def test_edge_cases_visualization_best_fit_lines(self):
"""`visualization.Visualization.best_fit_lines`: Edge Case Validator.
Tests the behavior of `Visualization.best_fit_lines` with edge cases.
Raises:
Exception: If at least one `Exception` raised is not of the expected
kind.
"""
x = _compose(list, _np.random.uniform)(0.0,
100.0,
size=self.data_shape[0])
"""list of float: Random x-values."""
values = {
name: _compose(list, _np.random.uniform)(0.0, 100.0,
size=self.data_shape[0]) \
for name in ["observations", "predictions"]
}
""":obj:`list of float`: Contains all y-values."""
v = Visualization("Title", (1, 1))
"""Visualization: Plotter instance."""
with self.assertRaises(TypeError):
# Empty `x`-values.
v._best_fit_lines([], values, 0)
with self.assertRaises(ValueError):
# Empty `y`-values.
v._best_fit_lines(x, {}, 0)
with self.assertRaises(IndexError):
# Suplot index out of range.
v._best_fit_lines(x, values, 1)
v.close()
def test_random_visualization_best_fit_lines(self):
"""`visualization.Visualization.best_fit_lines`: Randomized Validator.
Tests the behavior of `Visualization.best_fit_lines` by feeding it
randomly generated arguments.
Raises:
AssertionError: If `Visualization.best_fit_lines` needs debugging.
"""
for i in range(self.n_tests):
v = Visualization("Title", (3, 3))
"""Visualization: Plotter instance."""
for i in range(9):
x = _compose(list, _np.random.uniform)(0.0,
100.0,
size=self.data_shape[0])
"""list of float: Random x-values."""
values = {
name: _compose(list,
_np.random.uniform)(0.0, 100.0,
size=self.data_shape[0]) \
for name in ["observations", "predictions"]
}
""":obj:`list of float`: Contains all y-values."""
lines = v._best_fit_lines(x, values, i)
"""list of matplotlib.lines.Line2D: Test input."""
# Best-fit lines should be a list of Line2D instances.
self.assertIsInstance(lines, list)
map(_appendargs(self.assertIsInstance, _Line2D), lines)
unique_x = _compose(list, _np.unique)(x)
"""list of float: X-values with duplicates removed."""
for line in lines:
x_data = line.get_xdata()
"""list of float: X-values in actual best-fit line."""
y_data = line.get_ydata()
"""list of float: Y-values in actual best-fit line."""
# Number of x- and y-values in all lines should match number
# of unique values in `x`.
self.assertEqual(*map(_np.linalg.norm, [x_data, unique_x]))
self.assertEqual(*map(len, [y_data, unique_x]))
m = (y_data[1] - y_data[0]) / (x_data[1] - x_data[0])
"""float: Best-fit line gradient."""
b = y_data[0] - m * x_data[0]
"""float: Y-intercept."""
computed_y_vals = map(lambda val: m * val + b, x_data)
"""list of float: Y-values computed analytically."""
# All lines should be rewritable in linear terms.
self.assertAlmostEqual(
*map(_np.linalg.norm, [computed_y_vals, y_data]))
v.close()
def test_invalid_args_visualization_best_fit_lines(self):
"""`visualization.Visualization.best_fit_lines`: Argument Validator.
Tests the behavior of `Visualization.best_fit_lines` with invalid
argument counts and values.
Raises:
Exception: If at least one `Exception` raised is not of the expected
kind.
"""
x = _compose(list, _np.random.uniform)(0.0,
100.0,
size=self.data_shape[0])
"""list of float: Random x-values."""
values = {
name: _compose(list, _np.random.uniform)(0.0, 100.0,
size=self.data_shape[0]) \
for name in ["observations", "predictions"]
}
""":obj:`list of float`: Contains all y-values."""
v = Visualization("Title", (1, 1))
"""Visualization: Plotter instance."""
with self.assertRaises(TypeError):
# No arguments.
v._best_fit_lines()
with self.assertRaises(TypeError):
# Only one argument.
v._best_fit_lines(x)
with self.assertRaises(TypeError):
# Only two arguments.
v._best_fit_lines(x, values)
with self.assertRaises(TypeError):
# Too many arguments.
v._best_fit_lines(x, values, 0, "extra")
with self.assertRaises(TypeError):
# With keyword.
v._best_fit_lines(x, values, 0, key="value")
with self.assertRaises(TypeError):
# `None` instead of list `x`.
v._best_fit_lines(None, values, 0)
with self.assertRaises(TypeError):
# np.matrix instead of list `x`.
v._best_fit_lines(_np.matrix(x), values, 0)
with self.assertRaises(TypeError):
# `None` instead of dict of lists `values`.
v._best_fit_lines(x, None, 0)
with self.assertRaises(TypeError):
value_list = {k: _np.matrix(val) for k, val in values.iteritems()}
""":obj:`np.matrix`: Lists in `values` mapped to matrices."""
# Dict of np.matrix instead of dict of lists `values`.
v._best_fit_lines(x, value_list, 0)
with self.assertRaises(TypeError):
# List instead of dict of lists `values`.
v._best_fit_lines(x, x, 0)
with self.assertRaises(TypeError):
# `None` instead of int `subplot`.
v._best_fit_lines(x, values, None)
with self.assertRaises(TypeError):
# Float instead of int `subplot`.
v._best_fit_lines(x, values, 5.6)