def _parse(self, fname):
crdfile = open(fname, 'r')
readingHeader = True
while readingHeader:
line = crdfile.readline()
if not len(line):
raise CharmmFileError('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) as e:
raise CharmmFileError('Problem parsing CHARMM restart')
self._scan(crdfile, '!XOLD')
self._get_formatted_crds(crdfile, self.positionsold)
self._scan(crdfile, '!VX')
self._get_formatted_crds(crdfile, self.velocities)
self._scan(crdfile, '!X')
self._get_formatted_crds(crdfile, self.positions)
# Convert velocities to angstroms/ps
self.velocities = [v * ONE_TIMESCALE for v in self.velocities]
# Add units to positions and velocities
self.positions = u.Quantity(self.positions, u.angstroms)
self.positionsold = u.Quantity(self.positionsold, u.angstroms)
self.velocities = u.Quantity(self.velocities,
u.angstroms / u.picoseconds)
def _parse(self, fname):
with open(fname, 'r') as crdfile:
line = crdfile.readline()
while len(line.strip()) == 0: # Skip whitespace, as a precaution
line = crdfile.readline()
intitle = True
while intitle:
self.title.append(line.strip())
line = crdfile.readline()
if len(line.strip()) == 0:
intitle = False
elif line.strip()[0] != '*':
intitle = False
else:
intitle = True
while len(line.strip()) == 0: # Skip whitespace
line = crdfile.readline()
try:
self.natom = int(line.strip().split()[0])
for _ in range(self.natom):
line = crdfile.readline().strip().split()
self.atomno.append(int(line[0]))
self.resno.append(int(line[1]))
self.resname.append(line[2])
self.attype.append(line[3])
pos = Vec3(float(line[4]), float(line[5]), float(line[6]))
self.positions.append(pos)
self.segid.append(line[7])
self.weighting.append(float(line[9]))
if self.natom != len(self.positions):
raise CharmmFileError(
"Error parsing CHARMM .crd file: %d "
"atoms requires %d positions (not %d)" %
(self.natom, self.natom, len(self.positions)))
except (ValueError, IndexError):
raise CharmmFileError('Error parsing CHARMM coordinate file')
# Apply units to the positions now. Do it this way to allow for
# (possible) numpy functionality in the future.
self.positions = u.Quantity(self.positions, u.angstroms)
def _convert(data, type, msg=''):
"""
Converts a data type to a desired type, raising CharmmFileError if it
fails
"""
try:
return type(data)
except ValueError:
raise CharmmFileError('Could not convert %s to %s' % (msg, type))
def _get_formatted_crds(self, crdfile, crds):
for row in range(self.natom):
line = crdfile.readline()
if not line:
raise CharmmFileError('Premature end of file')
if len(line) < 3 * CHARMMLEN:
raise CharmmFileError("Less than 3 coordinates present in "
"coordinate row or positions may be "
"truncated.")
line = line.replace('D', 'E') # CHARMM uses 'D' for exponentials
# CHARMM uses fixed format (len = CHARMMLEN = 22) for crds in .rst's
c = Vec3(float(line[0:CHARMMLEN]),
float(line[CHARMMLEN:2 * CHARMMLEN]),
float(line[2 * CHARMMLEN:3 * CHARMMLEN]))
crds.append(c)
def _scan(self, handle, str, r=0): # read lines in file until str is found
scanning = True
if (r): handle.seek(0)
while scanning:
line = handle.readline()
if not line:
raise CharmmFileError('Premature end of file')
if len(line.strip()) != 0:
if line.strip().split()[0][0:len(str)] == str:
scanning = False
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'
try:
self._handle = open(fname, mode, encoding='utf-8')
except IOError as e:
raise CharmmFileError(str(e))
self.closed = False
self.line_number = 0
def readTopologyFile(self, tfile):
"""Reads _only_ the atom type definitions from a topology file. This is
unnecessary for versions 36 and later of the CHARMM force field.
