def test_load_psf():
top = psf.load_psf(get_fn('ala_ala_ala.psf'))
ref_top = md.load(get_fn('ala_ala_ala.pdb')).topology
eq(top, ref_top)
# Test the XPLOR psf format parsing
top2 = psf.load_psf(get_fn('ala_ala_ala.xpsf'))
eq(top2, ref_top)
def test_seek():
reference = TRRTrajectoryFile(get_fn("frame0.trr")).read()[0]
with TRRTrajectoryFile(get_fn("frame0.trr")) as f:
eq(len(f), len(reference))
eq(len(f.offsets), len(reference))
eq(f.tell(), 0)
eq(f.read(1)[0][0], reference[0])
eq(f.tell(), 1)
xyz = f.read(1)[0][0]
eq(xyz, reference[1])
eq(f.tell(), 2)
f.seek(0)
eq(f.tell(), 0)
xyz = f.read(1)[0][0]
eq(f.tell(), 1)
eq(xyz, reference[0])
f.seek(5)
eq(f.read(1)[0][0], reference[5])
eq(f.tell(), 6)
f.seek(-5, 1)
eq(f.tell(), 1)
eq(f.read(1)[0][0], reference[1])
def test_iterload_skip():
files = [
"frame0.nc",
"frame0.h5",
"frame0.xtc",
"frame0.trr",
"frame0.dcd",
"frame0.binpos",
"frame0.xyz",
"frame0.lammpstrj",
]
if not (on_win and on_py3):
files.append("legacy_msmbuilder_trj0.lh5")
err_msg = "failed for file %s with chunksize %i and skip %i"
for file in files:
for cs in [0, 1, 11, 100]:
for skip in [0, 1, 20, 101]:
print("testing file %s with skip=%i" % (file, skip))
t_ref = md.load(get_fn(file), top=get_fn("native.pdb"))
t = functools.reduce(
lambda a, b: a.join(b), md.iterload(get_fn(file), skip=skip, top=get_fn("native.pdb"), chunk=cs)
)
eq(t_ref.xyz[skip:], t.xyz, err_msg=err_msg % (file, cs, skip))
eq(t_ref.time[skip:], t.time, err_msg=err_msg % (file, cs, skip))
eq(t_ref.topology, t.topology, err_msg=err_msg % (file, cs, skip))
def test_dihedral_pbc():
traj_uncorrected = md.load(get_fn('1am7_uncorrected.xtc'), top=get_fn('1am7_protein.pdb'))
traj_corrected = md.load(get_fn('1am7_corrected.xtc'), top=get_fn('1am7_protein.pdb'))
epsilon = 6E-3
ang1 = md.geometry.compute_phi(traj_uncorrected, opt=False, periodic=True)[1]
ang2 = md.geometry.compute_phi(traj_corrected, opt=False, periodic=True)[1]
assert np.max(np.abs(ang1 - ang2)) < epsilon
ang1 = md.geometry.compute_phi(traj_uncorrected, opt=True, periodic=True)[1]
ang2 = md.geometry.compute_phi(traj_corrected, opt=True, periodic=True)[1]
assert np.max(np.abs(ang1 - ang2)) < epsilon
ang1 = md.geometry.compute_phi(traj_uncorrected, opt=True, periodic=True)[1]
ang2 = md.geometry.compute_phi(traj_corrected, opt=True, periodic=False)[1]
assert np.max(np.abs(ang1 - ang2)) < epsilon
ang1 = md.geometry.compute_phi(traj_uncorrected, opt=False, periodic=True)[1]
ang2 = md.geometry.compute_phi(traj_corrected, opt=False, periodic=False)[1]
assert np.max(np.abs(ang1 - ang2)) < epsilon
ang1 = md.geometry.compute_phi(traj_uncorrected, opt=True, periodic=True)[1]
ang2 = md.geometry.compute_phi(traj_corrected, opt=False, periodic=False)[1]
assert np.max(np.abs(ang1 - ang2)) < epsilon
ang1 = md.geometry.compute_phi(traj_uncorrected, opt=False, periodic=True)[1]
ang2 = md.geometry.compute_phi(traj_corrected, opt=True, periodic=True)[1]
assert np.max(np.abs(ang1 - ang2)) < epsilon
ang1 = md.geometry.compute_phi(traj_uncorrected, opt=False, periodic=True)[1]
ang2 = md.geometry.compute_phi(traj_corrected, opt=True, periodic=False)[1]
assert np.max(np.abs(ang1 - ang2)) < epsilon
def test_seek():
reference = XTCTrajectoryFile(get_fn('frame0.xtc')).read()[0]
with XTCTrajectoryFile(get_fn('frame0.xtc')) as f:
eq(f.tell(), 0)
eq(f.read(1)[0][0], reference[0])
eq(f.tell(), 1)
xyz = f.read(1)[0][0]
eq(xyz, reference[1])
eq(f.tell(), 2)
f.seek(0)
eq(f.tell(), 0)
xyz = f.read(1)[0][0]
eq(f.tell(), 1)
eq(xyz, reference[0])
f.seek(5)
eq(f.read(1)[0][0], reference[5])
eq(f.tell(), 6)
f.seek(-5, 1)
eq(f.tell(), 1)
eq(f.read(1)[0][0], reference[1])
def test_kappa():
traj = md.load(get_fn("tip3p_300K_1ATM.xtc"), top=get_fn("tip3p_300K_1ATM.pdb"))
kappa = md.geometry.isothermal_compressability_kappa_T(traj, temperature)
reference = 2.05427E-10 * 1E5 # m^3 / J to 1 / bar. Data from gromacs. See above comment
