def ri2py_listvector(obj):
if 'data.frame' in obj.rclass:
items = zip(obj.do_slot('names'), (numpy2ri.ri2py(x) for x in obj))
res = PandasDataFrame.from_items(items)
else:
res = numpy2ri.ri2py(obj)
return res
def ri2py_listvector(obj):
if 'data.frame' in obj.rclass:
items = zip(obj.do_slot('names'), (numpy2ri.ri2py(x) for x in obj))
res = PandasDataFrame.from_items(items)
else:
res = numpy2ri.ri2py(obj)
return res
def ri2py_dataframe(obj):
# use the numpy converter
recarray = numpy2ri.ri2py(obj)
try:
idx = numpy2ri.ri2py(obj.do_slot('row.names'))
except LookupError as le:
idx = None
res = PandasDataFrame.from_records(recarray, index=idx)
return res
def ri2py_dataframe(obj):
# use the numpy converter
recarray = numpy2ri.ri2py(obj)
try:
idx = numpy2ri.ri2py(obj.do_slot('row.names'))
except LookupError as le:
idx = None
res = PandasDataFrame.from_records(recarray,
index=idx)
return res
def pval_grenander_fit(pvals):
fdrtool = importr('fdrtool')
pval_vec = ro.FloatVector(pvals)
pval_ecdf = ecdf_pkg.ecdf_pval(pval_vec)
gre_fit = fdrtool.grenander(pval_ecdf, type='decreasing')
x_knots = rpyn.ri2py(gre_fit.rx2('x.knots'))
f_knots = rpyn.ri2py(gre_fit.rx2('f.knots'))
if len(f_knots) == 0:
x_knots = np.array([0, 1])
f_knots = np.array([1, 1])
assert len(f_knots) == len(x_knots)
return x_knots, f_knots
def predict(self, X, eval_MSE=False):
"""
X should be a dataframe
"""
X = self._check_X(X)
_ = self.pkg.predict_randomForest(self.rf, X, predict_all=eval_MSE)
if eval_MSE:
y_hat = numpy2ri.ri2py(_[0])
mse = std(numpy2ri.ri2py(_[1]), axis=1, ddof=1) ** 2.
return y_hat, mse
else:
return numpy2ri.ri2py(_)
def r_spectrum(signal, r_periodogram):
filtered_signal = detrend(signal, type='linear')
filtered_signal = recursive_filter_function(filtered_signal)
filtered_signal = demean(filtered_signal)
res = r_periodogram(numpy2ri.numpy2ri(filtered_signal),
pad=0.3,
tap=0.3,
span=2,
plot=False,
detrend=True,
demean=True)
freq = numpy2ri.ri2py(res.rx2('freq'))
psd = numpy2ri.ri2py(res.rx2('spec'))
return 1 / freq, psd
def _infer_network(self, data):
"""
Infer the network.
Args:
data (pd.DataFrame): data to be used for the inference.