Parameters
----------
tfile : str
: Name of the CHARMM TOPology file to read
Notes
-----
The CHARMM TOPology file is also called a Residue Topology File
"""
conv = CharmmParameterSet._convert
if isinstance(tfile, str):
own_handle = True
f = CharmmFile(tfile)
else:
own_handle = False
f = tfile
for line in f:
line = line.strip()
if line[:4] != 'MASS': continue
words = line.split()
try:
idx = conv(words[1], int, 'atom type')
name = words[2]
mass = conv(words[3], float, 'atom mass')
except IndexError:
raise CharmmFileError('Could not parse MASS section of %s' %
tfile)
# The parameter file might or might not have an element name
try:
elem = words[4]
atomic_number = get_by_symbol(elem).atomic_number
except (IndexError, KeyError):
# Figure it out from the mass
masselem = Element.getByMass(mass)
if masselem is None:
atomic_number = 0 # Extra point or something
else:
atomic_number = masselem.atomic_number
atype = AtomType(name=name,
number=idx,
mass=mass,
atomic_number=atomic_number)
self.atom_types_str[atype.name] = atype
self.atom_types_int[atype.number] = atype
self.atom_types_tuple[(atype.name, atype.number)] = atype
if own_handle: f.close()
def readParameterFile(self, pfile, permissive=False):
"""Reads all of the parameters from a parameter file. Versions 36 and later
of the CHARMM force field files have an ATOMS section defining all of
the atom types. Older versions need to load this information from the
RTF/TOP files.
Parameters
----------
pfile : str
Name of the CHARMM PARameter file to read
permissive : bool
Accept non-bonbded parameters for undefined atom types (default:
False).
Notes
-----
The atom types must all be loaded by the end of this routine. Either
supply a PAR file with atom definitions in them or read in a RTF/TOP
file first. Failure to do so will result in a raised RuntimeError.
"""
conv = CharmmParameterSet._convert
if isinstance(pfile, str):
own_handle = True
f = CharmmFile(pfile)
else:
own_handle = False
f = pfile
# What section are we parsing?
section = None
# The current cmap we are building (these span multiple lines)
current_cmap = None
current_cmap_data = []
current_cmap_res = 0
nonbonded_types = dict() # Holder
parameterset = None
read_first_nonbonded = False
previous = ''
for line in f:
line = previous+line.strip()
previous = ''
if line.endswith('-'):
# This will be continued on the next line.
previous = line[:-1]
continue
if line.startswith('!'):
# This is a comment.
continue
if not line:
# This is a blank line
continue
if parameterset is None and line.strip().startswith('*>>'):
parameterset = line.strip()[1:78]
continue
# Set section if this is a section header
if line.startswith('ATOM'):
section = 'ATOMS'
continue
if line.startswith('BOND'):
section = 'BONDS'
continue
if line.startswith('ANGL') or line.startswith('THET'):
section = 'ANGLES'
continue
if line.startswith('DIHE') or line.startswith('PHI'):
section = 'DIHEDRALS'
continue
if line.startswith('IMPR') or line.startswith('IMPH'):
section = 'IMPROPER'
continue
if line.startswith('CMAP'):
section = 'CMAP'
continue
if line.startswith('NONB') or line.startswith('NBON'):
read_first_nonbonded = False
section = 'NONBONDED'
fields = line.upper().split()
if 'NBXMOD' in fields:
nbxmod = int(fields[fields.index('NBXMOD')+1])
if nbxmod not in list(range(-5, 6)):
raise CharmmFileError('Unsupported value for NBXMOD: %d' % nbxmod)
self.nbxmod = nbxmod
if 'E14FAC' in fields:
self.e14fac = float(fields[fields.index('E14FAC')+1])
continue
if line.startswith('NBFIX'):
section = 'NBFIX'
continue
if line.startswith('THOLE'):
section = 'NBTHOLE'
continue
if line.startswith('HBOND'):
section = None
continue
# It seems like files? sections? can be terminated with 'END'
if line.startswith('END'): # should this be case-insensitive?
section = None
continue
# If we have no section, skip
if section is None: continue
# Now handle each section specifically
if section == 'ATOMS':
if not line.startswith('MASS'): continue # Should this happen?
words = line.split()
try:
idx = conv(words[1], int, 'atom type')
name = words[2]
mass = conv(words[3], float, 'atom mass')
except IndexError:
raise CharmmFileError('Could not parse MASS section.')
