def test_individual_ofv(pheno, pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1, model=pheno)
iofv = res.individual_ofv
assert len(iofv) == 59
assert pytest.approx(iofv[1], 1e-15) == 5.9473520242962552
assert pytest.approx(iofv[57], 1e-15) == 5.6639479151436394
assert res.plot_iofv_vs_iofv(res)
def test_eta_shrinkage(pheno, pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1, model=pheno)
shrinkage = res.eta_shrinkage()
assert len(shrinkage) == 2
assert pytest.approx(shrinkage['ETA(1)'], 0.0001) == 7.2048E+01 / 100
assert pytest.approx(shrinkage['ETA(2)'], 0.0001) == 2.4030E+01 / 100
shrinkage = res.eta_shrinkage(sd=True)
assert len(shrinkage) == 2
assert pytest.approx(shrinkage['ETA(1)'], 0.0001) == 4.7130E+01 / 100
assert pytest.approx(shrinkage['ETA(2)'], 0.0001) == 1.2839E+01 / 100
def test_individual_estimates(pheno, pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1, model=pheno)
ie = res.individual_estimates
assert len(ie) == 59
assert pytest.approx(ie['ETA(1)'][1], 1e-15) == -0.0438608
assert pytest.approx(ie['ETA(2)'][1], 1e-15) == 0.00543031
assert pytest.approx(ie['ETA(1)'][28], 1e-15) == 7.75957e-04
assert pytest.approx(ie['ETA(2)'][28], 1e-15) == 8.32311E-02
def test_individual_estimates_covariance(pheno, pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1, model=pheno)
cov = res.individual_estimates_covariance
assert len(cov) == 59
names = ['ETA(1)', 'ETA(2)']
correct = pd.DataFrame([[2.48833E-02, -2.99920E-03], [-2.99920E-03, 7.15713E-03]],
index=names, columns=names)
pd.testing.assert_frame_equal(cov[1], correct)
correct2 = pd.DataFrame([[2.93487E-02, -1.95747E-04], [-1.95747E-04, 8.94118E-03]],
index=names, columns=names)
pd.testing.assert_frame_equal(cov[43], correct2)
def test_tool_files(pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst)
names = [str(p.name) for p in res.tool_files]
assert names == [
'pheno_real.lst',
'pheno_real.ext',
'pheno_real.cov',
'pheno_real.cor',
'pheno_real.coi',
'pheno_real.phi',
]
def modelfit_results(self):
if self._modelfit_results is None:
if self.source.path.is_file():
ext_path = self.source.path.with_suffix('.ext')
if ext_path.exists() and stat(ext_path).st_size > 0:
self._modelfit_results = NONMEMChainedModelfitResults(
ext_path, model=self)
return self._modelfit_results
else:
return None
else:
return self._modelfit_results
def test_condition_number(testdata, pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst)
assert res.condition_number == pytest.approx(4.39152)
maxeval3 = Model(testdata / 'nonmem' / 'modelfit_results' / 'onePROB' /
'oneEST' / 'noSIM' / 'maxeval3.mod')
assert maxeval3.modelfit_results.condition_number == 4.77532e06
maxeval0 = Model(testdata / 'nonmem' / 'modelfit_results' / 'onePROB' /
'oneEST' / 'noSIM' / 'maxeval0.mod')
assert maxeval0.modelfit_results.condition_number is None
def modelfit_results(self):
try:
return self._modelfit_results
except AttributeError:
None
if self.source.path.is_file():
lst_path = self.source.path.with_suffix('.lst')
if lst_path.exists():
num_est = len(self.control_stream.get_records('ESTIMATION'))
self._modelfit_results = NONMEMChainedModelfitResults(
lst_path, num_est, model=self)
return self._modelfit_results
else:
return None
def test_individual_shrinkage(pheno, pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1, model=pheno)
ishr = res.individual_shrinkage
assert len(ishr) == 59
assert pytest.approx(ishr['ETA(1)'][1], 1e-15) == 0.84778949807160287
def test_ofv(pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1)
assert res.ofv == 586.27605628188053
def test_standard_errors(pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1)
ses = res.standard_errors
assert len(ses) == 6
assert pytest.approx(ses['THETA(1)'], 1e-13) == 2.10036E-04
def test_correlation(pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1)
corr = res.correlation_matrix
assert len(corr) == 6
assert corr.loc['THETA(1)', 'THETA(1)'] == 1.0
assert pytest.approx(corr.loc['OMEGA(2,2)', 'THETA(2)'], 1e-13) == 3.56662E-01
def test_information(pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1)
m = res.information_matrix
assert len(m) == 6
assert pytest.approx(m.loc['THETA(1)', 'THETA(1)'], 1e-13) == 2.99556E+07
assert pytest.approx(m.loc['OMEGA(2,2)', 'THETA(2)'], 1e-13) == -2.80082E+03
def test_covariance(pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1)
cov = res.covariance_matrix
assert len(cov) == 6
assert pytest.approx(cov.loc['THETA(1)', 'THETA(1)'], 1e-13) == 4.41151E-08
assert pytest.approx(cov.loc['OMEGA(2,2)', 'THETA(2)'], 1e-13) == 7.17184E-05
def test_parameter_estimates(pheno_lst):
res = NONMEMChainedModelfitResults(pheno_lst, 1)
pe = res.parameter_estimates
assert len(pe) == 6
assert pe['THETA(1)'] == 4.69555e-3
assert pe['OMEGA(2,2)'] == 2.7906e-2