"""
# activate implicit conversion from pandas to R objects
pandas2ri.activate()
genie3 = importr('GENIE3')
importr('foreach')
importr('doParallel')
# transform pandas dataframe into GENIE3 input format
globalenv['r_matrix'] = numpy2ri.py2ri(data.T.values)
globalenv['r_rows'] = data.columns
globalenv['r_cols'] = data.index
r('''
rownames(r_matrix) <- c(r_rows)
colnames(r_matrix) <- c(r_cols)
''')
expr_matrix = globalenv['r_matrix']
# run GENIE3
values = numpy2ri.ri2py(
genie3.GENIE3(expr_matrix, self.regulators, self.targets,
self.tree_method, self.k, self.n_trees, self.n_cores,
self.verbose))
weight_matrix = pd.DataFrame(values,
columns=data.columns,
index=data.columns)
self.graph = Graph(adjacency=weight_matrix)
logger.debug('inferred with {}'.format(self.method))
def _pr_rc_curve_r(observations, predictions, FDRth=0.05):
"""
:param observations: known truth set
:param predictions: all data
:param FDRth:
:return:
"""
obs_rtbl = numpy2ri.py2ri(observations)
prd_rtbl = numpy2ri.py2ri(predictions)
curve_prm = {'scores.class0': prd_rtbl, 'weights.class0': obs_rtbl, 'curve': True, 'sorted': True}
prc = PRROC.pr_curve(**curve_prm)
auc = prc.rx2('auc.integral')[0]
curve = numpy2ri.ri2py(prc.rx2('curve'))
cols = ['recall', 'precision', 'threshold']
df = pd.DataFrame(curve, columns=cols)
FDR5percTh = - (df[df.precision >= (1 - FDRth)])['threshold'].min()
if not np.isnan(FDR5percTh):
index_min = min(df[df.precision >= (1 - FDRth)].index.tolist())
else:
index_min = 0
SENS = df.at[index_min, 'recall']
threshold = -FDR5percTh
return df, auc, SENS, FDR5percTh
def _roc_curve_r(observations, predictions, FDRth=0.05):
"""
:param observations: known truth set
:param predictions: all data
:param FDRth:
:return:
"""
obs_rtbl = numpy2ri.py2ri(observations)
prd_rtbl = numpy2ri.py2ri(predictions)
roc_prm = {'direction': '>'}
RES = pROC.roc(obs_rtbl, prd_rtbl, **roc_prm)
auc = pandas2ri.ri2py(RES.rx2('auc'))[0]
columns = ['threshold', 'ppv', 'sensitivity', 'specificity']
coor_prm = {'ret': r.c('threshold', 'ppv', 'sensitivity', 'specificity')}
COORS = pROC.coords(RES, 'all', **coor_prm)
cords = numpy2ri.ri2py(COORS)
df = pd.DataFrame(cords.T, columns=columns)
FDR5percTh = (df[df.ppv >= (1 - FDRth)])['threshold'].max()
if not np.isnan(FDR5percTh):
index_min = min(df[df.threshold <= FDR5percTh].index.tolist())
else:
index_min = 0
threshold = df.at[index_min, 'threshold']
SENS = df.at[index_min, 'sensitivity']
SPEC = df.at[index_min, 'specificity']
return df, auc, SENS, FDR5percTh
def _exec_r_module(self):
try:
import rpy2.robjects
from rpy2.robjects import numpy2ri
from rpy2.robjects import pandas2ri
from rpy2.robjects.packages import importr
except ImportError:
raise ImportError('R module cannot be run, because '
'"rpy2" package is not installed.')
module_name = os.path.splitext(os.path.basename(self.source_file))[0]
logger.debug('import module "%s" from source file: %s',
self.source_file)
logger.debug('source module: "%s"', self.source_file)
rpy2.robjects.r('source("{0}")'.format(self.source_file))
module = rpy2.robjects.r[module_name]
version = module.get('VERSION')[0]
if version != self.handles.version:
raise PipelineRunError(
'Version of source and handles is not the same.')
func = module.get('main')
numpy2ri.activate() # enables use of numpy arrays
pandas2ri.activate() # enable use of pandas data frames
kwargs = self.keyword_arguments
logger.debug('evaluate main() function with INPUTS: "%s"',
'", "'.join(kwargs.keys()))
# R doesn't have unsigned integer types
for k, v in kwargs.iteritems():
if isinstance(v, np.ndarray):
if v.dtype == np.uint16 or v.dtype == np.uint8:
logging.debug(
'module "%s" input argument "%s": '
'convert unsigned integer data type to integer',
self.name, k)
kwargs[k] = v.astype(int)
elif isinstance(v, pd.DataFrame):
# TODO: We may have to translate pandas data frames explicitly
# into the R equivalent.
# pandas2ri.py2ri(v)
kwargs[k] = v
args = rpy2.robjects.ListVector({k: v for k, v in kwargs.iteritems()})
base = importr('base')
r_out = base.do_call(func, args)
for handle in self.handles.output:
