len(psfsections['NDON'][1]) != ndon * 2:
raise CharmmError(
'Got %d indexes for %d donors' % # pragma: no cover
(len(psfsections['NDON'][1]), ndon))
it = iter(psfsections['NDON'][1])
for i, j in zip(it, it):
self.donors.append(
AcceptorDonor(self.atoms[i - 1], self.atoms[j - 1]))
# Now handle the acceptors (what is this used for??)
nacc = conv(psfsections['NACC'][0], int, 'number of acceptors')
if len(psfsections['NACC'][1]) != nacc * 2:
raise CharmmError(
'Got %d indexes for %d acceptors' % # pragma: no cover
(len(psfsections['NACC'][1]), nacc))
it = iter(psfsections['NACC'][1])
for i, j in zip(it, it):
self.acceptors.append(
AcceptorDonor(self.atoms[i - 1], self.atoms[j - 1]))
# Now get the group sections
try:
ngrp, nst2 = psfsections['NGRP NST2'][0]
except ValueError: # pragma: no cover
raise CharmmError(
'Could not unpack GROUP pointers') # pragma: no cover
tmp = psfsections['NGRP NST2'][1]
self.groups.nst2 = nst2
# Now handle the groups
if len(psfsections['NGRP NST2'][1]) != ngrp * 3:
raise CharmmError(
'Got %d indexes for %d groups' % # pragma: no cover
(len(tmp), ngrp))
it = iter(psfsections['NGRP NST2'][1])
for i, j, k in zip(it, it, it):
self.groups.append(Group(self.atoms[i], j, k))
# Assign all of the atoms to molecules recursively
tmp = psfsections['MOLNT'][1]
set_molecules(self.atoms)
molecule_list = [a.marked for a in self.atoms]
if len(tmp) == len(self.atoms):
if molecule_list != tmp:
warnings.warn(
'Detected PSF molecule section that is WRONG. '
'Resetting molecularity.', CharmmWarning)
# We have a CHARMM PSF file; now do NUMLP/NUMLPH sections
numlp, numlph = psfsections['NUMLP NUMLPH'][0]
if numlp != 0 or numlph != 0:
raise NotImplementedError(
'Cannot currently handle PSFs with '
'lone pairs defined in the NUMLP/'
'NUMLPH section.')
# Now do the CMAPs
ncrterm = conv(psfsections['NCRTERM'][0], int,
'Number of cross-terms')
if len(psfsections['NCRTERM'][1]) != ncrterm * 8:
raise CharmmError('Got %d CMAP indexes for %d cmap terms'
% # pragma: no cover
(len(psfsections['NCRTERM']), ncrterm))
it = iter(psfsections['NCRTERM'][1])
for i, j, k, l, m, n, o, p in zip(it, it, it, it, it, it, it, it):
self.cmaps.append(
Cmap.extended(self.atoms[i - 1], self.atoms[j - 1],
self.atoms[k - 1], self.atoms[l - 1],
self.atoms[m - 1], self.atoms[n - 1],
self.atoms[o - 1], self.atoms[p - 1]))
self.unchange()
self.flags = psf_flags
def test_read_ftp_URL(self):
""" Tests genopen reading a ftp remote file """
url = 'ftp://ftp.wwpdb.org/pub/pdb/data/structures/divided/mmCIF/05/205l.cif.gz'
with closing(genopen(url, 'r')) as f:
self.assertEqual(f.read(), genopen(get_fn('205l.cif.gz')).read())
def write(struct, dest, precision=3, nobox=False, combine=False):
""" Write a Gromacs Topology File from a Structure
Parameters
----------
struct : :class:`Structure`
The structure to write to a Gromacs GRO file (must have coordinates)
dest : str or file-like
The name of a file or a file object to write the Gromacs topology to
precision : int, optional
The number of decimal places to print in the coordinates. Default 3
nobox : bool, optional
If the system does not have a periodic box defined, and this option
is True, no box will be written. If False, the periodic box will be
defined to enclose the solute with 0.5 nm clearance on all sides. If
periodic box dimensions *are* defined, this variable has no effect.
combine : 'all', None, or list of iterables, optional
Equivalent to the combine argument of the GromacsTopologyFile.write
method. If None, system atom order may be changed to meet the need
for contiguously bonded groups of atoms to be part of a single
moleculetype. All other values leave the atom order unchanged.
Default is None.