# The parameter file might or might not have an element name
try:
elem = words[4]
atomic_number = get_by_symbol(elem).atomic_number
except (IndexError, KeyError):
# Figure it out from the mass
masselem = Element.getByMass(mass)
if masselem is None:
atomic_number = 0 # Extra point or something
else:
atomic_number = masselem.atomic_number
atype = AtomType(name=name, number=idx, mass=mass,
atomic_number=atomic_number)
self.atom_types_str[atype.name] = atype
self.atom_types_int[atype.number] = atype
self.atom_types_tuple[(atype.name, atype.number)] = atype
continue
if section == 'BONDS':
words = line.split()
try:
type1 = words[0]
type2 = words[1]
k = conv(words[2], float, 'bond force constant')
req = conv(words[3], float, 'bond equilibrium dist')
except IndexError:
raise CharmmFileError('Could not parse bonds.')
key = (min(type1, type2), max(type1, type2))
self.bond_types[key] = BondType(k, req)
continue
if section == 'ANGLES':
words = line.split()
try:
type1 = words[0]
type2 = words[1]
type3 = words[2]
k = conv(words[3], float, 'angle force constant')
theteq = conv(words[4], float, 'angle equilibrium value')
except IndexError:
raise CharmmFileError('Could not parse angles.')
key = (min(type1, type3), type2, max(type1, type3))
self.angle_types[key] = AngleType(k, theteq)
# See if we have a urey-bradley
try:
ubk = conv(words[5], float, 'Urey-Bradley force constant')
ubeq = conv(words[6], float, 'Urey-Bradley equil. value')
ubtype = UreyBradleyType(ubk, ubeq)
except IndexError:
ubtype = NoUreyBradley
self.urey_bradley_types[key] = ubtype
continue
if section == 'DIHEDRALS':
words = line.split()
try:
type1 = words[0]
type2 = words[1]
type3 = words[2]
type4 = words[3]
k = conv(words[4], float, 'dihedral force constant')
n = conv(words[5], float, 'dihedral periodicity')
phase = conv(words[6], float, 'dihedral phase')
except IndexError:
raise CharmmFileError('Could not parse dihedrals.')
# Torsion can be in either direction. Sort by end groups first,
# then sort by middle 2
if type1 < type4:
key = (type1, type2, type3, type4)
elif type1 > type4:
key = (type4, type3, type2, type1)
else:
# OK, we need to sort by the middle atoms now
if type2 < type3:
key = (type1, type2, type3, type4)
else:
key = (type4, type3, type2, type1)
# See if this is a second (or more) term of the dihedral group
# that's already present.
dihedral = DihedralType(k, n, phase)
if key in self.dihedral_types:
# See if the existing dihedral type list has a term with
# the same periodicity -- If so, replace it
replaced = False
for i, dtype in enumerate(self.dihedral_types[key]):
if dtype.per == dihedral.per:
# Replace. Warn if they are different
if dtype != dihedral:
warnings.warn('Replacing dihedral %r with %r' %
(dtype, dihedral))
self.dihedral_types[key][i]=dihedral
replaced = True
break
if not replaced:
self.dihedral_types[key].append(dihedral)
else: # key not present
self.dihedral_types[key] = [dihedral]
continue
if section == 'IMPROPER':
words = line.split()
try:
type1 = words[0]
type2 = words[1]
type3 = words[2]
type4 = words[3]
k = conv(words[4], float, 'improper force constant')
theteq = conv(words[5], float, 'improper equil. value')
except IndexError:
raise CharmmFileError('Could not parse dihedrals.')
# If we have a 7th column, that is the real psi0 (and the 6th
# is just a dummy 0)
try:
tmp = conv(words[6], float, 'improper equil. value')
theteq = tmp
except IndexError:
pass # Do nothing
if type1 < type4:
key = (type1, type2, type3, type4)
elif type1 > type4:
key = (type4, type3, type2, type1)
else:
# OK, we need to sort by the middle atoms now
if type2 < type3:
key = (type1, type2, type3, type4)
else:
key = (type4, type3, type2, type1)
# check repeat improper dihedral
if key in self.improper_types:
warnings.warn('Repeat improper dihedral found %r and New param. k: %f ,thetaq: %f will use' %
(key,k,theteq))
self.improper_types[key] = ImproperType(k, theteq)
continue
if section == 'CMAP':
# This is the most complicated part, since cmap parameters span
# many lines. We won't do much error catching here.
words = line.split()
try:
holder = [float(w) for w in words]
current_cmap_data.extend(holder)
except ValueError:
# We assume this is a definition of a new CMAP, so
# terminate the last CMAP if applicable
if current_cmap is not None:
# We have a map to terminate
ty = CmapType(current_cmap_res, current_cmap_data)
self.cmap_types[current_cmap] = ty
try:
type1 = words[0]
type2 = words[1]
type3 = words[2]
type4 = words[3]
type5 = words[4]
type6 = words[5]
type7 = words[6]
type8 = words[7]
res = conv(words[8], int, 'CMAP resolution')
except IndexError:
raise CharmmFileError('Could not parse CMAP data.')