# NOTE: R functions are supposed to return a list. Therefore
# we can extract the output argument using rx2().
# The R equivalent would be indexing the list with "[[]]".
if isinstance(r_out.rx2(handle.name),
rpy2.robjects.vectors.DataFrame):
handle.value = pandas2ri.ri2py(r_out.rx2(handle.name))
# handle.value = pd.DataFrame(r_out.rx2(handle.name))
else:
# NOTE: R doesn't have an unsigned integer data type.
# So we cast to uint16.
handle.value = numpy2ri.ri2py(r_out.rx2(handle.name)).astype(
np.uint16)
# handle.value = np.array(r_out.rx2(handle.name), np.uint16)
return self.handles.output
def ri2py_intvector(obj):
# special case for factors
if 'factor' in obj.rclass:
res = pandas.Categorical.from_codes(numpy.asarray(obj) - 1,
categories=obj.do_slot('levels'),
ordered='ordered' in obj.rclass)
else:
res = numpy2ri.ri2py(obj)
return res
def ri2py_intvector(obj):
# special case for factors
if 'factor' in obj.rclass:
res = pandas.Categorical.from_codes(numpy.asarray(obj) - 1,
categories = obj.do_slot('levels'),
ordered = 'ordered' in obj.rclass)
else:
res = numpy2ri.ri2py(obj)
return res
def fit(self, X, Y):
numpy2ri.activate()
rPMA = importr('PMA')
typex, typez = _check_penalty_type(self.penalty)
X, x_mean, x_std = _center_data(X)
Y, y_mean, y_std = _center_data(Y)
if self.n_component is None:
self.n_component = np.min([X.shape[1], Y.shape[1]])
out = rPMA.CCA(x=X, z=Y, K=self.n_component, \
niter=self.n_iter, standardize=False, \
typex=typex, typez=typez, \
penaltyx=self.C[0], penaltyz=self.C[1], \
trace=False)
self.u = numpy2ri.ri2py(out[0])
self.v = numpy2ri.ri2py(out[1])
self._x_score, self._y_score = self.transform(X, Y)
self._cancorr = _cancorr(X, Y, self.u, self.v)
numpy2ri.deactivate()
return self
def get_distance(phyloseq_d, dist_method):
R_phyloseq = importr('phyloseq')
R_base = importr('base')
distances = R_phyloseq.distance(phyloseq_d, method=dist_method)
distance_mat = R_base.as_matrix(distances)
distance_df = pd.DataFrame(numpy2ri.ri2py(distance_mat),
index=pandas2ri.ri2py(R_phyloseq.sample_names(phyloseq_d)),
columns=pandas2ri.ri2py(R_phyloseq.sample_names(phyloseq_d))
)
return distance_df
def get_wunifrac_distance(phyloseq_d):
R_phyloseq = importr('phyloseq')
R_base = importr('base')
distances = R_phyloseq.UniFrac(phyloseq_d, weighted=True, normalized=True, fast=True, parallel=False)
distance_mat = R_base.as_matrix(distances)
distance_df = pd.DataFrame(numpy2ri.ri2py(distance_mat),
index=pandas2ri.ri2py(R_phyloseq.sample_names(phyloseq_d)),
columns=pandas2ri.ri2py(R_phyloseq.sample_names(phyloseq_d))
)
return distance_df
def processData(data,prompts):
expandedMatrix = buildMatrices(data)
competencies = {}
eigenRatios = {}
meanCompetencies = {}
negativeCompetencies = {}
for dataSet, matrix in expandedMatrix.items():
sys.err.println("Processing " + dataSet)
prompt = prompts['Answer.prompt_' + dataSet.split('_')[1]]
# Calculate competencies for this scale
factors = CCT(matrix, prompt)
competencies[dataSet] = np2ri.ri2py((factors.rx('loadings')[0])).flatten()
eigen = np2ri.ri2py((factors.rx('values')[0])).flatten()
eigenRatios[dataSet] = eigen[0]/eigen[1]
meanCompetencies[dataSet] = competencies[dataSet].mean()
negativeCompetencies[dataSet] = (competencies[dataSet] < 0).sum() * 1.0/len(competencies[dataSet])
return (competencies,eigenRatios,meanCompetencies,negativeCompetencies)
def fit(self, X, y):
self.X = self._check_X(X)
y = array(y).astype(float)
self.columns = numpy2ri.ri2py(self.X.colnames)