"""
def _write_atom_line(atom, atid, resid, has_vels, dest, precision):
varwidth = 5 + precision
crdfmt = '%%%d.%df' % (varwidth, precision)
velfmt = '%%%d.%df' % (varwidth, precision + 1)
dest.write('%5d%-5s%5s%5d' %
(resid, atom.residue.name[:5], atom.name[:5], atid))
dest.write((crdfmt % (atom.xx / 10))[:varwidth])
dest.write((crdfmt % (atom.xy / 10))[:varwidth])
dest.write((crdfmt % (atom.xz / 10))[:varwidth])
if has_vels:
dest.write((velfmt % (atom.vx / 10))[:varwidth])
dest.write((velfmt % (atom.vy / 10))[:varwidth])
dest.write((velfmt % (atom.vz / 10))[:varwidth])
dest.write('\n')
own_handle = False
if isinstance(dest, string_types):
dest = genopen(dest, 'w')
own_handle = True
elif not hasattr(dest, 'write'):
raise TypeError('dest must be a file name or file-like object')
dest.write('GROningen MAchine for Chemical Simulation\n')
dest.write('%5d\n' % len(struct.atoms))
has_vels = all(hasattr(a, 'vx') for a in struct.atoms)
if combine != 'all':
resid, atid = 0, 0
# use struct.split to get residue order as per topology file
split_struct = struct.split()
n_mols = sum(len(mol[1]) for mol in split_struct)
unused_atoms = list(struct.atoms)
for molid in range(n_mols):
# loop through molids so we can get the correct molecule
# according to the order they appear
molecule = [mol[0] for mol in split_struct
if molid in mol[1]][0]
new_molecule = set() # track atoms added
last_found_atom = None # track when gro and top diverge
for residue in molecule.residues:
resid += 1
for atom in residue.atoms:
# for each atom in split topology get the first
# matching occurrence in the original structure
for original_atom in unused_atoms:
if atom.type == original_atom.type and \
atom.name == original_atom.name and \
atom.residue.name == original_atom.residue.name:
if last_found_atom is not None and \
original_atom.idx != last_found_atom.idx + 1:
# a rearrangement has occurred! Need to do
# extra check that we've found the correct
# original_atom
if len(
new_molecule.intersection(
original_atom.bond_partners)
) == 0:
# original_atom must be bonded to at
# least one atom in the molecule we
# are currently writing otherwise find
# next candidate
continue
atid += 1
_write_atom_line(original_atom, atid % 100000,
resid % 100000, has_vels,
dest, precision)
new_molecule.add(original_atom)
last_found_atom = original_atom
unused_atoms.remove(original_atom)
break
else:
raise Exception("Could not find %s" % atom)
else:
for atom in struct.atoms:
resid = (atom.residue.idx + 1) % 100000
atid = (atom.idx + 1) % 100000
_write_atom_line(atom, atid, resid, has_vels, dest, precision)
# Box, in the weird format...
if struct.box is not None:
a, b, c = reduce_box_vectors(*box_lengths_and_angles_to_vectors(
*struct.box))
if all([abs(x - 90) < TINY for x in struct.box[3:]]):
dest.write('%10.5f' * 3 % (a[0] / 10, b[1] / 10, c[2] / 10))
else:
dest.write('%10.5f' * 9 %
(a[0] / 10, b[1] / 10, c[2] / 10, a[1] / 10, a[2] /
10, b[0] / 10, b[2] / 10, c[0] / 10, c[1] / 10))
dest.write('\n')
elif not nobox and struct.atoms:
# Find the extent of the molecule in all dimensions, and buffer it
# by 5 A
crds = struct.coordinates
diff = (crds.max(axis=1) - crds.min(axis=1)) / 10 + 0.5
dest.write('%10.5f' * 3 % (diff[0], diff[1], diff[2]))
dest.write('\n')
if own_handle:
dest.close()
def test_read_normal_URL(self):
""" Tests genopen reading a remote file """
url = 'https://github.com/ParmEd/ParmEd/raw/master/test/files/tripos1.mol2'
with closing(genopen(url, 'r')) as f:
self.assertEqual(f.read(), open(get_fn('tripos1.mol2')).read())
def parse(filename, skip_bonds=False):
""" Parses a Gromacs GRO file
Parameters
----------
filename : str or file-like
Name of the file or the GRO file object
skip_bonds : bool, optional
If True, skip trying to assign bonds. This can save substantial time
when parsing large files with non-standard residue names. However,
no bonds are assigned. This is OK if, for instance, the GRO file is
being parsed simply for its coordinates. This will also reduce the
accuracy of assigned atomic numbers for typical ions. Default is
False.