# order the torsions independently
k1 = [type1, type2, type3, type4]
k2 = [type4, type3, type2, type1]
key1 = min(k1, k2)
k1 = [type5, type6, type7, type8]
k2 = [type8, type7, type6, type5]
key2 = min(k1, k2)
current_cmap = tuple(key1 + key2)
current_cmap_res = res
current_cmap_data = []
continue
if section == 'NONBONDED':
# Now get the nonbonded values
words = line.split()
try:
atype = words[0]
# 1st column is ignored
epsilon = conv(words[2], float, 'vdW epsilon term')
rmin = conv(words[3], float, 'vdW Rmin/2 term')
except IndexError:
# If we haven't read our first nonbonded term yet, we may
# just be parsing the settings that should be used. So
# soldier on
if not read_first_nonbonded: continue
raise CharmmFileError('Could not parse nonbonded terms.')
except CharmmFileError as e:
if not read_first_nonbonded: continue
raise CharmmFileError(str(e))
else:
# OK, we've read our first nonbonded section for sure now
read_first_nonbonded = True
# See if we have 1-4 parameters
try:
# 4th column is ignored
eps14 = conv(words[5], float, '1-4 vdW epsilon term')
rmin14 = conv(words[6], float, '1-4 vdW Rmin/2 term')
except IndexError:
eps14 = rmin14 = None
nonbonded_types[atype] = [epsilon, rmin, eps14, rmin14]
continue
if section == 'NBFIX':
words = line.split()
try:
at1 = words[0]
at2 = words[1]
emin = abs(conv(words[2], float, 'NBFIX Emin'))
rmin = conv(words[3], float, 'NBFIX Rmin')
try:
emin14 = abs(conv(words[4], float, 'NBFIX Emin 1-4'))
rmin14 = conv(words[5], float, 'NBFIX Rmin 1-4')
except IndexError:
emin14 = rmin14 = None
try:
self.atom_types_str[at1].add_nbfix(at2, rmin, emin,
rmin14, emin14)
self.atom_types_str[at2].add_nbfix(at1, rmin, emin,
rmin14, emin14)
except KeyError:
# Some stream files define NBFIX terms with an atom that
# is defined in another toppar file that does not
# necessarily have to be loaded. As a result, not every
# NBFIX found here will necessarily need to be applied.
# If we can't find a particular atom type, don't bother
# adding that nbfix and press on
pass
except IndexError:
raise CharmmFileError('Could not parse NBFIX terms.')
self.nbfix_types[(min(at1, at2), max(at1, at2))] = (emin, rmin)
continue
# Here parse the possible nbthole section
if section == 'NBTHOLE':
words = line.split()
try:
at1 = words[0]
at2 = words[1]
nbt = abs(conv(words[2], float, 'NBTHOLE a'))
try:
self.atom_types_str[at1].add_nbthole(at2, nbt)
self.atom_types_str[at2].add_nbthole(at1, nbt)
except KeyError:
pass
except IndexError:
raise CharmmFileError('Could not parse NBTHOLE terms.')
self.nbthole_types[(min(at1, at2), max(at1, at2))] = (nbt)
# If there were any CMAP terms stored in the parameter set, the last one
# defined will not have been added to the set. Add it now.
if current_cmap is not None:
ty = CmapType(current_cmap_res, current_cmap_data)
self.cmap_types[current_cmap] = ty
# If in permissive mode create an atomtype for every type used in
# the nonbonded parameters. This is a work-around for when all that's
# available is a CHARMM22 inp file, which has no ATOM/MASS fields
if permissive:
try:
idx = max(self.atom_types_int.keys())+1000
except ValueError:
idx = 10000
for key in nonbonded_types:
if not key in self.atom_types_str:
atype =AtomType(name=key, number=idx, mass= float('NaN'), atomic_number= 1 )
self.atom_types_str[key] = atype
self.atom_types_int[idx] = atype
idx=idx+1
# Now we're done. Load the nonbonded types into the relevant AtomType
# instances. In order for this to work, all keys in nonbonded_types
# must be in the self.atom_types_str dict. Raise a RuntimeError if this
# is not satisfied
try:
for key in nonbonded_types:
self.atom_types_str[key].set_lj_params(*nonbonded_types[key])
except KeyError:
raise RuntimeError('Atom type %s not present in AtomType list' %
key)
if parameterset is not None: self.parametersets.append(parameterset)
if own_handle: f.close()