n_sample, self.n_feature = self.X.nrow, self.X.ncol
if isinstance(self.param['mtry'], basestring):
p = self.n_feature
self.param['mtry'] = eval(self.param['mtry'])
self.rf = self.pkg.randomForest(x=self.X, y=y, **self.param)
return self
def processData(dataFrame):
originalCompetencies = {}
competencies = {}
eigenRatios = {}
matrices = buildMatrices(dataFrame)
for dataSet, matrix in matrices.items():
factors = CCT(matrix)
results = np2ri.ri2py((factors.rx('loadings')[0])).flatten()
originalCompetencies[dataSet] = results
matrix = reverseNegatives(matrix,results)
factors = CCT(matrix)
results = np2ri.ri2py((factors.rx('loadings')[0])).flatten()
competencies[dataSet] = results
eigen = np2ri.ri2py((factors.rx('values')[0])).flatten()
eigenRatios[dataSet] = eigen
return (competencies, eigenRatios, originalCompetencies)
def fit_betabinom_ab(n, k, weights=None):
assert np.all((k >= 0) & (k <= n))
n_r = ro.FloatVector(n)
k_r = ro.FloatVector(k)
if weights is not None:
assert len(weights) == len(k)
assert len(weights) == len(n)
weights_r = ro.FloatVector(weights)
result_r = bbfit.fit_betabinom_w(n_r, k_r, weights_r)
else:
result_r = bbfit.fit_betabinom(n_r, k_r)
result = rpyn.ri2py(result_r)
log_a, log_b = result.flatten()
return np.exp(log_a), np.exp(log_b)
def ri2py_floatvector(obj):
# special case for POSIXct date objects
if 'POSIXct' in obj.rclass:
tzone_name = obj.do_slot('tzone')[0]
if tzone_name == '':
# R is implicitly using the local timezone, while Python time libraries
# will assume UTC.
tzone = get_timezone()
else:
tzone = pytz.timezone(tzone_name)
foo = (tzone.localize(datetime.fromtimestamp(x)) for x in obj)
res = pandas.to_datetime(tuple(foo))
else:
res = numpy2ri.ri2py(obj)
return res
def run_auto_arima(window, horizons, args):
numpy2ri.activate()
forecast = importr('forecast')
_, y_train, _, _ = window
extra_param = args['extras']
fit = forecast.auto_arima(y_train,
stationary=extra_param.get('stationary', False),
seasonal=extra_param.get('seasonal', True),
stepwise=extra_param.get('stepwise', False))
order = list(fit[6])
arima_str = 'ARIMA({0},{5},{1}) ({2},{6},{3})_{4}'
print arima_str.format(*order)
result = forecast.forecast(fit, h=max_h, level=np.array([0.95]))
y_pred = numpy2ri.ri2py(result[3])
y_arima = y_pred[np.array(horizons) - 1]
return y_arima
def output_pheno_beta(self, meffil=False):
"""Get pheno and beta dataframe objects stored as attributes for input to MethylationArray object.
Parameters
----------
meffil
True if ran meffil pipeline."""
self.pheno_py=pandas2ri.ri2py(robjects.r['as'](self.pheno,'data.frame'))
if not meffil:
self.beta_py=pd.DataFrame(pandas2ri.ri2py(self.beta_final),index=numpy2ri.ri2py(robjects.r("featureNames")(self.RSet)),columns=numpy2ri.ri2py(robjects.r("sampleNames")(self.RSet))).transpose()
self.pheno_py['Sample_Name']=np.vectorize(lambda x: x.split('/')[-1])(self.pheno_py['Basename'])
self.pheno_py = self.pheno_py.set_index('Sample_Name').loc[self.beta_py.index,:]
else:
self.beta_py=pd.DataFrame(pandas2ri.ri2py(self.beta_final),index=robjects.r("rownames")(self.beta_final),columns=robjects.r("colnames")(self.beta_final)).transpose()
print(self.beta_py)
print(self.beta_py.index)
print(self.pheno_py)
self.pheno_py = self.pheno_py.set_index('Sample_Name').loc[self.beta_py.index,:]
def convert_r_response_to_python(response,
convert_scalars=True,
to_list=False,
native_to_json=False):
"""Converts rpy2 types to Python objects.