Returns
-------
struct : :class:`Structure`
The Structure instance instantiated with *just* residues and atoms
populated (with coordinates)
"""
struct = Structure()
if isinstance(filename, string_types):
fileobj = genopen(filename, 'r')
own_handle = True
else:
fileobj = filename
own_handle = False
try:
# Ignore the title line
fileobj.readline()
try:
natom = int(fileobj.readline().strip())
except ValueError:
raise GromacsError('Could not parse %s as GRO file' % filename)
line_parser = _AtomLineParser()
for i, line in enumerate(fileobj):
if i == natom: break
try:
atom, resname, resnum = line_parser.read(line)
except (ValueError, IndexError):
raise GromacsError('Could not parse the atom record of '
'GRO file %s' % filename)
struct.add_atom(atom, resname, resnum)
else:
# If no box exists, the break did not hit, so line still
# contains the last atom (which cannot be interpreted as a box).
# This wipes out line (IFF fileobj reached the line)
line = fileobj.readline()
if i+1 != natom:
raise GromacsError('Truncated GRO file. Found %d of %d '
'atoms' % (i+1, natom))
# Get the box from the last line if it's present
if line.strip():
try:
box = [float(x) for x in line.split()]
except ValueError:
raise GromacsError('Could not understand box line of GRO '
'file %s' % filename)
if len(box) == 3:
struct.box = [box[0]*10, box[1]*10, box[2]*10,
90.0, 90.0, 90.0]
elif len(box) == 9:
# Assume we have vectors
leng, ang = box_vectors_to_lengths_and_angles(
[box[0], box[3], box[4]]*u.nanometers,
[box[5], box[1], box[6]]*u.nanometers,
[box[7], box[8], box[2]]*u.nanometers)
a, b, c = leng.value_in_unit(u.angstroms)
alpha, beta, gamma = ang.value_in_unit(u.degrees)
struct.box = [a, b, c, alpha, beta, gamma]
finally:
if own_handle:
fileobj.close()
# Assign bonds (and improved element guesses)
if not skip_bonds:
struct.assign_bonds()
return struct
def id_format(filename):
""" Identifies the file type as an Amber restart/inpcrd file
Parameters
----------
filename : str
Name of the file to check format for
Returns
-------
is_fmt : bool
True if it is an Amber restart/inpcrd file. False otherwise
"""
if isinstance(filename, string_types):
f = genopen(filename, 'r')
lines = [f.readline() for i in range(5)]
f.close()
elif (hasattr(filename, 'readline') and hasattr(filename, 'seek')
and hasattr(filename, 'tell')):
cur = filename.tell()
lines = [filename.readline() for i in range(5)]
filename.seek(cur)
# Look for natom
words = lines[1].split()
if len(words) > 2 or len(words) < 1:
return False
try:
natom = int(words[0])
float(words[1])
except ValueError:
return False
except IndexError:
pass
# Next 3 lines, make sure we have %12.7f format. This only works if we
# have at least 6 atoms. Any fewer than that means the restart file is
# shorter than that.
try:
if natom <= 0:
return False
i = 0
for line in lines[2:]:
i += 1
if i > natom: break
for j in range(3):
j12 = j * 12
if line[j12 + 4] != '.': return False
float(line[j12:j12 + 12])
if line[j12 + 11] not in '0123456789':
return False
i += 1
if i > natom: break
for j in range(3):
j12 = j * 12 + 36
if line[j12 + 4] != '.': return False
float(line[j12:j12 + 12])
if line[j12 + 11] not in '0123456789':
return False
except (IndexError, ValueError):
