def gmx_water_particle_without_energy():
dataset = alchemtest.gmx.load_water_particle_without_energy()
u_nk = pd.concat([
gmx.extract_u_nk(filename, T=300)
for filename in dataset['data']['AllStates']
])
return u_nk
def gmx_benzene_coul_u_nk():
dataset = alchemtest.gmx.load_benzene()
u_nk = pd.concat([
gmx.extract_u_nk(filename, T=300)
for filename in dataset['data']['Coulomb']
])
return u_nk
def gmx_expanded_ensemble_case_3():
dataset = alchemtest.gmx.load_expanded_ensemble_case_3()
u_nk = pd.concat([
gmx.extract_u_nk(filename, T=300)
for filename in dataset['data']['AllStates']
])
return u_nk
def gmx_benzene_vdw_u_nk():
dataset = alchemtest.gmx.load_benzene()
u_nk = alchemlyb.concat([
gmx.extract_u_nk(filename, T=300)
for filename in dataset['data']['VDW']
])
return u_nk
def estimaters():
bz = load_benzene().data
dHdl_coul = alchemlyb.concat(
[extract_dHdl(xvg, T=300) for xvg in bz['Coulomb']])
ti = TI().fit(dHdl_coul)
u_nk_coul = alchemlyb.concat(
[extract_u_nk(xvg, T=300) for xvg in bz['Coulomb']])
mbar = MBAR().fit(u_nk_coul)
return ti, mbar
def test_u_nk():
"""Test that u_nk has the correct form when extracted from files.
"""
dataset = load_benzene()
for leg in dataset['data']:
for filename in dataset['data'][leg]:
u_nk = extract_u_nk(filename, T=300)
assert u_nk.index.names == ['time', 'fep-lambda']
if leg == 'Coulomb':
assert u_nk.shape == (4001, 5)
elif leg == 'VDW':
assert u_nk.shape == (4001, 16)
def test_u_nk_with_potential_energy():
"""Test that the reduced potential is calculated correctly when the potential energy is given.
"""
# Load dataset
dataset = load_water_particle_with_potential_energy()
# Check if the sum of values on the diagonal has the correct value
assert_almost_equal(_diag_sum(dataset), 16674040406778.867, decimal=2)
# Check one specific value in the dataframe
assert_almost_equal(extract_u_nk(dataset['data']['AllStates'][0],
T=300).iloc[0][0],
-15656.557252200757,
decimal=6)
def test_u_nk_with_total_energy():
"""Test that the reduced potential is calculated correctly when the total energy is given.
"""
# Load dataset
dataset = load_water_particle_with_total_energy()
# Check if the sum of values on the diagonal has the correct value
assert_almost_equal(_diag_sum(dataset), 47611374980.34574, decimal=4)
# Check one specific value in the dataframe
assert_almost_equal(extract_u_nk(dataset['data']['AllStates'][0],
T=300).iloc[0][0],
-11211.577658852531,
decimal=6)
def test_u_nk_case2():
"""Test that u_nk has the correct form when extracted from expanded ensemble files (case 2).
"""
dataset = load_expanded_ensemble_case_2()
for leg in dataset['data']:
for filename in dataset['data'][leg]:
u_nk = extract_u_nk(filename, T=300, filter=False)
assert u_nk.index.names == [
'time', 'fep-lambda', 'coul-lambda', 'vdw-lambda',
'restraint-lambda'
]
assert u_nk.shape == (25001, 28)
def test_u_nk_without_energy():
"""Test that the reduced potential is calculated correctly when no energy is given.
