def _guess_atomic_number(name, residue=None):
""" Guesses the atomic number """
# Special-case single-atom residues, which are almost always ions
name = ''.join(c for c in name if c.isalpha())
if residue is None or len(residue.atoms) == 1:
if len(name) > 1:
try:
return AtomicNum[name[0].upper() + name[1].lower()]
except KeyError:
return AtomicNum[element_by_name(name)]
return AtomicNum[element_by_name(name)]
def _guess_atomic_number(name, residue=None):
""" Guesses the atomic number """
# Special-case single-atom residues, which are almost always ions
name = ''.join(c for c in name if c.isalpha())
if residue is None or len(residue.atoms) == 1:
if len(name) > 1:
try:
return AtomicNum[name[0].upper() + name[1].lower()]
except KeyError:
return AtomicNum[element_by_name(name)]
return AtomicNum[element_by_name(name)]
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 __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 to_parmed(self, title='', **kwargs):
"""Create a ParmEd Structure from a Compound. """
structure = pmd.Structure()
structure.title = title if title else self.name
atom_mapping = {} # For creating bonds below
guessed_elements = set()
for atom in self.particles():
atomic_number = None
name = ''.join(char for char in atom.name if not char.isdigit())
try:
atomic_number = AtomicNum[atom.name]
except KeyError:
element = element_by_name(atom.name)
if name not in guessed_elements:
warn('Guessing that "{}" is element: "{}"'.format(
atom, element))
guessed_elements.add(name)
else:
element = atom.name
atomic_number = atomic_number or AtomicNum[element]
mass = Mass[element]
pmd_atom = pmd.Atom(atomic_number=atomic_number,
name=atom.name,
mass=mass)
pmd_atom.xx, pmd_atom.xy, pmd_atom.xz = atom.pos * 10 # Angstroms
structure.add_atom(pmd_atom, resname='RES', resnum=1)
atom_mapping[atom] = pmd_atom
for atom1, atom2 in self.bonds():
bond = pmd.Bond(atom_mapping[atom1], atom_mapping[atom2])
structure.bonds.append(bond)
box = self.boundingbox
box_vector = np.empty(6)
box_vector[3] = box_vector[4] = box_vector[5] = 90.0
for dim, val in enumerate(self.periodicity):
if val:
box_vector[dim] = val * 10
else:
box_vector[dim] = box.lengths[dim] * 10 + 5
structure.box = box_vector
return structure
def read(self, line):
""" Reads a line
Parameters
----------
line : str
A line with an atom record from a GRO file
Returns
-------
atom, resname, resnum : Atom, str, int
The Atom instance, residue name, and residue number containing the
atom
"""
resnum = int(line[:5])
resname = line[5:10].strip()
atomname = line[10:15].strip()
elem = element_by_name(atomname)
atomic_number = AtomicNum[elem]
mass = Mass[elem]
atnum = int(line[15:20])
if atomic_number == 0:
atom = ExtraPoint(name=atomname, number=atnum)
else:
atom = Atom(atomic_number=atomic_number, name=atomname,
number=atnum, mass=mass)
if self._digits is None:
self._pdeci = line.index('.', 20)
self._ndeci = line.index('.', self._pdeci+1)
self._digits = self._ndeci - self._pdeci
atom.xx, atom.xy, atom.xz = (
float(line[20+i*self._digits:20+(i+1)*self._digits])*10
for i in range(3)
)
wbeg = 20 + self._digits * 3
wend = wbeg + self._digits
if line[wbeg:wend].strip():
atom.vx, atom.vy, atom.vz = (
float(line[wbeg+i*self._digits:wend+i*self._digits])*10
for i in range(3)
)
return atom, resname, resnum
def to_parmed(self, title='', **kwargs):
"""Create a ParmEd Structure from a Compound. """
structure = pmd.Structure()
structure.title = title if title else self.name
atom_mapping = {} # For creating bonds below
for atom in self.particles():
atomic_number = None
try:
atomic_number = AtomicNum[atom.name]
except KeyError:
element = element_by_name(atom.name)
warn('Guessing that {} is element: {}'.format(atom, element))
else:
element = atom.name
atomic_number = atomic_number or AtomicNum[element]
mass = Mass[element]
pmd_atom = pmd.Atom(atomic_number=atomic_number, name=atom.name,
mass=mass)
pmd_atom.xx, pmd_atom.xy, pmd_atom.xz = atom.pos * 10 # Angstroms
structure.add_atom(pmd_atom, resname='RES', resnum=1)
atom_mapping[atom] = pmd_atom
for atom1, atom2 in self.bonds():
bond = pmd.Bond(atom_mapping[atom1], atom_mapping[atom2])
structure.bonds.append(bond)
box = self.boundingbox
box_vector = np.empty(6)
box_vector[3] = box_vector[4] = box_vector[5] = 90.0
for dim, val in enumerate(self.periodicity):
if val:
box_vector[dim] = val * 10
else:
box_vector[dim] = box.lengths[dim] * 10 + 5
structure.box = box_vector
return structure
def parse(filename):
""" Parses a Gromacs GRO file
Parameters
----------
filename : str or file-like
Name of the file or the GRO file object
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)
digits = None
for i, line in enumerate(fileobj):
if i == natom: break
try:
resnum = int(line[:5])
resname = line[5:10].strip()
atomname = line[10:15].strip()
elem = element_by_name(atomname)
atomic_number = AtomicNum[elem]
mass = Mass[elem]
atnum = int(line[15:20])
if atomic_number == 0:
atom = ExtraPoint(name=atomname, number=atnum)
else:
atom = Atom(atomic_number=atomic_number, name=atomname,
number=atnum, mass=mass)
if digits is None:
pdeci = line.index('.', 20)
ndeci = line.index('.', pdeci+1)
digits = ndeci - pdeci
atom.xx, atom.xy, atom.xz = (
float(line[20+i*digits:20+(i+1)*digits])*10
for i in range(3)
)
i = 4
wbeg = (pdeci-4)+(5+ndeci)*(i-1)
wend = (pdeci-4)+(5+ndeci)*i
if line[wbeg:wend].strip():
atom.vx, atom.vy, atom.vz = (
float(line[(pdeci-3)+(6+ndeci)*i:
(pdeci-3)+(6+ndeci)*(i+1)])*10
for i in range(3, 6)
)
except (ValueError, IndexError):
raise GromacsError('Could not parse the atom record of '
'GRO file %s' % filename)
struct.add_atom(atom, resname, resnum)
# 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()
return struct
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 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