return False
# Must be a restart...
return True
def write(struct, dest, vmd=False):
"""
Writes a PSF file from the stored molecule
Parameters
----------
struct : :class:`Structure`
The Structure instance from which the PSF should be written
dest : str or file-like
The place to write the output PSF file. If it has a "write"
attribute, it will be used to print the PSF file. Otherwise, it will
be treated like a string and a file will be opened, printed, then
closed
vmd : bool
If True, it will write out a PSF in the format that VMD prints it in
(i.e., no NUMLP/NUMLPH or MOLNT sections)
Examples
--------
>>> cs = CharmmPsfFile('testfiles/test.psf')
>>> cs.write_psf('testfiles/test2.psf')
"""
# See if this is an extended format
try:
ext = 'EXT' in struct.flags
except AttributeError:
ext = True
# See if this is an XPLOR format
try:
xplor = 'XPLOR' in struct.flags
except AttributeError:
for atom in struct.atoms:
if isinstance(atom.type, string_types):
xplor = True
break
else:
xplor = False
own_handle = False
# Index the atoms and residues TODO delete
if isinstance(dest, string_types):
own_handle = True
dest = genopen(dest, 'w')
# Assign the formats we need to write with
if ext:
atmfmt1 = ('%10d %-8s %-8i %-8s %-8s %6d %10.6f %13.4f' + 11 * ' ')
atmfmt2 = ('%10d %-8s %-8i %-8s %-8s %-6s %10.6f %13.4f' +
11 * ' ')
intfmt = '%10d' # For pointers
else:
atmfmt1 = ('%8d %-4s %-4i %-4s %-4s %4d %10.6f %13.4f' + 11 * ' ')
atmfmt2 = ('%8d %-4s %-4i %-4s %-4s %-4s %10.6f %13.4f' + 11 * ' ')
intfmt = '%8d' # For pointers
# Now print the header then the title
dest.write('PSF CHEQ ')
if hasattr(struct, 'flags'):
dest.write(' '.join(f for f in struct.flags
if f not in ('CHEQ', )))
else:
dest.write('EXT') # EXT is always active if no flags present
if xplor:
dest.write(' XPLOR')
dest.write('\n\n')
if isinstance(struct.title, string_types):
dest.write(intfmt % 1 + ' !NTITLE\n')
dest.write('%s\n\n' % struct.title)
else:
dest.write(intfmt % len(struct.title) + ' !NTITLE\n')
dest.write('\n'.join(struct.title) + '\n\n')
# Now time for the atoms
dest.write(intfmt % len(struct.atoms) + ' !NATOM\n')
# atmfmt1 is for CHARMM format (i.e., atom types are integers)
# atmfmt is for XPLOR format (i.e., atom types are strings)
add = 0 if struct.residues[
0].number > 0 else 1 - struct.residues[0].number
for i, atom in enumerate(struct.atoms):
typ = atom.type
if isinstance(atom.type, str):
fmt = atmfmt2
if not atom.type: typ = atom.name
else:
fmt = atmfmt1
segid = atom.residue.segid or 'SYS'
atmstr = fmt % (i + 1, segid, atom.residue.number + add,
atom.residue.name, atom.name, typ, atom.charge,
atom.mass)
if hasattr(atom, 'props'):
dest.write(atmstr + ' '.join(atom.props) + '\n')
else:
dest.write('%s\n' % atmstr)
dest.write('\n')
# Bonds
dest.write(intfmt % len(struct.bonds) + ' !NBOND: bonds\n')
for i, bond in enumerate(struct.bonds):
dest.write((intfmt * 2) % (bond.atom1.idx + 1, bond.atom2.idx + 1))
if i % 4 == 3: # Write 4 bonds per line
dest.write('\n')
# See if we need to terminate
if len(struct.bonds) % 4 != 0 or len(struct.bonds) == 0:
dest.write('\n')
dest.write('\n')
# Angles
dest.write(intfmt % len(struct.angles) + ' !NTHETA: angles\n')
for i, angle in enumerate(struct.angles):
dest.write(
(intfmt * 3) % (angle.atom1.idx + 1, angle.atom2.idx + 1,
angle.atom3.idx + 1))
if i % 3 == 2: # Write 3 angles per line
dest.write('\n')
# See if we need to terminate
if len(struct.angles) % 3 != 0 or len(struct.angles) == 0:
dest.write('\n')
dest.write('\n')
