def testEvaluateTwoFormulas(self):
if IGNORE_TEST:
return
self.column_a.setFormula(
SECOND_VALID_FORMULA) # Make A a formula column
evaluator = TableEvaluator(self.table)
error = evaluator.evaluate(user_directory=TEST_DIR)
self.assertIsNone(error)
def export(self, **kwargs): """ Exports the table to a python program :return: error - string from the file export """ table_evaluator = TableEvaluator(self) error = table_evaluator.export(**kwargs) return error
def testEvaluateError(self):
if IGNORE_TEST:
return
column_invalid_formula = cl.Column(COLUMN_INVALID_FORMULA)
column_invalid_formula.setFormula(INVALID_FORMULA)
self.table.addColumn(column_invalid_formula)
evaluator = TableEvaluator(self.table)
error = evaluator.evaluate(user_directory=TEST_DIR)
self.assertIsNotNone(error)
def evaluate(self, user_directory=None):
"""
Evaluates formulas in the table
:param user_directory: full directory path where user modules
are placed
:return: error from table evaluation or None
"""
evaluator = TableEvaluator(self)
error = evaluator.evaluate(user_directory=user_directory)
return error
def main():
cat_cols_test = ['firstname', 'lastname', 'address']
real, fake = load_data('data/test-data-real.csv',
'data/test-data-fake.csv')
table_evaluator_test = TableEvaluator(real, fake, cat_cols=cat_cols_test)
for category in cat_cols_test:
print(category + " Statistics")
table_evaluator_test.evaluate(target_col=category)
print("-------------------------------")
def testRunningSheetWithExport(self):
if IGNORE_TEST:
return
"""
Imports two sheets. The first is the Michaelis-Menten sheet. The second
is a sheet with S, V, V_MAX, K_M. This test:
1. Exports the first sheet
2. Changes the V_MAX formula in the second sheet to use the export
3. Evaluates the second sheet
"""
def isEqualColumns(table1, table2, colnm):
"""
Checks that the tables have the same value for the column name
:param Table table1:
:param Table table2:
:param str colnm:
"""
value1 = table1.columnFromName(colnm).getCells()[0]
value2 = table2.columnFromName(colnm).getCells()[0]
self.assertEqual(value1, value2)
function_name = "MM"
# Read the tables
table1 = api_util.readObjectFromFile(join(TEST_DIR, FILE1))
table2 = api_util.readObjectFromFile(join(TEST_DIR, FILE1))
# Set V_MAX in the second table to a dummy value
column_v_max_second = table2.columnFromName("V_MAX")
column_v_max_second.setFormula("range(10)")
evaluator = TableEvaluator(table2)
evaluator.evaluate(user_directory=TEST_DIR)
# Use the exported first table for the values of V_MAX, K_M
table1.export(function_name=function_name,
inputs=["S", "V"],
outputs=["V_MAX", "K_M"],
user_directory=TEST_DIR)
formula = "V_MAX, K_M = %s(S, V)" % function_name
column_v_max_second.setFormula(formula)
evaluator = TableEvaluator(table2)
evaluator.evaluate(user_directory=TEST_DIR)
isEqualColumns(table1, table2, "V_MAX")
isEqualColumns(table1, table2, "K_M")
def setUp(self):
"""
Creates the following structure:
DUMMY_TABLE
row:
DUMMY1:
A:
B:
C:
VALID_FORMULA:
"""
self.table = createTable(TABLE_NAME)
self._addColumn(COLUMN1, cells=COLUMN1_CELLS)
self.column_a = self._addColumn(COLUMN2, cells=COLUMN2_CELLS)
self.column_b = self._addColumn(COLUMNB, cells=COLUMNB_CELLS)
self.column_c = self._addColumn(COLUMNC, cells=COLUMNC_CELLS)
self.column_valid_formula = self._addColumn(COLUMN_VALID_FORMULA,
formula=VALID_FORMULA)
api_util.writeObjectToFile(self.table)
self.evaluator = TableEvaluator(self.table)
ax.set_zlim(data['Test'][z].min() - 10,
data['Test'][z].max() + 10)
if i == 4:
ax.set_xlabel(x)
ax.set_ylabel(y)
ax.set_zlabel(z)
ax.set_title(dataset, fontsize=14)
plt.tight_layout()
plt.show()
dataset = 'MIAMI'
cols = ['age', 'workclass', 'marital.status']
table_evaluator = TableEvaluator(data['Test'][cols],
data[dataset][cols])
table_evaluator.plot_distributions()
#==========================================
# Correlation matrices
#==========================================
w_ratios = [1] * len(plot_datasets) + [0.08]
fig, axs = plt.subplots(1, len(plot_datasets) + 1, \
gridspec_kw={'width_ratios':w_ratios}, figsize = (15,10))
axs[0].get_shared_y_axes().join(*axs[1:-1])
#fig.suptitle(design + ' ' + ' (preds' + str(idx) + ')',\
#fontsize=16)
for idx, mod in enumerate(ord_modalities):
train_new['education'] = np.where(
train_new['education'] == str(float(idx)), mod,
train_new['education'])
# Keep only the women that have more than 60 years in the test
if design in ['Absent', 'Unb']:
test = test[(test['age'] >= 60) & (test['sex'] == 'Female')]
# Store the predictions
train_new.to_csv('pseudo_adult/' + design + '/' + 'preds' + str(filenum) + '.csv',\
index = False)
acceptance_rate = pd.DataFrame(acceptance_rate)
acceptance_rate.to_csv('pseudo_adult/acceptance_rate.csv')
acceptance_rate[['Unbalanced', 'Absent']].astype(float).boxplot()
plt.title('Acceptance rate of MIAMI in the Absent and Unbalanced designs')
plt.ylabel('Acceptance rate')
plt.xlabel('Design')
# Visualise the predictions
# Use table_evaluator
plt.plot([0])
plt.title('Design:' + design)
plt.show()
cat_features = (~(pd.Series(var_distrib) != 'continuous')).to_list()
table_evaluator = TableEvaluator(test, train_new)
table_evaluator.visual_evaluation()
def testConstructor(self):
if IGNORE_TEST:
return
evaluator = TableEvaluator(self.table)
self.assertEqual(evaluator._table.getName(is_global_name=False),
TABLE_NAME)