# 20% tolerance.
assert abs((kappa - reference) / reference) < 2E-1, "Compressability tolerance not met!"
def test_load():
filenames = [
"frame0.xtc",
"frame0.trr",
"frame0.dcd",
"frame0.binpos",
"traj.h5",
"frame0.nc",
"traj.h5",
"frame0.lammpstrj",
"frame0.xyz",
]
num_block = 3
for filename in filenames:
t0 = md.load(get_fn(filename), top=nat, discard_overlapping_frames=True)
t1 = md.load(get_fn(filename), top=nat, discard_overlapping_frames=False)
t2 = md.load([get_fn(filename) for i in xrange(num_block)], top=nat, discard_overlapping_frames=False)
t3 = md.load([get_fn(filename) for i in xrange(num_block)], top=nat, discard_overlapping_frames=True)
# these don't actually overlap, so discard_overlapping_frames should
# have no effect. the overlap is between the last frame of one and the
# first frame of the next.
yield lambda: eq(t0.n_frames, t1.n_frames)
yield lambda: eq(t0.n_frames * num_block, t2.n_frames)
yield lambda: eq(t3.n_frames, t2.n_frames)
def test_lprmsd_5():
t = md.load(get_fn('frame0.h5'))
t1 = md.load(get_fn('frame0.h5'))
r = md.rmsd(t, t1, 0)
a = md.lprmsd(t, t1, 0, permute_groups=[[]], superpose=True)
eq(a, r, decimal=3)
def test_seek():
reference = BINPOSTrajectoryFile(get_fn('frame0.binpos')).read()
with BINPOSTrajectoryFile(get_fn('frame0.binpos')) as f:
xyz = f.read(1)[0]
eq(xyz, reference[0])
eq(f.tell(), 1)
xyz = f.read(1)[0]
eq(xyz, reference[1])
eq(f.tell(), 2)
f.seek(0)
eq(f.tell(), 0)
xyz = f.read(1)[0]
eq(f.tell(), 1)
eq(xyz, reference[0])
f.seek(5)
eq(f.read(1)[0], reference[5])
eq(f.tell(), 6)
f.seek(-5, 1)
eq(f.tell(), 1)
xyz = f.read(1)[0]
eq(xyz, reference[1])
f.seek(0, 2)
eq(f.tell(), len(reference))
def deprecated_test_fah_core17_1():
from mdtraj.utils import six
from mdtraj.testing import get_fn, eq
filename = get_fn('frame0.xtc')
tempdir = tempfile.mkdtemp()
tar_filename = os.path.join(tempdir, "results-000.tar.bz2")
archive = tarfile.open(tar_filename, mode='w:bz2')
tar = tarfile.open(tar_filename, "w:bz2")
tar.add(filename, arcname="positions.xtc")
tar.close()
shutil.copy(tar_filename, os.path.join(tempdir, "results-001.tar.bz2"))
trj0 = md.load(get_fn("frame0.xtc"), top=get_fn("frame0.h5"))
output_filename = os.path.join(tempdir, "traj.h5")
fah.concatenate_core17(tempdir, trj0, output_filename)
trj = md.load(output_filename)
eq(trj.n_atoms, trj0.n_atoms)
eq(trj.n_frames, trj0.n_frames * 2)
shutil.copy(tar_filename, os.path.join(tempdir, "results-002.tar.bz2"))
fah.concatenate_core17(tempdir, trj0, output_filename)