The output is either a Python built-in, NumPy or Pandas object.
Note:
Rpy2 returns scalars as vectors. These are converted back to scalars with this conversion, which
may not be desirable. This behavior can be switched of.
Args:
response (rpy2 type): object from the R session to be converted
convert_scalars (bool): convert vectors with length 1 to scalars
Returns:
Converted object
"""
original_response_type = type(response)
new_res = response
if issubclass(original_response_type, rpy2.robjects.vectors.Array):
new_res = numpy2ri.ri2py(response)
elif issubclass(original_response_type,
rpy2.robjects.vectors.DataFrame):
new_res = pandas2ri.ri2py(response)
if native_to_json:
json_res = new_res.to_json(orient='records')
json_res = json_res.replace('-2147483648', 'null')
new_res = json.loads(json_res)
elif issubclass(original_response_type, rpy2.robjects.vectors.Vector):
if convert_scalars and len(response) == 1:
new_res = response[0]
elif to_list:
new_res = list(response)
return new_res
def testAtomicVectorToNumpy(self):
v = robjects.vectors.IntVector((1, 2, 3))
a = rpyn.ri2py(v)
self.assertTrue(isinstance(a, numpy.ndarray))
self.assertEqual(1, v[0])
def ri2py_listvector(obj):
if 'data.frame' in obj.rclass:
res = ri2py.registry[DataFrame](obj)
else:
res = numpy2ri.ri2py(obj)
return res
def ri2py_listvector(obj):
if 'data.frame' in obj.rclass:
res = ri2py.registry[DataFrame](obj)
else:
res = numpy2ri.ri2py(obj)
return res
def ri2py_vector(obj):
res = numpy2ri.ri2py(obj)
return res
def testAtomicVectorToNumpy(self):
v = robjects.vectors.IntVector((1,2,3))
a = rpyn.ri2py(v)
self.assertTrue(isinstance(a, numpy.ndarray))
self.assertEqual(1, v[0])
from rpy2.robjects.numpy2ri import numpy2rpy as numpy2ri
from rpy2.robjects.numpy2ri import rpy2py as ri2py
#from rpy2.robjects.numpy2ri import numpy2ri, ri2py
from rpy2.robjects.packages import importr, data
import openturns as ot
# Require
stats = importr("stats")
faraway = importr("faraway")
savings_data = data(faraway).fetch("savings")["savings"]
#data(faraway).fetch("savings")["savings"]
# Model 1 : 2 param, non intercept
sr = ri2py(savings_data)["sr"]
r.assign('sr', numpy2ri(sr))
pop15 = ri2py(savings_data)["pop15"]
r.assign('pop15', numpy2ri(pop15))
pop75 = ri2py(savings_data)["pop75"]
r.assign('pop75', numpy2ri(pop75))
formula = Formula('sr ~ pop75 + pop15 - 1')
fit = stats.lm(formula)
summary = stats.summary_lm(fit)
"""
list(summary.names) provides
['call', 'terms', 'residuals', 'coefficients', 'aliased', 'sigma', 'df', 'r.squared', 'adj.r.squared', 'fstatistic', 'cov.unscaled']
"""
r2 = summary[7][0]
ar2 = summary[8][0]
def ri2py_vector(obj):
res = numpy2ri.ri2py(obj)
return res
def ri2py_dataframe(obj):
# use the numpy converter
recarray = numpy2ri.ri2py(obj)
res = PandasDataFrame.from_records(recarray)
return res
def risk_assessment_plot(X_binarized_norm, y, ROC_models):
"""
Function that generates risk assessment plot. Takes as the input matrix of metrics.
