def test_count_plots(self):
df = self.iris
ax = df.sns.countplot()
self.assertIsInstance(ax, matplotlib.axes.Axes)
self.assertEqual(ax.get_xlabel(), '.target')
self.assertEqual(ax.get_ylabel(), 'count')
tm.close()
return # ToDo: only fails on Travis
ax = df.sns.countplot(df.columns[1])
self.assertIsInstance(ax, matplotlib.axes.Axes)
self.assertEqual(ax.get_xlabel(), df.columns[1])
self.assertEqual(ax.get_ylabel(), 'count')
tm.close()
ax = df.sns.countplot(x=df.columns[1])
self.assertIsInstance(ax, matplotlib.axes.Axes)
self.assertEqual(ax.get_xlabel(), df.columns[1])
self.assertEqual(ax.get_ylabel(), 'count')
tm.close()
ax = df.sns.countplot(y=df.columns[1])
self.assertIsInstance(ax, matplotlib.axes.Axes)
self.assertEqual(ax.get_xlabel(), 'count')
self.assertEqual(ax.get_ylabel(), df.columns[1])
tm.close()
with tm.assertRaises(TypeError):
# can't pass both x and y
df.sns.countplot(x=df.columns[1], y=df.columns[2])
def test_frame_data_proparty_series(self):
df = pdml.ModelFrame({
'A': [1, 2, 3],
'B': [4, 5, 6]
},
target=[7, 8, 9],
index=['a', 'b', 'c'])
df.data = df['A']
exp = pdml.ModelFrame({'A': [1, 2, 3]},
target=[7, 8, 9],
index=['a', 'b', 'c'])
self.assert_frame_equal(df, exp)
df = pdml.ModelFrame({
'A': [1, 2, 3],
'B': [4, 5, 6]
},
target=[7, 8, 9],
index=['a', 'b', 'c'])
df.data = pd.Series([1, 2, 3], name='x', index=['a', 'b', 'c'])
exp = pdml.ModelFrame({'x': [1, 2, 3]},
target=[7, 8, 9],
index=['a', 'b', 'c'])
self.assert_frame_equal(df, exp)
df = pdml.ModelFrame({
'A': [1, 2, 3],
'B': [4, 5, 6]
},
target=[7, 8, 9],
index=['a', 'b', 'c'])
with tm.assertRaises(TypeError):
df.data = [1, 2, 3]
def test_plot():
try:
import matplotlib.pyplot # noqa
except ImportError:
import nose
raise nose.SkipTest()
y_true = [
'rabbit', 'cat', 'rabbit', 'rabbit', 'cat', 'dog', 'dog', 'rabbit',
'rabbit', 'cat', 'dog', 'rabbit'
]
y_pred = [
'cat', 'cat', 'rabbit', 'dog', 'cat', 'rabbit', 'dog', 'cat', 'rabbit',
'cat', 'rabbit', 'rabbit'
]
cm = ConfusionMatrix(y_true, y_pred)
# check plot works
cm.plot()
cm.plot(backend='seaborn')
with tm.assertRaises(ValueError):
cm.plot(backend='xxx')
def test_count_plots(self):
df = self.iris
ax = df.sns.countplot()
self.assertIsInstance(ax, matplotlib.axes.Axes)
self.assertEqual(ax.get_xlabel(), '.target')
self.assertEqual(ax.get_ylabel(), 'count')
tm.close()
return # ToDo: only fails on Travis
ax = df.sns.countplot(df.columns[1])
self.assertIsInstance(ax, matplotlib.axes.Axes)
self.assertEqual(ax.get_xlabel(), df.columns[1])
self.assertEqual(ax.get_ylabel(), 'count')
tm.close()
ax = df.sns.countplot(x=df.columns[1])
self.assertIsInstance(ax, matplotlib.axes.Axes)
self.assertEqual(ax.get_xlabel(), df.columns[1])
self.assertEqual(ax.get_ylabel(), 'count')
tm.close()
ax = df.sns.countplot(y=df.columns[1])
self.assertIsInstance(ax, matplotlib.axes.Axes)
self.assertEqual(ax.get_xlabel(), 'count')
self.assertEqual(ax.get_ylabel(), df.columns[1])
tm.close()
with tm.assertRaises(TypeError):
# can't pass both x and y
df.sns.countplot(x=df.columns[1], y=df.columns[2])
def test_frame_data_proparty_series(self):
df = pdml.ModelFrame({'A': [1, 2, 3],
'B': [4, 5, 6]},
target=[7, 8, 9],
index=['a', 'b', 'c'])
df.data = df['A']
exp = pdml.ModelFrame({'A': [1, 2, 3]},
target=[7, 8, 9],
index=['a', 'b', 'c'])
self.assert_frame_equal(df, exp)
df = pdml.ModelFrame({'A': [1, 2, 3],
'B': [4, 5, 6]},
target=[7, 8, 9],
index=['a', 'b', 'c'])
df.data = pd.Series([1, 2, 3], name='x', index=['a', 'b', 'c'])
exp = pdml.ModelFrame({'x': [1, 2, 3]},
target=[7, 8, 9],
index=['a', 'b', 'c'])
self.assert_frame_equal(df, exp)
df = pdml.ModelFrame({'A': [1, 2, 3],
'B': [4, 5, 6]},
target=[7, 8, 9],
index=['a', 'b', 'c'])
with tm.assertRaises(TypeError):
df.data = [1, 2, 3]
def test_plotting(self):
iris = datasets.load_iris()
df = pdml.ModelFrame(iris)
df.fit(df.svm.SVC())
# raises if df.estimator is not XGBModel
with tm.assertRaises(ValueError):
df.xgb.plot_importance()
with tm.assertRaises(ValueError):
df.xgb.to_graphviz()