"""
# Load dataset
dataset = load_water_particle_without_energy()
# Check if the sum of values on the diagonal has the correct value
assert_almost_equal(_diag_sum(dataset), 20572986867158.184, decimal=2)
# Check one specific value in the dataframe
assert_almost_equal(extract_u_nk(dataset['data']['AllStates'][0],
T=300).iloc[0][0],
0.0,
decimal=6)
def _diag_sum(dataset):
"""Calculate the sum of diagonal elements (i, i)
"""
# Initialize the sum variable
ds = 0.0
for leg in dataset['data']:
for filename in dataset['data'][leg]:
u_nk = extract_u_nk(filename, T=300)
# Calculate the sum of diagonal elements:
for i in range(len(dataset['data'][leg])):
ds += u_nk.iloc[i][i]
return ds
def test_plot_mbar_omatrix():
'''Just test if the plot runs'''
bz = load_benzene().data
u_nk_coul = pd.concat([extract_u_nk(xvg, T=300) for xvg in bz['Coulomb']])
mbar_coul = MBAR()
mbar_coul.fit(u_nk_coul)
assert isinstance(plot_mbar_overlap_matrix(mbar_coul.overlap_matrix),
matplotlib.axes.Axes)
assert isinstance(
plot_mbar_overlap_matrix(mbar_coul.overlap_matrix, [
1,
]), matplotlib.axes.Axes)
# Bump up coverage
overlap_maxtrix = mbar_coul.overlap_matrix
overlap_maxtrix[0, 0] = 0.0025
overlap_maxtrix[-1, -1] = 0.9975
assert isinstance(plot_mbar_overlap_matrix(overlap_maxtrix),
matplotlib.axes.Axes)
def test_plot_convergence():
bz = load_benzene().data
data_list = [extract_u_nk(xvg, T=300) for xvg in bz['Coulomb']]
forward = []
forward_error = []
backward = []
backward_error = []
num_points = 10
for i in range(1, num_points + 1):
# Do the forward
slice = int(len(data_list[0]) / num_points * i)
u_nk_coul = alchemlyb.concat([data[:slice] for data in data_list])
estimate = MBAR().fit(u_nk_coul)
forward.append(estimate.delta_f_.iloc[0, -1])
forward_error.append(estimate.d_delta_f_.iloc[0, -1])
# Do the backward
u_nk_coul = alchemlyb.concat([data[-slice:] for data in data_list])
estimate = MBAR().fit(u_nk_coul)
backward.append(estimate.delta_f_.iloc[0, -1])
backward_error.append(estimate.d_delta_f_.iloc[0, -1])
ax = plot_convergence(forward, forward_error, backward, backward_error)
assert isinstance(ax, matplotlib.axes.Axes)
plt.close(ax.figure)
def u_nk():
bz = load_benzene().data
u_nk_coul = alchemlyb.concat(
[extract_u_nk(xvg, T=300) for xvg in bz['Coulomb']])
u_nk_coul.attrs = extract_u_nk(load_benzene().data['Coulomb'][0], T=300).attrs
return u_nk_coul
def n_uk_list(self):
n_uk_list = [
gmx.extract_u_nk(dhdl, T=300)
for dhdl in load_ABFE()['data']['complex']
]
return n_uk_list
def gmx_benzene_u_nk_full():
dataset = alchemtest.gmx.load_benzene()
return pd.concat(
[gmx.extract_u_nk(i, T=300) for i in dataset['data']['Coulomb']])
def gmx_benzene_u_nk():
dataset = alchemtest.gmx.load_benzene()
return gmx.extract_u_nk(dataset['data']['Coulomb'][0], T=300)
def test_plot_dF_state():
'''Just test if the plot runs'''
bz = load_benzene().data
u_nk_coul = alchemlyb.concat(
[extract_u_nk(xvg, T=300) for xvg in bz['Coulomb']])
dHdl_coul = alchemlyb.concat(
[extract_dHdl(xvg, T=300) for xvg in bz['Coulomb']])
u_nk_vdw = alchemlyb.concat(
[extract_u_nk(xvg, T=300) for xvg in bz['VDW']])
dHdl_vdw = alchemlyb.concat(
[extract_dHdl(xvg, T=300) for xvg in bz['VDW']])
ti_coul = TI().fit(dHdl_coul)
ti_vdw = TI().fit(dHdl_vdw)
bar_coul = BAR().fit(u_nk_coul)
bar_vdw = BAR().fit(u_nk_vdw)
mbar_coul = MBAR().fit(u_nk_coul)
mbar_vdw = MBAR().fit(u_nk_vdw)
dhdl_data = [
(ti_coul, ti_vdw),
(bar_coul, bar_vdw),
(mbar_coul, mbar_vdw),
]
fig = plot_dF_state(dhdl_data, orientation='portrait')
assert isinstance(fig, matplotlib.figure.Figure)
plt.close(fig)
fig = plot_dF_state(dhdl_data, orientation='landscape')
assert isinstance(fig, matplotlib.figure.Figure)
plt.close(fig)