:return: plot
"""
# for recurrence problem
lr_probabilities, rf_probabilities, mlp_probabilities, xgb_probabilities = me.best_models_ROC_curves(
X_binarized_norm,
y, ROC_models,
False)
# for recurrence problem
output_list1 = me.risk_assessment_data(lr_probabilities, rf_probabilities, y)
output_list2 = me.risk_assessment_data(mlp_probabilities, xgb_probabilities, y)
import rpy2.robjects.numpy2ri as rpyn
rpyn.activate()
[xrf_sens, yrf_sens, xrf_spec, yrf_spec, xlr_sens, ylr_sens, xlr_spec, ylr_spec] = rpyn.ri2py(output_list1)
[xmlp_sens, ymlp_sens, xmlp_spec, ymlp_spec, xxgb_sens, yxgb_sens, xxgb_spec, yxgb_spec] = rpyn.ri2py(
output_list2)
plt.figure()
plt.rcParams["figure.figsize"] = (10, 6)
# for recurrence problem
plt.plot(xlr_sens, ylr_sens, color='b',
label='Events LR',
lw=2, alpha=.8)
plt.plot(xlr_spec, ylr_spec, color='b', linestyle='--',
label='Non-events LR',
lw=2, alpha=.8)
plt.plot(xrf_sens, yrf_sens, color='g',
label='Events RF',
lw=2, alpha=.8)
plt.plot(xrf_spec, yrf_spec, color='g', linestyle='--',
label=r'Non-events RF',
lw=2, alpha=.8)
plt.plot(xmlp_sens, ymlp_sens, color='k',
label=r'Events MLP',
lw=2, alpha=.8)
plt.plot(xmlp_spec, ymlp_spec, color='k', linestyle='--',
label=r'Non-events MLP',
lw=2, alpha=.8)
plt.plot(xxgb_sens, yxgb_sens, color='magenta',
label=r'Events XGBoost',
lw=2, alpha=.8)
plt.plot(xxgb_spec, yxgb_spec, color='magenta', linestyle='--',
label=r'Non-events XGBoost',
lw=2, alpha=.8)
plt.xlabel('Calculated risk')
plt.ylabel('Sensitivity, 1-Specificity')
plt.legend(loc="best")
plt.show()
def ri2py_vector(obj):
# use the numpy converter first
res = numpy2ri.ri2py(obj)
if isinstance(res, recarray):
res = PandasDataFrame.from_records(res)
return res
def _exec_r_module(self):
try:
import rpy2.robjects
from rpy2.robjects import numpy2ri
from rpy2.robjects import pandas2ri
from rpy2.robjects.packages import importr
except ImportError:
raise ImportError(
'R module cannot be run, because '
'"rpy2" package is not installed.'
)
module_name = os.path.splitext(os.path.basename(self.source_file))[0]
logger.debug(
'import module "%s" from source file: %s', self.source_file
)
logger.debug('source module: "%s"', self.source_file)
rpy2.robjects.r('source("{0}")'.format(self.source_file))
module = rpy2.robjects.r[module_name]
version = module.get('VERSION')[0]
if version != self.handles.version:
raise PipelineRunError(
'Version of source and handles is not the same.'
)
func = module.get('main')
numpy2ri.activate() # enables use of numpy arrays
pandas2ri.activate() # enable use of pandas data frames
kwargs = self.keyword_arguments
logger.debug(
'evaluate main() function with INPUTS: "%s"',
'", "'.join(kwargs.keys())
)
# R doesn't have unsigned integer types
for k, v in kwargs.iteritems():
if isinstance(v, np.ndarray):
if v.dtype == np.uint16 or v.dtype == np.uint8:
logging.debug(
'module "%s" input argument "%s": '
'convert unsigned integer data type to integer',
self.name, k
)
kwargs[k] = v.astype(int)
elif isinstance(v, pd.DataFrame):
# TODO: We may have to translate pandas data frames explicitly
# into the R equivalent.
# pandas2ri.py2ri(v)
kwargs[k] = v
args = rpy2.robjects.ListVector({k: v for k, v in kwargs.iteritems()})
base = importr('base')
r_out = base.do_call(func, args)
for handle in self.handles.output:
# NOTE: R functions are supposed to return a list. Therefore
# we can extract the output argument using rx2().
# The R equivalent would be indexing the list with "[[]]".
if isinstance(r_out.rx2(handle.name), rpy2.robjects.vectors.DataFrame):
handle.value = pandas2ri.ri2py(r_out.rx2(handle.name))
# handle.value = pd.DataFrame(r_out.rx2(handle.name))
else:
# NOTE: R doesn't have an unsigned integer data type.
# So we cast to uint16.
handle.value = numpy2ri.ri2py(r_out.rx2(handle.name)).astype(
np.uint16
)
# handle.value = np.array(r_out.rx2(handle.name), np.uint16)
return self.handles.output
def ri2py_dataframe(obj):
# use the numpy converter
recarray = numpy2ri.ri2py(obj)
res = PandasDataFrame.from_records(recarray)
return res
def approx_qei(X, model, maxima, x_pending = None,
num_sampled_points = 5,
num_batches_eval = 400,
strategy_batch_selection = 'random'):
"""
Use Mickael Binois approximation to qEI function
This is used to calculate qEI score for batches
Parameters
----------
* `X` [array-like, shape=(n_samples, n_features)]:
Values where the acquisition function should be computed.
* `model` [sklearn estimator that implements predict with ``return_std``]:
The fit estimator that approximates the function through the
method ``predict``.
It should have a ``return_std`` parameter that returns the standard
deviation.
* `maxima`[float, default 0]:
Previous minimum value which we would like to improve upon.
* `num_sampled_points` [int, default 5]:
Number of points to sample in parallel
* `num_batches_eval`[int, default 400]
Number of batches to evaluate
* `strategy_batch_selection` [default 'random']:
Strategy for selection of elements in batches
Returns
-------
* `values`: [array-like, shape=(len(num_batches_eval),)]:
qEI values for each batch
"""
# Converting x_pending list to numpy array
if x_pending is not None:
x_pending = np.array(x_pending)
batches = []
cc_vec = np.zeros(num_batches_eval)
# Batch preparation
for i in range(num_batches_eval):
if strategy_batch_selection == 'random':
rel_ind = np.random.choice(X.shape[0], num_sampled_points, replace=False)
b = X[rel_ind,:]
if x_pending is not None:
b = np.vstack([x_pending, b])
else:
raise ValueError ("No such sampling strategy exists ..")
batches.append(b)
mean, covar = model.predict(b, return_cov=True)
#print ('covar')
#print(np.isnan(covar).any())
cc = qEI.qEI_approx(mean, covar, maxima)
cc_num = rpyn.ri2py(cc)
cc_vec[i] = cc_num
#print(cc_vec)
max_qEI_val = np.nanmax(cc_vec)
print('max')
print(max_qEI_val)
max_qEI_val_ind = np.argmax(cc_vec)
best_batch = batches[max_qEI_val_ind]
return best_batch, batches, cc_vec, max_qEI_val
################################################################################
################################################################################
# code with sample from R objects : it is easier if we want to compare directly in R
print('')
print('-' * 10)
print('')
rnorm = r['rnorm']
set_seed = r['set.seed']
## generate regressor and dependent variable
set_seed(0)
X1 = rnorm(n_points)
X2 = rnorm(n_points)
err1 = rnorm(n_points)
output = numpy2ri(ri2py(X1) + ri2py(X1) - ri2py(X2) + ri2py(err1) + 1)
r.assign('X1', X1)
r.assign('X2', X2)
r.assign('output', output)
## perform Durbin-Watson test
formula = Formula('output ~ X1 + X2')
dw_test_1 = lmtest.dwtest(formula, alternative=hypothesis['R'], exact=True)
print('Result from R :')
print('p-value :', dw_test_1[3][0])
print('Alternative hypothesis : ', dw_test_1[2][0])
print('dw stat : ', dw_test_1[0][0])
print('')
# transformation into openturns objects
firstSample = ot.Sample(np.column_stack((ri2py(X1), ri2py(X2))))