# Should notice the new file and append it to the HDF file.
trj = md.load(output_filename)
eq(trj.n_atoms, trj0.n_atoms)
eq(trj.n_frames, trj0.n_frames * 3)
def test_seek():
reference = XTCTrajectoryFile(get_fn('frame0.xtc')).read()[0]
with XTCTrajectoryFile(get_fn('frame0.xtc')) as f:
eq(f.tell(), 0)
eq(f.read(1)[0][0], reference[0])
eq(f.tell(), 1)
xyz = f.read(1)[0][0]
eq(xyz, reference[1])
eq(f.tell(), 2)
f.seek(0)
eq(f.tell(), 0)
xyz = f.read(1)[0][0]
eq(f.tell(), 1)
eq(xyz, reference[0])
f.seek(5) # offset array is going to be built
assert len(f.offsets) == len(reference)
eq(f.read(1)[0][0], reference[5])
eq(f.tell(), 6)
f.seek(-5, 1)
eq(f.tell(), 1)
eq(f.read(1)[0][0], reference[1])
def test_seek_read_mode():
# Test the seek/tell capacity of the different TrajectoryFile objects in
# read mode. Basically, we just seek around the files and read different
# segments, keeping track of our location manually and checking with both
# tell() and by checking that the right coordinates are actually returned
files = [(md.formats.NetCDFTrajectoryFile, 'frame0.nc'),
(md.formats.HDF5TrajectoryFile, 'frame0.h5'),
(md.formats.XTCTrajectoryFile, 'frame0.xtc'),
(md.formats.TRRTrajectoryFile, 'frame0.trr'),
(md.formats.DCDTrajectoryFile, 'frame0.dcd'),
(md.formats.MDCRDTrajectoryFile, 'frame0.mdcrd'),
(md.formats.BINPOSTrajectoryFile, 'frame0.binpos'),
(md.formats.LH5TrajectoryFile, 'legacy_msmbuilder_trj0.lh5'),
(md.formats.DTRTrajectoryFile,'frame0.dtr/clickme.dtr'),]
for a, b in files:
point = 0
xyz = md.load(get_fn(b), top=get_fn('native.pdb')).xyz
length = len(xyz)
kwargs = {}
if a is md.formats.MDCRDTrajectoryFile:
kwargs = {'n_atoms': 22}
with a(get_fn(b), **kwargs) as f:
for i in range(100):
r = np.random.rand()
if r < 0.25:
offset = np.random.randint(-5, 5)
if 0 < point + offset < length:
point += offset
f.seek(offset, 1)
else:
f.seek(0)
point = 0
if r < 0.5:
offset = np.random.randint(1, 10)
if point + offset < length:
read = f.read(offset)
if a not in [md.formats.BINPOSTrajectoryFile, md.formats.LH5TrajectoryFile]:
read = read[0]
readlength = len(read)
read = mdtraj.utils.in_units_of(read, f.distance_unit, 'nanometers')
eq(xyz[point:point+offset], read)
point += readlength
elif r < 0.75:
offset = np.random.randint(low=-100, high=0)
try:
f.seek(offset, 2)
point = length + offset
except NotImplementedError:
# not all of the *TrajectoryFiles currently support
# seeking from the end, so we'll let this pass if they
# say that they dont implement this.
pass
else:
offset = np.random.randint(100)
f.seek(offset, 0)
point = offset
eq(f.tell(), point)
def test_lprmsd_3():
# resolve rotation and permutation togetehr
t1 = md.load(get_fn('alanine-dipeptide-explicit.pdb'))[0]
t2 = md.load(get_fn('alanine-dipeptide-explicit.pdb'))[0]
h2o_o_indices = [a.index for a in t1.topology.atoms if a.residue.name == 'HOH' and a.element.symbol == 'O'][0:20]
h2o_h_indices = [a.index for a in t1.topology.atoms if a.residue.name == 'HOH' and a.element.symbol == 'H'][0:20]
backbone_indices = [a.index for a in t1.topology.atoms if a.element.symbol in ['C', 'N']][:5]
# scramble two groups of indices
t2.xyz[:, random.permutation(h2o_o_indices)] = t2.xyz[:, h2o_o_indices]
t2.xyz[:, random.permutation(h2o_h_indices)] = t2.xyz[:, h2o_h_indices]
# offset the backbone indices slightly
t2.xyz[:, backbone_indices] += 0.001 * random.randn(len(backbone_indices), 3)
# rotate everything
rot = rotation_matrix_from_quaternion(uniform_quaternion())
t2.xyz[0].dot(rot)
print('raw distinguishable indices', backbone_indices)
atom_indices = np.concatenate((h2o_o_indices, backbone_indices))
value = md.lprmsd(t2, t1, atom_indices=atom_indices, permute_groups=[h2o_o_indices])
