def test_general(self):
traj = pt.iterload("./data/Tc5b.x", "./data/Tc5b.top")
# with mask
saved_data = pt.radgyr(traj, '@CA')
data = pt.pmap(pt.radgyr, traj, '@CA')
data = pt.tools.dict_to_ndarray(data)
aa_eq(saved_data, data)
# with a series of functions
func_list = [pt.radgyr, pt.molsurf, pt.rmsd]
ref = traj[-3]
for n_cores in [2, 3]:
for func in func_list:
if func in [pt.rmsd, ]:
pout = pt.tools.dict_to_ndarray(pt.pmap(func=func,
traj=traj,
ref=ref,
n_cores=n_cores))
serial_out = flatten(func(traj, ref=ref))
else:
pout = pt.tools.dict_to_ndarray(pt.pmap(n_cores=n_cores,
func=func,
traj=traj))
serial_out = flatten(func(traj))
aa_eq(pout[0], serial_out)
# test worker
# need to test this since coverages seems not recognize partial func
from pytraj.parallel.multiprocessing_ import worker_byfunc
data = worker_byfunc(rank=2, n_cores=8, func=pt.radgyr, traj=traj, args=(), kwd={'mask': '@CA'}, iter_options={})
assert data[0] == 2, 'rank must be 2'
assert data[2] == 1, 'n_frames for rank=2 should be 1 (only 10 frames in total)'
def dssp_allresidues(traj, *args, **kwd):
'''calculate dssp for all residues. Mostly used for visulization.
Returns
-------
ndarray, shape=(n_frames, n_residues)
Examples
--------
>>> import pytraj as pt
>>> traj = pt.datafiles.load_dpdp()
>>> x = pt.dssp_allresidues(traj, simplified=True)
>>> x[0].tolist()
['C', 'E', 'E', 'E', 'E', 'C', 'C', 'C', 'C', 'E', 'E', 'E', 'E', 'E', 'C', 'C', 'E', 'E', 'E', 'E', 'C', 'C']
>>> len(x[0]) == traj.top.n_residues
True
>>> # load trajectory having waters
>>> traj = pt.datafiles.load_tz2_ortho()
>>> x = pt.dssp_allresidues(traj, simplified=True)
>>> len(x[0]) == traj.top.n_residues
True
>>> len(x[0])
1704
>>> # only calculate protein residues, use `pytraj.dssp`
>>> y = pt.dssp(traj, simplified=True)
>>> len(y[0])
13
Notes
-----
this method is not well optimized for speed.
See also
--------
calc_dssp
'''
res_labels, data = calc_dssp(traj, *args, **kwd)[:2]
top = get_topology(traj, kwd.get('top', None))
# do not need to compute again if there is no solvent or weird residues
if len(res_labels) == top.n_residues:
return data
res_indices = [int(x.split(':')[-1]) - 1 for x in res_labels]
if not PY3:
new_data = np.empty((traj.n_frames, traj.top.n_residues), dtype='S2')
else: # pragma no cover
new_data = np.empty((traj.n_frames, traj.top.n_residues), dtype='U2')
simplified = kwd.get('simplified', False)
for fid, arr in enumerate(data):
new_data[fid][:] = tools.flatten(
get_ss_per_frame(arr,
top,
res_indices,
simplified,
all_atoms=False))
return new_data
def dssp_allatoms(traj, *args, **kwd):
'''calculate dssp for all atoms
Returns
-------
ndarray, shape=(n_frames, n_atoms)
Notes
-----
this method is not well optimized for speed.
Examples
--------
>>> import pytraj as pt
>>> traj = pt.fetch_pdb('1l2y')
>>> x = pt.dssp_allatoms(traj, simplified=True)
>>> x[0, :3].tolist()
['C', 'C', 'C']
See also
--------
dssp
'''
res_labels, data = dssp(traj, *args, **kwd)[:2]
top = get_topology(traj, kwd.get('top'))
res_indices = [int(x.split(':')[-1]) - 1 for x in res_labels]
new_data = np.empty((traj.n_frames, traj.n_atoms), dtype='U2')
simplified = kwd.get('simplified', False)
for fid, arr in enumerate(data):
new_data[fid][:] = tools.flatten(
get_ss_per_frame(arr, top, res_indices, simplified,
all_atoms=True))
return new_data
def dssp_allresidues(traj, *args, **kwd):
'''calculate dssp for all residues. Mostly used for visulization.
Returns
-------
ndarray, shape=(n_frames, n_residues)
Examples
--------
>>> import pytraj as pt
>>> traj = pt.datafiles.load_dpdp()
>>> x = pt.dssp_allresidues(traj, simplified=True)
>>> x[0].tolist()
['C', 'E', 'E', 'E', 'E', 'C', 'C', 'C', 'C', 'E', 'E', 'E', 'E', 'E', 'C', 'C', 'E', 'E', 'E', 'E', 'C', 'C']
>>> len(x[0]) == traj.top.n_residues
True
>>> # load trajectory having waters
>>> traj = pt.datafiles.load_tz2_ortho()
>>> x = pt.dssp_allresidues(traj, simplified=True)
>>> len(x[0]) == traj.top.n_residues
True
>>> len(x[0])
1704
>>> # only calculate protein residues, use `pytraj.dssp`
>>> y = pt.dssp(traj, simplified=True)
>>> len(y[0])
13
Notes
-----
this method is not well optimized for speed.