fig = plot_dF_state(dhdl_data, labels=['MBAR', 'TI', 'BAR'])
assert isinstance(fig, matplotlib.figure.Figure)
plt.close(fig)
with pytest.raises(ValueError):
fig = plot_dF_state(dhdl_data, labels=[
'MBAR',
'TI',
])
fig = plot_dF_state(dhdl_data, colors=['#C45AEC', '#33CC33', '#F87431'])
assert isinstance(fig, matplotlib.figure.Figure)
plt.close(fig)
with pytest.raises(ValueError):
fig = plot_dF_state(dhdl_data, colors=['#C45AEC', '#33CC33'])
with pytest.raises(ValueError):
fig = plot_dF_state(dhdl_data, orientation='xxx')
fig = plot_dF_state(ti_coul, orientation='landscape')
assert isinstance(fig, matplotlib.figure.Figure)
plt.close(fig)
fig = plot_dF_state(ti_coul, orientation='portrait')
assert isinstance(fig, matplotlib.figure.Figure)
plt.close(fig)
fig = plot_dF_state([ti_coul, bar_coul])
assert isinstance(fig, matplotlib.figure.Figure)
plt.close(fig)
fig = plot_dF_state([(ti_coul, ti_vdw)])
assert isinstance(fig, matplotlib.figure.Figure)
plt.close(fig)
def gmx_ABFE_u_nk():
dataset = alchemtest.gmx.load_ABFE()
return gmx.extract_u_nk(dataset['data']['complex'][-1], T=300)
def extract_data(self, dir, temp, dt):
# extract and subsample dHdl using equilibrium_detection
dHdl_state = [] # dHdl_state is for collecting data for a single state
u_nk_state = [] # u_nk_state is for collecting data fro a single state
if os.path.isfile('temporary.xvg') is True:
os.system("rm temporary.xvg")
files = glob.glob(os.path.join(dir, '*dhdl.xvg*'))
files = natsort.natsorted(files, reverse=False)
file_idx = -1
n = 0 # counter for the number of files of a specific state
self.n_state = 0 # counter for the number of states
for i in track(files):
n += 1
file_idx += 1
logger(f"Parsing {files[file_idx]} and collecting data ...")
os.system(f"head -n-1 {i} > temporary.xvg"
) # delete the last line in case it is incomplete
dHdl_state.append(extract_dHdl('temporary.xvg', T=temp))
u_nk_state.append(extract_u_nk('temporary.xvg', T=temp))
if n > 1: # for discard the overlapped time frames of the previous file
upper_t = dHdl_state[-2].iloc[
dHdl_state[-2].shape[0] -
1].name[0] # the last time frame of file n
lower_t = dHdl_state[-1].iloc[0].name[
0] # the first time frame of file n + 1
# upper_t and lower_t should be the same for both dHdl and u_nk
if lower_t != 0: # in case that the file n+1 is the first file of the next replica
n_discard = int(
(upper_t - lower_t) / dt +
1) # number of data frames to discard in file n
dHdl_state[-2] = dHdl_state[-2].iloc[:-n_discard]
u_nk_state[-2] = u_nk_state[-2].iloc[:-n_discard]
else: # lower_t == 0 means that we have gathered dHdl for the previous state
self.n_state += 1
dHdl_data = pd.concat(dHdl_state[:-1])
u_nk_data = pd.concat(u_nk_state[:-1])
dHdl.append(
equilibrium_detection(dHdl_data, dHdl_data.iloc[:, 0]))
dHdl_state = [dHdl_state[-1]]
logger(
f'Subsampling dHdl data of the {ordinal(self.n_state)} state ...'
)
u_nk.append(
equilibrium_detection(u_nk_data, u_nk_data.iloc[:, 0]))
u_nk_state = [u_nk_state[-1]]
logger(
f'Subsampling u_nk data of the {ordinal(self.n_state)} state ...'
)
n = 1 # now there is only one file loaded in dHdl_state/u_nk_state
# dealing with the last state with equilibrium_detection
self.n_state += 1
dHdl_data = pd.concat(dHdl_state)
u_nk_data = pd.concat(u_nk_state)
return dHdl_data, u_nk_data
def test_extract_u_nk_unit():
'''Test if extract_u_nk assign the attr correctly'''
dataset = load_benzene()
u_nk = extract_u_nk(dataset['data']['Coulomb'][0], 310)
assert u_nk.attrs['temperature'] == 310
assert u_nk.attrs['energy_unit'] == 'kT'
def gmx_benzene():
dataset = load_benzene()
return [gmx.extract_dHdl(dhdl, T=300) for dhdl in dataset['data']['Coulomb']], \
[gmx.extract_u_nk(dhdl, T=300) for dhdl in dataset['data']['Coulomb']]
