def evaluate(self, data_frame, array_datas, classifier_algo, pre_evaluate_results=None):
"""
Train a classifier on multiple arrays
"""
result_object = ResultObject(None, None, None, CommandStatus.Error)
if type(pre_evaluate_results) is not list:
Printer.Print("Pre evaluation results failed! Attach bug report!")
return result_object
win = Window.window()
if data_frame is not None:
result_object = VizContainer.createResult(win, data_frame, ['cval'])
elif array_datas is not None:
result_object = VizContainer.createResult(win, array_datas, ['cval'])
else:
Printer.Print("Provide one of data frame or array datas")
return result_object
cv_output, aux_output = pre_evaluate_results
properties, model_data = aux_output.data
result_object.data = [win, properties, model_data, self.processkFoldCV]
self.printkValueMessage(cv_output.data[0])
self.updateWindow(win, cv_output.data[1], cv_output.data[2], model_data[1], properties["title"])
self.modify_figure.evaluate(result_object)
return result_object
def preEvaluate(self, data_frame, array_datas, classifier_algo):
result_object = ResultObject(None, None, None, CommandStatus.Error)
# Get the data frame
sns.set(color_codes=True)
if data_frame is not None:
df = data_frame.data
cname = data_frame.name
elif array_datas is not None:
command_status, df, kl1, cname = DataGuru.transformArray_to_dataFrame(
array_datas)
if command_status == CommandStatus.Error:
print("Error in getting dataframe!")
result_object.data = "Error in getting dataframe!"
return result_object
else:
result_object.data = "Please provide data frame or arrays to analyze"
return result_object
# Get the ground truth array
if StatContainer.ground_truth is None:
result_object.data = ("Please set a feature vector to ground truth by" +
"typing set ground truth before using this command")
return result_object
else:
df = DataGuru.removeGT(df, StatContainer.ground_truth)
Y = StatContainer.ground_truth.data
# Remove nans:
df, Y = DataGuru.removenan(df, Y)
# Get the classifier model
model = classifier_algo.data[0]
# Code to run the classifier
X = df.values
# Get a standard scaler for the extracted data X
scaler = preprocessing.StandardScaler().fit(X)
X = scaler.transform(X)
properties = self.createDefaultProperties()
properties['title'] = cname
cv_output = self.performCV(properties, X, Y, model)
aux_output = (properties, [X, Y, model])
return [ResultObject(cv_output, None),
ResultObject(aux_output, None)]
def evaluate(self, array_datas):
"""
Create a bar plot between multiple variables
"""
result_object = ResultObject(None, None, None, CommandStatus.Error)
sns.set(color_codes=True)
command_status, df, kl1, cname = DataGuru.transformArray_to_dataFrame(
array_datas)
if command_status == CommandStatus.Error:
return ResultObject(None, None, None, CommandStatus.Error)
if StatContainer.ground_truth is None:
gtVals = np.ones(df.shape[0])
ground_truth = 'ground_truth'
else:
gtVals = StatContainer.filterGroundTruth()
ground_truth = StatContainer.ground_truth.name
if len(gtVals) != df.shape[0]:
print("ground truth does not match with df shape")
print(len(gtVals), df.shape[0])
gtVals = np.ones(df.shape[0])
ground_truth = 'ground_truth'
# Remove nans:
df[ground_truth] = gtVals
df.dropna(inplace=True)
gtVals = df[ground_truth]
uniqVals = StatContainer.isCategorical(gtVals)
binned_ground_truth = False
if uniqVals is None and np.issubdtype(gtVals.dtype, np.number):
# Convert to categorical
df[ground_truth] = pd.cut(gtVals, 10)
binned_ground_truth = True
if binned_ground_truth is True or uniqVals is not None:
gb = df.groupby(ground_truth)
df_mean = gb.mean()
df_errors = gb.std()
if uniqVals is not None and isinstance(uniqVals[0], str):
truncated_uniqVals, _ = StatContainer.removeCommonNames(
df_mean.index)
df_mean.index = truncated_uniqVals
df_errors.index = truncated_uniqVals
# Number of uniq_vals x number of arrs
df_mean_shape = df_mean.shape
if (not binned_ground_truth
and df_mean_shape[1] >= df_mean_shape[0]):
df_mean = df_mean.T
df_errors = df_errors.T
else:
Printer.Print("Ground truth could not be mapped to",
"categorical array\n")
Printer.Print("Please clear or select appropriate ground truth")
return result_object
properties = self.createDefaultProperties()
properties['title'] = cname
if uniqVals is not None and isinstance(uniqVals[0], str):
max_len = max([len(uniqVal) for uniqVal in uniqVals])
else:
max_len = 0
if (binned_ground_truth or
(uniqVals is not None and len(uniqVals) > 5 and max_len > 8)):
properties["horizontal"] = True
if binned_ground_truth:
properties["overwrite_labels"] = True
properties["ylabel"] = StatContainer.ground_truth.name
win = Window.window()
result_object = VizContainer.createResult(win, array_datas, ['bar'])
result_object.data = [
win, properties, [df_mean, df_errors], self.updateFigure
]
self.updateFigure(result_object.data)
self.modify_figure.evaluate(result_object)
return result_object