See also
--------
calc_dssp
'''
res_labels, data = calc_dssp(traj, *args, **kwd)[:2]
top = get_topology(traj, kwd.get('top', None))
# do not need to compute again if there is no solvent or weird residues
if len(res_labels) == top.n_residues:
return data
res_indices = [int(x.split(':')[-1]) - 1 for x in res_labels]
if not PY3:
new_data = np.empty((traj.n_frames, traj.top.n_residues), dtype='S2')
else: # pragma no cover
new_data = np.empty((traj.n_frames, traj.top.n_residues), dtype='U2')
simplified = kwd.get('simplified', False)
for fid, arr in enumerate(data):
new_data[fid][:] = tools.flatten(get_ss_per_frame(arr,
top,
res_indices,
simplified,
all_atoms=False))
return new_data
def test_general(self):
traj = pt.iterload("./data/Tc5b.x", "./data/Tc5b.top")
# with mask
saved_data = pt.radgyr(traj, '@CA')
data = pt.pmap(pt.radgyr, traj, '@CA')
data = pt.tools.dict_to_ndarray(data)
aa_eq(saved_data, data)
# with a series of functions
func_list = [pt.radgyr, pt.molsurf, pt.rmsd]
ref = traj[-3]
for n_cores in [2, 3]:
for func in func_list:
if func in [
pt.rmsd,
]:
pout = pt.tools.dict_to_ndarray(
pt.pmap(func=func, traj=traj, ref=ref,
n_cores=n_cores))
serial_out = flatten(func(traj, ref=ref))
else:
pout = pt.tools.dict_to_ndarray(
pt.pmap(n_cores=n_cores, func=func, traj=traj))
serial_out = flatten(func(traj))
aa_eq(pout[0], serial_out)
# test worker
# need to test this since coverages seems not recognize partial func
from pytraj.parallel.multiprocessing_ import worker_byfunc
data = worker_byfunc(rank=2,
n_cores=8,
func=pt.radgyr,
traj=traj,
args=(),
kwd={'mask': '@CA'},
iter_options={})
assert data[0] == 2, 'rank must be 2'
assert data[
2] == 1, 'n_frames for rank=2 should be 1 (only 10 frames in total)'
def test_different_references(self):
traj = self.traj
func = pt.rmsd
for i in range(0, 8, 2):
ref = self.traj[i]
for n_cores in [2, 3, ]:
pout = pt.tools.dict_to_ndarray(pt.pmap(n_cores=n_cores,
func=func,
traj=traj,
ref=ref))
serial_out = flatten(func(traj, ref=ref))
aa_eq(pout[0], serial_out)
def test_different_references(self):
traj = self.traj
func = pt.rmsd
for i in range(0, 8, 2):
ref = self.traj[i]
for n_cores in [
2,
3,
]:
pout = pt.tools.dict_to_ndarray(
pt.pmap(n_cores=n_cores, func=func, traj=traj, ref=ref))
serial_out = flatten(func(traj, ref=ref))
aa_eq(pout[0], serial_out)
def dssp_allatoms(traj, *args, **kwd):
'''calculate dssp for all atoms
Returns
-------
ndarray, shape=(n_frames, n_atoms)
Notes
-----
this method is not well optimized for speed.
Examples
--------
>>> import pytraj as pt
>>> traj = pt.fetch_pdb('1l2y')
>>> x = pt.dssp_allatoms(traj, simplified=True)
>>> x[0, :3].tolist()
['C', 'C', 'C']
See also
--------
calc_dssp
'''
res_labels, data = calc_dssp(traj, *args, **kwd)[:2]
top = get_topology(traj, kwd.get('top'))
res_indices = [int(x.split(':')[-1]) - 1 for x in res_labels]
if not PY3:
new_data = np.empty((traj.n_frames, traj.n_atoms), dtype='S2')
else: # pragma: no cover
new_data = np.empty((traj.n_frames, traj.n_atoms), dtype='U2')
simplified = kwd.get('simplified', False)
for fid, arr in enumerate(data):
new_data[fid][:] = tools.flatten(get_ss_per_frame(arr,
top,
res_indices,
simplified,
all_atoms=True))
return new_data
'''print unique residue names from all mol2 files in current folder
python ./residues_from_mol2.py
'''
import parmed as pmd
from glob import glob
from pytraj.tools import flatten
plist = [pmd.load_file(fname, structure=True) for fname in glob('*.mol2')]
reslist = flatten([[res.name for res in p.residues] for p in plist])
print(set(reslist))