with tm.assertRaises(ValueError):
df.xgb.plot_tree()
df.fit(df.xgb.XGBClassifier())
from matplotlib.axes import Axes
from graphviz import Digraph
try:
ax = df.xgb.plot_importance()
except ImportError:
import nose
# matplotlib.use doesn't work on Travis
# PYTHON=3.4 PANDAS=0.17.1 SKLEARN=0.16.1
raise nose.SkipTest()
self.assertIsInstance(ax, Axes)
assert ax.get_title() == 'Feature importance'
assert ax.get_xlabel() == 'F score'
assert ax.get_ylabel() == 'Features'
assert len(ax.patches) == 4
g = df.xgb.to_graphviz(num_trees=0)
self.assertIsInstance(g, Digraph)
ax = df.xgb.plot_tree(num_trees=0)
self.assertIsInstance(ax, Axes)
def test_plotting(self):
import matplotlib
matplotlib.use('Agg')
iris = datasets.load_iris()
df = pdml.ModelFrame(iris)
df.fit(df.svm.SVC())
# raises if df.estimator is not XGBModel
with tm.assertRaises(ValueError):
df.xgb.plot_importance()
with tm.assertRaises(ValueError):
df.xgb.to_graphviz()
with tm.assertRaises(ValueError):
df.xgb.plot_tree()
df.fit(df.xgb.XGBClassifier())
from matplotlib.axes import Axes
from graphviz import Digraph
ax = df.xgb.plot_importance()
assert isinstance(ax, Axes)
assert ax.get_title() == 'Feature importance'
assert ax.get_xlabel() == 'F score'
assert ax.get_ylabel() == 'Features'
assert len(ax.patches) == 4
g = df.xgb.to_graphviz(num_trees=0)
assert isinstance(g, Digraph)
ax = df.xgb.plot_tree(num_trees=0)
assert isinstance(ax, Axes)
def test_plotting(self):
import matplotlib
matplotlib.use('Agg')
iris = datasets.load_iris()
df = pdml.ModelFrame(iris)
df.fit(df.svm.SVC())
# raises if df.estimator is not XGBModel
with tm.assertRaises(ValueError):
df.xgb.plot_importance()
with tm.assertRaises(ValueError):
df.xgb.to_graphviz()
with tm.assertRaises(ValueError):
df.xgb.plot_tree()
df.fit(df.xgb.XGBClassifier())
from matplotlib.axes import Axes
from graphviz import Digraph
ax = df.xgb.plot_importance()
assert isinstance(ax, Axes)
assert ax.get_title() == 'Feature importance'
assert ax.get_xlabel() == 'F score'
assert ax.get_ylabel() == 'Features'
assert len(ax.patches) == 4
g = df.xgb.to_graphviz(num_trees=0)
assert isinstance(g, Digraph)
ax = df.xgb.plot_tree(num_trees=0)
assert isinstance(ax, Axes)
def test_not_supported_estimator(self):
df = pdml.ModelFrame(datasets.load_iris())
df.fit(df.cluster.KMeans(n_clusters=3))
with tm.assertRaises(NotImplementedError):
df.plot_estimator()
def test_no_estimator(self):
df = pdml.ModelFrame(datasets.load_iris())
with tm.assertRaises(ValueError):
df.plot_estimator()
wpercent = (basewidth/float(im_pred.size[0]))
hsize = int((float(im_pred.size[1])*float(wpercent)))
im_pred = im_pred.resize((basewidth,hsize), Image.NEAREST)
im_pred = im_pred
wpercent = (basewidth/float(im_true.size[0]))
hsize = int((float(im_true.size[1])*float(wpercent)))
im_true = im_true.resize((basewidth,hsize), Image.NEAREST)
im_true = im_true
'''
pixels_pred_all += list(im_pred.getdata())
# pix_val_flat_pred = [x for sets in pix_val_pred for x in sets]
pixels_true_all += list(im_true.getdata())
# pix_val_flat_true = [x for sets in pix_val_true for x in sets]
# Print number of pixel count vs the colors
# print im_pred.getcolors()
# pixels_true_all = ['rabbit', 'cat', 'rabbit', 'rabbit', 'cat', 'dog', 'dog', 'rabbit', 'mouse', 'cat', 'dog', 'rabbit']
# pixels_pred_all = ['cat', 'cat', 'rabbit', 'dog', 'cat', 'rabbit', 'dog', 'cat', 'rabbit', 'cat', 'rabbit', 'rabbit']
# Print the confusion matrix
confusion_matrix = ConfusionMatrix(pixels_true_all, pixels_pred_all)
print("\nConfusion matrix:\n%s" % confusion_matrix)
confusion_matrix.plot()
confusion_matrix.plot(backend='seaborn')
with tm.assertRaises(ValueError):
confusion_matrix.plot(backend='xxx')
def test_not_supported_estimator(self):
df = pdml.ModelFrame(datasets.load_iris())
df.fit(df.cluster.KMeans(n_clusters=3))
with tm.assertRaises(NotImplementedError):
df.plot_estimator()
def test_no_estimator(self):
df = pdml.ModelFrame(datasets.load_iris())
with tm.assertRaises(ValueError):
df.plot_estimator()