t1.xyz[:, h2o_o_indices] += random.randn(len(h2o_o_indices), 3)
print('final value', value)
assert value[0] < 1e-2
def test_atom_indices_1():
atom_indices = np.arange(10)
top = md.load(get_fn("native.pdb"))
t0 = md.load(get_fn("frame0.mdcrd"), top=top)
t1 = md.load(get_fn("frame0.mdcrd"), top=top, atom_indices=atom_indices)
eq(t0.xyz[:, atom_indices], t1.xyz)
def test_read_1():
with MDCRDTrajectoryFile(get_fn("frame0.mdcrd"), n_atoms=22) as f:
xyz, _ = f.read()
with MDCRDTrajectoryFile(get_fn("frame0.mdcrd"), n_atoms=22) as f:
xyz3, _ = f.read(stride=3)
eq(xyz[::3], xyz3)
def test_atom_indices_0():
atom_indices = np.arange(10)
with MDCRDTrajectoryFile(get_fn("frame0.mdcrd"), n_atoms=22) as f:
xyz0, box0 = f.read(n_frames=1)
with MDCRDTrajectoryFile(get_fn("frame0.mdcrd"), n_atoms=22) as f:
xyz1, box1 = f.read(n_frames=1, atom_indices=atom_indices)
eq(xyz0[:, atom_indices], xyz1)
def test_multiread():
reference = md.load(get_fn('frame0.xyz'), top=get_fn('native.pdb'))
with XYZTrajectoryFile(get_fn('frame0.xyz')) as f:
xyz0 = f.read(n_frames=1)
xyz1 = f.read(n_frames=1)
eq(reference.xyz[0], xyz0[0]/10)
eq(reference.xyz[1], xyz1[0]/10)
def test_load_mol2():
trj = md.load(get_fn('imatinib.mol2'))
ref_trj = md.load(get_fn('imatinib.pdb'))
eq(trj.xyz, ref_trj.xyz)
ref_top, ref_bonds = ref_trj.top.to_dataframe()
top, bonds = trj.top.to_dataframe()
eq(bonds, ref_bonds)
def test_load():
t = load(get_fn('test_hoomdxml.dcd'), top=get_fn('well-mixed.hoomdxml'))
top = load(get_fn('well-mixed.hoomdxml')).topology
eq(t.topology, top)
eq(top.n_atoms, 6600)
eq(top.n_bonds, 3200)
eq(top.n_chains, 4000)
eq(top.n_chains, top.n_residues) # hoomdxml assumes each chain is 1 residue
def test_read_0():
with MDCRDTrajectoryFile(get_fn("frame0.mdcrd"), n_atoms=22) as f:
xyz, box = f.read()
assert box is None
with MDCRDTrajectoryFile(get_fn("frame0.mdcrdbox"), n_atoms=22) as f:
xyz2, box = f.read()
eq(box.shape, (1002, 3))
eq(xyz, xyz2)
def test_multiread():
reference = md.load(get_fn("frame0.mdcrd"), top=get_fn("native.pdb"))
with MDCRDTrajectoryFile(get_fn("frame0.mdcrd"), n_atoms=22) as f:
xyz0, box0 = f.read(n_frames=1)
xyz1, box1 = f.read(n_frames=1)
eq(reference.xyz[0], xyz0[0] / 10)
eq(reference.xyz[1], xyz1[0] / 10)
def test_sasa_3():
traj_ref = np.loadtxt(get_fn('gmx_sasa.dat'))
traj = md.load(get_fn('frame0.h5'))
traj_sasa = md.geometry.shrake_rupley(traj, probe_radius=0.14, n_sphere_points = 960)
# the algorithm used by gromacs' g_sas is slightly different than the one
# used here, so the results are not exactly the same
assert_array_almost_equal(traj_sasa, traj_ref, decimal=1)
def test_multiread():
reference = md.load(get_fn('frame0.lammpstrj'), top=get_fn('native.pdb'))
with LAMMPSTrajectoryFile(get_fn('frame0.lammpstrj')) as f:
xyz0, _, _ = f.read(n_frames=1)
xyz1, _, _ = f.read(n_frames=1)
eq(reference.xyz[0], xyz0[0]/10)
eq(reference.xyz[1], xyz1[0]/10)
def test_compute_neighbors_3():
traj = md.load(get_fn('test_good.nc'), top=get_fn('test.parm7'))
query_indices = traj.top.select('residue 1')
cutoff = 1.0
value = md.compute_neighbors(traj, cutoff, query_indices)
reference = compute_neighbors_reference(traj, cutoff, query_indices)
for i in range(traj.n_frames):
eq(value[i], reference[i])
def test_load():
t = load(get_fn("test_hoomdxml.dcd"), top=get_fn("well-mixed.hoomdxml"))
top = load(get_fn("well-mixed.hoomdxml")).topology
eq(t.topology, top)
eq(top.n_atoms, 6600)
eq(top.n_bonds, 3200)
eq(top.n_chains, 4000)
eq(top.n_chains, top.n_residues) # hoomdxml makes each chain 1 residue
def test_load_unknown_topology():
try:
md.load(get_fn('frame0.dcd'), top=get_fn('frame0.dcd'))
except IOError as e:
