def test_simeval(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
simeval_res = read_results(testdata / 'nonmem' / 'qa' /
'simeval_results.json')
cdd_res = read_results(testdata / 'nonmem' / 'qa' / 'cdd_results.json')
calculate_results(orig,
base,
simeval_results=simeval_res,
cdd_results=cdd_res)
def test_tdist(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
td = Model(testdata / 'nonmem' / 'qa' / 'tdist.mod')
res = calculate_results(orig, base, tdist_model=td)
correct = """df,new_sd,old_sd
ETA(1),3.77,0.344951,0.333246
ETA(2),3.77,0.400863,0.448917
"""
correct = pd.read_csv(StringIO(correct), index_col=[0])
pd.testing.assert_frame_equal(res.tdist_parameters, correct)
assert res.dofv['dofv']['parameter_variability', 'tdist',
np.nan] == pytest.approx(730.89472681373070 -
729.45800311609150)
assert res.dofv['df']['parameter_variability', 'tdist', np.nan] == 2
res = calculate_results(orig, base, tdist_model=None)
def test_boxcox(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
bc = Model(testdata / 'nonmem' / 'qa' / 'boxcox.mod')
res = calculate_results(orig, base, boxcox_model=bc)
correct = """lambda,new_sd,old_sd
ETA(1),-1.581460,0.296257,0.333246
ETA(2),0.645817,0.429369,0.448917
"""
correct = pd.read_csv(StringIO(correct), index_col=[0])
pd.testing.assert_frame_equal(res.boxcox_parameters, correct)
assert res.dofv['dofv']['parameter_variability', 'boxcox',
np.nan] == pytest.approx(730.89472681373070 -
721.78812733817688)
assert res.dofv['df']['parameter_variability', 'boxcox', np.nan] == 2
res = calculate_results(orig, base, boxcox_model=None)
assert res.boxcox_parameters is None
def test_fullblock(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
fb = Model(testdata / 'nonmem' / 'qa' / 'fullblock.mod')
res = calculate_results(orig, base, fullblock_model=fb)
correct = """,new,old
"OMEGA(1,1)",0.486600,0.333246
"OMEGA(2,1)",0.846728,NaN
"OMEGA(2,2)",0.423262,0.448917
"""
correct = pd.read_csv(StringIO(correct), index_col=[0])
pd.testing.assert_frame_equal(res.fullblock_parameters, correct)
assert res.dofv['dofv']['parameter_variability', 'fullblock',
np.nan] == pytest.approx(730.89472681373070 -
706.36113798726512)
assert res.dofv['df']['parameter_variability', 'fullblock', np.nan] == 1
res = calculate_results(orig, base, fullblock_model=None)
assert res.fullblock_parameters is None
def test_iov(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
iov = Model(testdata / 'nonmem' / 'qa' / 'iov.mod')
res = calculate_results(orig, base, iov_model=iov)
correct = """new_iiv_sd,orig_iiv_sd,iov_sd
ETA(1),0.259560,0.333246,0.555607
ETA(2),0.071481,0.448917,0.400451
"""
correct = pd.read_csv(StringIO(correct), index_col=[0])
pd.testing.assert_frame_equal(res.iov_parameters, correct)
assert res.dofv['dofv']['parameter_variability', 'iov',
np.nan] == pytest.approx(42.314986)
assert res.dofv['df']['parameter_variability', 'iov', np.nan] == 2
def test_scm(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
scm_res = read_results(testdata / 'nonmem' / 'qa' / 'scm_results.json')
res = calculate_results(orig, base, scm_results=scm_res)
correct = """,,dofv,coeff
ETA(1),APGR,2.48792,-0.033334
ETA(1),WGT,0.48218,0.052342
ETA(2),APGR,0.59036,0.008371
ETA(2),WGT,0.00887,-0.003273
"""
correct = pd.read_csv(StringIO(correct), index_col=[0, 1])
correct.index.set_names(['parameter', 'covariate'], inplace=True)
pd.testing.assert_frame_equal(res.covariate_effects, correct, atol=1e-6)
assert res.dofv['dofv']['covariates', 'ET1APGR-2',
np.nan] == pytest.approx(2.48792)
assert res.dofv['df']['covariates', 'ET1APGR-2', np.nan] == 1
def test_resmod(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
resmod_res = read_results(testdata / 'nonmem' / 'qa' /
'resmod_results.json')
res = calculate_results(orig, base, resmod_idv_results=resmod_res)
assert list(
res.residual_error['additional_parameters']) == [2, 2, 1, 1, 1, 1]
assert list(
res.residual_error['dOFV']) == [13.91, 8.03, 5.53, 3.34, 1.31, 0.03]
assert res.residual_error.index.tolist() == [
(1, 'dtbs'),
(1, 'time_varying'),
(1, 'tdist'),
(1, 'autocorrelation'),
(1, 'IIV_on_RUV'),
(1, 'power'),
]
assert res.dofv['dofv']['residual_error_model', 'dtbs',
1] == pytest.approx(13.91)
def test_add_etas(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
add_etas = Model(testdata / 'nonmem' / 'qa' / 'add_etas_linbase.mod')
res = calculate_results(orig,
base,
add_etas_model=add_etas,
etas_added_to=['CL', 'V'])
correct = """added,new_sd,orig_sd
ETA(1),True,0.338974,0.333246
ETA(2),True,0.449430,0.448917
CL,False,0.010001,NaN
V,False,0.010000,NaN
"""
correct = pd.read_csv(StringIO(correct), index_col=[0])
pd.testing.assert_frame_equal(res.add_etas_parameters, correct, atol=1e-6)
assert res.dofv['dofv']['parameter_variability', 'add_etas',
np.nan] == pytest.approx(730.89472681373070 -
730.84697789365532)
assert res.dofv['df']['parameter_variability', 'add_etas', np.nan] == 2
def test_resmod_dvid(testdata):
orig = Model(testdata / 'nonmem' / 'pheno.mod')
base = Model(testdata / 'nonmem' / 'qa' / 'pheno_linbase.mod')
resmod_res = psn_resmod_results(testdata / 'psn' / 'resmod_dir2')
res = calculate_results(orig, base, resmod_idv_results=resmod_res)
assert res.residual_error.loc[("4", "tdist"), 'dOFV'] == 2.41