# Dihedrals
# impropers need to be split off in the "improper" section.
# PSF files need to have each dihedral listed *only* once. So count the
# number of unique dihedrals
nnormal = 0
torsions = set()
for dih in struct.dihedrals:
if dih.improper: continue
a1, a2, a3, a4 = dih.atom1, dih.atom2, dih.atom3, dih.atom4
if (a1, a2, a3, a4) in torsions or (a4, a3, a2, a1) in torsions:
continue
nnormal += 1
torsions.add((a1, a2, a3, a4))
nimprop = sum(1 for dih in struct.dihedrals if dih.improper)
dest.write(intfmt % nnormal + ' !NPHI: dihedrals\n')
torsions = set()
c = 0
for dih in struct.dihedrals:
if dih.improper: continue
a1, a2, a3, a4 = dih.atom1, dih.atom2, dih.atom3, dih.atom4
if (a1, a2, a3, a4) in torsions or (a4, a3, a2, a1) in torsions:
continue
dest.write((intfmt * 4) %
(a1.idx + 1, a2.idx + 1, a3.idx + 1, a4.idx + 1))
torsions.add((a1, a2, a3, a4))
if c % 2 == 1: # Write 2 dihedrals per line
dest.write('\n')
c += 1
# See if we need to terminate
if nnormal % 2 != 0 or nnormal == 0:
dest.write('\n')
dest.write('\n')
# Impropers
nimprop += len(struct.impropers)
dest.write(intfmt % (nimprop) + ' !NIMPHI: impropers\n')
def improp_gen(struct):
for imp in struct.impropers:
yield (imp.atom1, imp.atom2, imp.atom3, imp.atom4)
for dih in struct.dihedrals:
if dih.improper:
yield (dih.atom1, dih.atom2, dih.atom3, dih.atom4)
for i, (a1, a2, a3, a4) in enumerate(improp_gen(struct)):
dest.write((intfmt * 4) %
(a1.idx + 1, a2.idx + 1, a3.idx + 1, a4.idx + 1))
if i % 2 == 1: # Write 2 dihedrals per line
dest.write('\n')
# See if we need to terminate
if nimprop % 2 != 0 or nimprop == 0:
dest.write('\n')
dest.write('\n')
# Donor section
dest.write(intfmt % len(struct.donors) + ' !NDON: donors\n')
for i, don in enumerate(struct.donors):
dest.write((intfmt * 2) % (don.atom1.idx + 1, don.atom2.idx + 1))
if i % 4 == 3: # 4 donors per line
dest.write('\n')
if len(struct.donors) % 4 != 0 or len(struct.donors) == 0:
dest.write('\n')
dest.write('\n')
# Acceptor section
dest.write(intfmt % len(struct.acceptors) + ' !NACC: acceptors\n')
for i, acc in enumerate(struct.acceptors):
dest.write((intfmt * 2) % (acc.atom1.idx + 1, acc.atom2.idx + 1))
if i % 4 == 3: # 4 donors per line
dest.write('\n')
if len(struct.acceptors) % 4 != 0 or len(struct.acceptors) == 0:
dest.write('\n')
dest.write('\n')
# NNB section ??
dest.write(intfmt % 0 + ' !NNB\n\n')
for i in range(len(struct.atoms)):
dest.write(intfmt % 0)
if i % 8 == 7: # Write 8 0's per line
dest.write('\n')
if len(struct.atoms) % 8 != 0: dest.write('\n')
dest.write('\n')
# Group section
try:
nst2 = struct.groups.nst2
except AttributeError:
nst2 = 0
dest.write((intfmt * 2) % (len(struct.groups) or 1, nst2))
dest.write(' !NGRP NST2\n')
if struct.groups:
for i, gp in enumerate(struct.groups):
dest.write((intfmt * 3) % (gp.atom.idx, gp.type, gp.move))
if i % 3 == 2: dest.write('\n')
if len(struct.groups) % 3 != 0 or len(struct.groups) == 0:
dest.write('\n')
else:
typ = 1 if abs(sum(a.charge for a in struct.atoms)) < 1e-4 else 2
dest.write((intfmt * 3) % (0, typ, 0))
dest.write('\n')
dest.write('\n')
# The next two sections are never found in VMD prmtops...
if not vmd:
# Molecule section; first set molecularity
set_molecules(struct.atoms)
mollist = [a.marked for a in struct.atoms]
dest.write(intfmt % max(mollist) + ' !MOLNT\n')
for i, atom in enumerate(struct.atoms):
dest.write(intfmt % atom.marked)
if i % 8 == 7: dest.write('\n')
if len(struct.atoms) % 8 != 0: dest.write('\n')
dest.write('\n')
# NUMLP/NUMLPH section
dest.write((intfmt * 2) % (0, 0) + ' !NUMLP NUMLPH\n')
dest.write('\n')
# CMAP section
dest.write(intfmt % len(struct.cmaps) + ' !NCRTERM: cross-terms\n')
for i, cmap in enumerate(struct.cmaps):
dest.write(
(intfmt * 8) %
(cmap.atom1.idx + 1, cmap.atom2.idx + 1, cmap.atom3.idx + 1,
cmap.atom4.idx + 1, cmap.atom2.idx + 1, cmap.atom3.idx + 1,
cmap.atom4.idx + 1, cmap.atom5.idx + 1))
dest.write('\n')
# Done!
# If we opened our own handle, close it
if own_handle:
dest.close()