# we want to make sure there's a nice error message than includes
# a list of the supported topology formats.
assert all(s in str(e) for s in ('.pdb', '.psf', '.prmtop'))
else:
assert False # fail
def test_chunk0_iterload():
filename = 'frame0.h5'
trj0 = md.load(get_fn(filename))
for trj in md.iterload(get_fn(filename), chunk=0):
pass
eq(trj0.n_frames, trj.n_frames)
def test_read_stride():
with TRRTrajectoryFile(get_fn('frame0.trr')) as f:
xyz, time, step, box, lambd = f.read()
with TRRTrajectoryFile(get_fn('frame0.trr')) as f:
xyz3, time3, step3, box3, lambd3 = f.read(stride=3)
yield lambda: eq(xyz[::3], xyz3)
yield lambda: eq(step[::3], step3)
yield lambda: eq(box[::3], box3)
yield lambda: eq(time[::3], time3)
def test_dihedral_pbc_fails2():
traj_uncorrected = md.load(get_fn("1am7_uncorrected.xtc"), top=get_fn("1am7_protein.pdb"))
traj_corrected = md.load(get_fn("1am7_corrected.xtc"), top=get_fn("1am7_protein.pdb"))
epsilon = 1e-2
ang1 = md.geometry.compute_phi(traj_uncorrected, opt=True, periodic=False)[1]
ang2 = md.geometry.compute_phi(traj_corrected, opt=True, periodic=False)[1]
assert np.max(np.abs(ang1 - ang2)) < epsilon
def test_0():
t1 = load(get_fn('native2.xml'), top=get_fn('native2.pdb'))
t2 = load(get_fn('native2.pdb'))
t1.center_coordinates()
t2.center_coordinates()
yield lambda: eq(t1.xyz, t2.xyz)
yield lambda: eq(t1.unitcell_vectors, t2.unitcell_vectors)
def test_select_pairs_args():
traj = md.load(get_fn('tip3p_300K_1ATM.pdb'))
assert len(traj.top.select_pairs(selection1='name O', selection2='name O')) == 258 * (258 - 1) // 2
assert (eq(traj.top.select_pairs(selection1="(name O) or (name =~ 'H.*')", selection2="(name O) or (name =~ 'H.*')").sort(),
traj.top.select_pairs(selection1='all', selection2='all').sort()))
assert (eq(traj.top.select_pairs(selection1="name O", selection2="name H1").sort(),
traj.top.select_pairs(selection1="name H1", selection2="name O").sort()))
assert (eq(traj.top.select_pairs(selection1=range(traj.n_atoms), selection2="(name O) or (name =~ 'H.*')").sort(),
traj.top.select_pairs(selection1='all', selection2='all').sort()))
def test_shape_when_none():
t = md.load(get_fn('frame0.h5'))
np.hstack((md.compute_phi(t)[1],
md.compute_psi(t)[1],
md.compute_chi1(t)[1],
md.compute_chi2(t)[1],
md.compute_chi3(t)[1],
md.compute_chi1(t)[1],
md.compute_omega(t)[1]))
def test_n_bonds():
t = md.load(get_fn('2EQQ.pdb'))
for atom in t.top.atoms:
if atom.element.symbol == 'H':
assert atom.n_bonds == 1
elif atom.element.symbol == 'C':
assert atom.n_bonds in [3, 4]
elif atom.element.symbol == 'O':
assert atom.n_bonds in [1, 2]
def test_read_after_close():
f = NetCDFTrajectoryFile(get_fn('mdcrd.nc'))
yield lambda: eq(f.n_atoms, 223)
yield lambda: eq(f.n_frames, 101)
f.close()
# should be an ioerror if you read a file that's closed
assert_raises(IOError, lambda: f.read())
def test_load_multiframe():
from mdtraj.pdb.pdbstructure import PdbStructure
with open(get_fn('multiframe.pdb')) as f:
pdb = PdbStructure(f)
yield lambda: eq(len(pdb.models), 2)
yield lambda: eq(len(pdb.models[0].chains), 1)
yield lambda: eq(len(pdb.models[0].chains[0].residues), 3)
yield lambda: eq(ilen(pdb.models[0].iter_atoms()), 22)
yield lambda: eq(len(pdb.models[1].chains), 1)
yield lambda: eq(len(pdb.models[1].chains[0].residues), 3)
yield lambda: eq(ilen(pdb.models[1].iter_atoms()), 22)
t = load(get_fn('multiframe.pdb'))
yield lambda: eq(t.n_frames, 2)
yield lambda: eq(t.n_atoms, 22)
yield lambda: eq(t.xyz[0], t.xyz[1])
def test_binpos():
t = md.load(get_fn('frame0.binpos'), top=nat)
for e in [tmpfns['xtc'], tmpfns['dcd'], tmpfns['binpos'], tmpfns['trr'], tmpfns['h5'], tmpfns['pdb'], tmpfns['pdb.gz'], tmpfns['nc']]:
def f():
t.save(e)
t2 = md.load(e, top=nat)
eq(t.xyz, t2.xyz, err_msg=e)
eq(t.time, t2.time, err_msg=e)
yield f
def test_box_load_save():
t = md.load(get_fn('native2.pdb'))
# these four tempfile have extensions (dcd, xtc, trr, h5) that
# should store the box information. lets make sure than through a load/save
# cycle, the box information is preserved:
for temp_fn in [tmpfns['xtc'], tmpfns['dcd'], tmpfns['trr'], tmpfns['h5']]:
t.save(temp_fn)
if temp_fn.endswith('.h5'):
t2 = md.load(temp_fn)
else:
t2 = md.load(temp_fn, top=get_fn('native.pdb'))
assert t.unitcell_vectors is not None
yield lambda: eq(t.xyz, t2.xyz, decimal=3)
yield lambda: eq(t.unitcell_vectors, t2.unitcell_vectors)
yield lambda: eq(t.unitcell_angles, t2.unitcell_angles)
yield lambda: eq(t.unitcell_lengths, t2.unitcell_lengths)
def test_segment_id():
top = md.load(get_fn('ala_ala_ala.pdb')).topology
assert next(
top.residues
).segment_id == "AAL", "Segment id is not being assigned correctly for ala_ala_ala.psf"
df = top.to_dataframe()[0]
assert len(df["segmentID"] == "AAL") == len(
df
), "Segment id is not being assigned correctly to topology data frame ala_ala_ala.psf"
def test_read_stride_2():
"read xtc with stride with n_frames"
iofile = io.loadh(get_fn('frame0.xtc.h5'), deferred=False)
with XTCTrajectoryFile(fn_xtc) as f:
xyz, time, step, box = f.read(n_frames=1000, stride=3)
yield lambda: eq(xyz, iofile['xyz'][::3])
yield lambda: eq(step, iofile['step'][::3])
yield lambda: eq(box, iofile['box'][::3])
yield lambda: eq(time, iofile['time'][::3])
def test_read_chunk1():
with XTCTrajectoryFile(fn_xtc, 'r', chunk_size_multiplier=0.5) as f:
xyz, time, step, box = f.read()
iofile = io.loadh(get_fn('frame0.xtc.h5'), deferred=False)
yield lambda: eq(xyz, iofile['xyz'])
yield lambda: eq(step, iofile['step'])
yield lambda: eq(box, iofile['box'])
yield lambda: eq(time, iofile['time'])
def test_tell():
with XTCTrajectoryFile(get_fn('frame0.xtc')) as f:
eq(f.tell(), 0)
f.read(101)
eq(f.tell(), 101)
f.read(3)
eq(f.tell(), 104)
def test_topology_pandas():
topology = md.load(get_fn('native.pdb')).topology
atoms, bonds = topology.to_dataframe()
topology2 = md.Topology.from_dataframe(atoms, bonds)
eq(topology, topology2)
topology3 = md.Topology.from_dataframe(
atoms) # Make sure you default arguement of None works, see issue #774
def test_6():
t = md.load(get_fn('alanine-dipeptide-explicit.pdb'))
a = md.compute_dssp(t, simplified=True)
protein_residues = np.array([
set(a.name for a in r.atoms).issuperset(('C', 'N', 'O', 'CA'))
for r in t.topology.residues
])
assert np.unique(a[:, protein_residues]) == "C"
assert np.unique(a[:, np.logical_not(protein_residues)]) == 'NA'
def test_iterload_skip():
files = ['frame0.nc', 'frame0.h5', 'frame0.xtc', 'frame0.trr',
'frame0.dcd', 'frame0.binpos', 'legacy_msmbuilder_trj0.lh5',
'frame0.xyz', 'frame0.lammpstrj']
err_msg = "failed for file %s with chunksize %i and skip %i"
for file in files:
for cs in [0, 1, 11, 100]:
for skip in [0, 1, 20, 101]:
print("testing file %s with skip=%i" % (file, skip))
t_ref = md.load(get_fn(file), top=get_fn('native.pdb'))
t = functools.reduce(lambda a, b: a.join(b),
md.iterload(get_fn(file), skip=skip,
top=get_fn('native.pdb'), chunk=cs))
eq(t_ref.xyz[skip:], t.xyz, err_msg=err_msg % (file, cs, skip))
eq(t_ref.time[skip:], t.time, err_msg=err_msg % (file, cs, skip))
eq(t_ref.topology, t.topology, err_msg=err_msg % (file, cs, skip))
def test_float_atom_indices_exception():
"Is an informative error message given when you supply floats for atom_indices?"
top = md.load(get_fn('native.pdb')).topology
for ext in md._FormatRegistry.loaders.keys():
try:
fn = get_fn('frame0' + ext)
except:
continue
try:
md.load(fn, atom_indices=[0.5, 1.3], top=top)
except ValueError as e:
if six.PY3:
assert e.args[0] == 'indices must be of an integer type. float64 is not an integer type'
else:
assert e.message == 'indices must be of an integer type. float64 is not an integer type'
except Exception as e:
raise
def test_array_vs_matrix():
top = md.load(get_fn('native.pdb')).topology
xyz = np.random.randn(1, 22, 3)
xyz_mat = np.matrix(xyz)
t1 = md.Trajectory(xyz, top)
t2 = md.Trajectory(xyz_mat, top)
eq(t1.xyz, xyz)
eq(t2.xyz, xyz)
def test_load():
tref = md.load(get_fn('4waters.pdb'))
with GroTrajectoryFile(temp, 'w') as f:
f.write(tref.xyz, tref.topology)
t = md.load(temp)
eq(t.xyz, tref.xyz, decimal=3)
eq(list(t.top.atoms), list(tref.top.atoms))
def test_dtr():
t = md.load(get_fn('ala_dipeptide_trj/clickme.dtr'),
top=get_fn('ala_dipeptide.pdb'))
for e in [
tmpfns['xtc'], tmpfns['dcd'], tmpfns['binpos'], tmpfns['trr'],
tmpfns['h5'], tmpfns['pdb'], tmpfns['pdb.gz'], tmpfns['nc']
]:
def f():
t.save(e)
t2 = md.load(e, top=get_fn('ala_dipeptide.pdb'))
# change decimal to 3 since the precision is different in different trajectory
# format
eq(t.xyz, t2.xyz, decimal=3, err_msg=e)
#eq(t.time, t2.time, err_msg=e)
yield f
def test_load_freesolv_gaffmol2_vs_sybylmol2_vs_obabelpdb():
with enter_temp_directory():
tar_filename = "freesolve_v0.3.tar.bz2"
tar = tarfile.open(get_fn(tar_filename), mode="r:bz2")
tar.extractall()
tar.close()
database = pickle.load(open("./v0.3/database.pickle"))
for key in database:
gaff_filename = "./v0.3/mol2files_gaff/%s.mol2" % key
pdb_filename = "./v0.3/mol2files_gaff/%s.pdb" % key
sybyl_filename = "./v0.3/mol2files_sybyl/%s.mol2" % key
cmd = "obabel -imol2 %s -opdb > %s 2>/dev/null" % (sybyl_filename,
pdb_filename)
assert os.system(cmd) == 0
t_pdb = md.load(pdb_filename)
t_gaff = md.load(gaff_filename)
t_sybyl = md.load(sybyl_filename)
eq(t_pdb.n_atoms, t_gaff.n_atoms)
eq(t_pdb.n_atoms, t_sybyl.n_atoms)
eq(t_pdb.n_frames, t_gaff.n_frames)
eq(t_pdb.n_frames, t_gaff.n_frames)
eq(t_pdb.xyz, t_gaff.xyz, decimal=4)
eq(t_pdb.xyz, t_sybyl.xyz, decimal=4)
top_pdb, bonds_pdb = t_pdb.top.to_dataframe()
top_gaff, bonds_gaff = t_gaff.top.to_dataframe()
top_sybyl, bonds_sybyl = t_sybyl.top.to_dataframe()
eq(top_sybyl.name.values, top_pdb.name.values)
# eq(top_gaff.name.values, top_sybyl.name.values) # THEY CAN HAVE DIFFERENT NAMES, so this isn't TRUE!
def make_bonds_comparable(bond_array):
"""Create a bond connectivity matrix from a numpy array of atom pairs. Avoids having to compare the order in which bonds are listed."""
n_bonds = len(bond_array)
data = np.ones(n_bonds)
i = bond_array[:, 0]
j = bond_array[:, 1]
matrix = scipy.sparse.coo_matrix(
(data, (i, j)),
shape=(t_pdb.n_atoms, t_pdb.n_atoms)).toarray()
return matrix + matrix.T # Symmetrize to account for (a ~ b) versus (b ~ a)
bond_matrix_pdb = make_bonds_comparable(bonds_pdb)
bond_matrix_gaff = make_bonds_comparable(bonds_gaff)
bond_matrix_sybyl = make_bonds_comparable(bonds_sybyl)
eq(bond_matrix_pdb, bond_matrix_gaff)
eq(bond_matrix_pdb, bond_matrix_sybyl)
def test_rmsd_ref_ainds():
native = md.load(get_fn('native.pdb'))
t1 = md.load(get_fn('traj.h5'))
atom_indices = np.arange(10)
dist1 = md.rmsd(t1,
native,
atom_indices=atom_indices,
ref_atom_indices=atom_indices)
bad_atom_indices = np.arange(10, 20)
t2 = md.load(get_fn('traj.h5'))
dist2 = md.rmsd(t2,
native,
atom_indices=atom_indices,
ref_atom_indices=bad_atom_indices)
assert np.all(dist2 > dist1)
def test_tell():
with BINPOSTrajectoryFile(get_fn('frame0.binpos')) as f:
eq(f.tell(), 0)
f.read(101)
eq(f.tell(), 101)
f.read(3)
eq(f.tell(), 104)
def test_binpos():
t = md.load(get_fn('frame0.binpos'), top=nat)
for e in [temp1, temp2, temp3, temp4, temp5, temp6, temp7]:
def f():
t.save(e)
t2 = md.load(e, top=nat)
eq(t.xyz, t2.xyz, err_msg=e)
eq(t.time, t2.time, err_msg=e)
yield f
def test_read_2():
with BINPOSTrajectoryFile(fn_binpos, chunk_size_multiplier=10) as f:
xyz = f.read()
with DCDTrajectoryFile(fn_dcd) as f:
xyz2 = f.read()[0]
xyz3 = io.loadh(get_fn('frame0.binpos.h5'), 'xyz')
yield lambda: eq(xyz[1:], xyz2)
yield lambda: eq(xyz, xyz3)
def test_chain():
top = md.load(get_fn('bpti.pdb')).topology
chain = top.chain(0)
assert chain.n_residues == len(list(chain.residues))
atoms = list(chain.atoms)
assert chain.n_atoms == len(atoms)
for i in range(chain.n_atoms):
assert atoms[i] == chain.atom(i)
def test_box():
t = md.load(get_fn('native.pdb'))
yield lambda: eq(t.unitcell_vectors, None)
yield lambda: eq(t.unitcell_lengths, None)
yield lambda: eq(t.unitcell_angles, None)
t.unitcell_vectors = np.array([[1, 0, 0], [0, 1, 0], [0, 0, 1]]).reshape(1, 3, 3)
yield lambda: eq(np.array([1.0, 1.0, 1.0]), t.unitcell_lengths[0])
yield lambda: eq(np.array([90.0, 90.0, 90.0]), t.unitcell_angles[0])
def test_3nch_conect():
# This has conect entries that use all available digits, good failure case.
t1 = load_pdb(get_fn('3nch.pdb.gz'))
top, bonds = t1.top.to_dataframe()
bonds = dict(((a, b), 1) for (a, b) in bonds)
eq(bonds[19782, 19783], 1) # Check that last SO4 molecule has right bonds
eq(bonds[19782, 19784], 1) # Check that last SO4 molecule has right bonds
eq(bonds[19782, 19785], 1) # Check that last SO4 molecule has right bonds
eq(bonds[19782, 19786], 1) # Check that last SO4 molecule has right bonds
def _run_amber_traj(trajname, ext_ref):
# Test triclinic case where simple approach in Tuckerman text does not
# always work
traj = md.load(get_fn(trajname), top=get_fn('test.parm7'))
distopt = md.compute_distances(traj, [[0, 9999]], opt=True)
distslw = md.compute_distances(traj, [[0, 9999]], opt=False)
dispopt = md.compute_displacements(traj, [[0, 9999]], opt=True)
dispslw = md.compute_displacements(traj, [[0, 9999]], opt=False)
eq(distopt, distslw, decimal=5)
eq(dispopt, dispslw, decimal=5)
assert_allclose(distopt.flatten(), ext_ref, atol=2e-5)
# Make sure distances from displacements are the same
eq(np.sqrt((dispopt.squeeze()**2).sum(axis=1)), distopt.squeeze())
eq(np.sqrt((dispslw.squeeze()**2).sum(axis=1)), distslw.squeeze())
eq(dispopt, dispslw, decimal=5)
def test_order_from_traj():
"""Made a perfectly aligned monolayer, should have S2 = 1
"""
traj = md.load(get_fn('alkane-monolayer.pdb'))
indices = [
list(range(1900 + x, 1900 + x + 36)) for x in range(0, 64 * 36, 36)
]
s2 = md.compute_nematic_order(traj, indices=indices)
assert_allclose(1.0, s2)
def test_rmsd_cluster_fixed_k_kcenters():
expected_size = (2, 501)
expected_k = 10
runhelper([
'--trajectories',
get_fn('frame0.xtc'),
get_fn('frame0.xtc'), '--topology',
get_fn('native.pdb'), '--n-clusters',
str(expected_k), '--atoms',
'(name N or name C or name CA or name H or name O)', '--algorithm',
'kcenters'
],
expected_size=expected_size,
algorithm='kcenters',
expected_k=expected_k)
def test_trajectory_rmsd():
t = md.load(get_fn('traj.h5'))
for parallel in [True, False]:
calculated = md.rmsd(t, t, 0, parallel=parallel)
reference = np.zeros(t.n_frames)
for i in range(t.n_frames):
reference[i] = rmsd_qcp(t.xyz[0], t.xyz[i])
eq(calculated, reference, decimal=3)
