def __init__(self,
table: Table,
agg_funs: Iterable[Functions],
row_var: Variable,
col_var: Variable = None,
val_var: Variable = None):
self._group_tables = self.Tables()
self._pivot_tables = self.Tables()
self._table = table
self._row_var = row_var
self._col_var = col_var if col_var else row_var
self.renamed = []
if not table:
return
if not self._row_var.is_primitive():
raise TypeError("Row variable should be DiscreteVariable"
" or ContinuousVariable")
if self._col_var and not self._col_var.is_discrete:
raise TypeError("Column variable should be DiscreteVariable")
self._row_var_col = table.get_column_view(row_var)[0].astype(np.float)
self._col_var_col = table.get_column_view(self._col_var)[0].astype(
np.float)
self._row_var_groups = nanunique(self._row_var_col)
self._col_var_groups = nanunique(self._col_var_col)
self._total_var = DiscreteVariable("Total", values=("total", ))
self._current_agg_functions = sorted(agg_funs)
self._indepen_agg_done = {} # type: Dict[Functions, int]
self._depen_agg_done = {} # type: Dict[Functions, Dict[Variable, int]]
self._initialize(agg_funs, val_var)
def test_nanunique_ignores_nans_in_counts(self, array):
# pylint: disable=bad-whitespace
x = array([[-1., 1., 0., 2., 3., np.nan], [0., 0., 0., 3., 5., np.nan],
[-1., 0., 0., 1., 7., 6.]])
expected = [2, 6, 2, 1, 2, 1, 1, 1]
np.testing.assert_equal(nanunique(x, return_counts=True)[1], expected)
def test_nanunique_ignores_nans_in_values(self, array):
# pylint: disable=bad-whitespace
x = array([[-1., 1., 0., 2., 3., np.nan], [0., 0., 0., 3., 5., np.nan],
[-1., 0., 0., 1., 7., 6.]])
expected = [-1, 0, 1, 2, 3, 5, 6, 7]
np.testing.assert_equal(nanunique(x, return_counts=False), expected)
def remove_unused_values(var, data):
column_data = Table.from_table(
Domain([var]),
data
)
unique = nanunique(column_data.X).astype(int)
if len(unique) == len(var.values):
return var
used_values = [var.values[i] for i in unique]
translation_table = np.array([np.NaN] * len(var.values))
translation_table[unique] = range(len(used_values))
base_value = -1
if 0 >= var.base_value < len(var.values):
base = translation_table[var.base_value]
if np.isfinite(base):
base_value = int(base)
return DiscreteVariable("{}".format(var.name),
values=used_values,
base_value=base_value,
compute_value=Lookup(var, translation_table),
sparse=var.sparse,
)
def test_nanunique_ignores_nans_in_counts(self, array):
# pylint: disable=bad-whitespace
x = array([[-1., 1., 0., 2., 3., np.nan],
[ 0., 0., 0., 3., 5., np.nan],
[-1., 0., 0., 1., 7., 6.]])
expected = [2, 6, 2, 1, 2, 1, 1, 1]
np.testing.assert_equal(nanunique(x, return_counts=True)[1], expected)
def test_nanunique_ignores_nans_in_values(self, array):
# pylint: disable=bad-whitespace
x = array([[-1., 1., 0., 2., 3., np.nan],
[ 0., 0., 0., 3., 5., np.nan],
[-1., 0., 0., 1., 7., 6.]])
expected = [-1, 0, 1, 2, 3, 5, 6, 7]
np.testing.assert_equal(nanunique(x, return_counts=False), expected)
def commit(self):
def send_outputs(pivot_table, filtered_data, grouped_data):
if self.data:
if grouped_data:
grouped_data.name = self.data.name
if pivot_table:
pivot_table.name = self.data.name
if filtered_data:
filtered_data.name = self.data.name
self.Outputs.grouped_data.send(grouped_data)
self.Outputs.pivot_table.send(pivot_table)
self.Outputs.filtered_data.send(filtered_data)
self.Warning.renamed_vars.clear()
self.Warning.too_many_values.clear()
self.Warning.cannot_aggregate.clear()
self.Warning.no_col_feature.clear()
self.table_view.clear()
if self.pivot is None:
if self.data:
if not self.data_has_primitives:
self.Warning.no_variables()
send_outputs(None, None, None)
return
if self.no_col_feature:
self.Warning.no_col_feature()
send_outputs(None, None, None)
return
if self.data:
col_var = self.col_feature or self.row_feature
col = self.data.get_column_view(col_var)[0].astype(float)
if len(nanunique(col)) >= self.MAX_VALUES:
self.table_view.clear()
self.Warning.too_many_values()
send_outputs(None, None, None)
return
self.pivot = Pivot(self.data, self.sel_agg_functions,
self.row_feature, self.col_feature,
self.val_feature)
if self.skipped_aggs:
self.Warning.cannot_aggregate(self.skipped_aggs)
self._update_graph()
send_outputs(self.pivot.pivot_table, self.get_filtered_data(),
self.pivot.group_table)
if self.pivot.renamed:
self.Warning.renamed_vars(self.pivot.renamed)
def remove_unused_values(var, data):
unique = nanunique(data.get_column_view(var)[0].astype(float)).astype(int)
if len(unique) == len(var.values):
return var
used_values = [var.values[i] for i in unique]
translation_table = np.array([np.NaN] * len(var.values))
translation_table[unique] = range(len(used_values))
return DiscreteVariable(var.name,
values=used_values,
sparse=var.sparse,
compute_value=Lookup(var, translation_table))
def remove_unused_values(var, data):
column_data = Table.from_table(Domain([var]), data)
unique = nanunique(column_data.X).astype(int)
if len(unique) == len(var.values):
return var
used_values = [var.values[i] for i in unique]
translation_table = np.array([np.NaN] * len(var.values))
translation_table[unique] = range(len(used_values))
return DiscreteVariable("{}".format(var.name),
values=used_values,
compute_value=Lookup(var, translation_table),
sparse=var.sparse)
def __init__(self,
data,
variable,
parent=None,
height=200,
width=300,
side_padding=5,
top_padding=20,
bottom_padding=0,
bar_spacing=4,
border=0,
border_color=None,
color_attribute=None,
n_bins=10):
super().__init__(parent)
self.height, self.width = height, width
self.padding = side_padding
self.bar_spacing = bar_spacing
self.data = data
self.attribute = data.domain[variable]
self.x = data.get_column_view(self.attribute)[0].astype(np.float64)
self.x_nans = np.isnan(self.x)
self.x = self.x[~self.x_nans]
if self.attribute.is_discrete:
self.n_bins = len(self.attribute.values)
elif self.attribute.is_continuous:
# If the attribute is continuous but contains fewer values than the
# bins, it is better to assign each their own bin. We will require
# at least 2 bins so that the histogram still visually makes sense
# except if there is only a single value, then we use 3 bins for
# symmetry
num_unique = ut.nanunique(self.x).shape[0]
if num_unique == 1:
self.n_bins = 3
else:
self.n_bins = min(max(2, num_unique), n_bins)
# Handle target variable index
self.color_attribute = color_attribute
if self.color_attribute is not None:
self.target_var = data.domain[color_attribute]
self.y = data.get_column_view(color_attribute)[0]
self.y = self.y[~self.x_nans]
if not np.issubdtype(self.y.dtype, np.number):
self.y = self.y.astype(np.float64)
else:
self.target_var, self.y = None, None
# Borders
self.border_color = border_color if border_color is not None else '#000'
if isinstance(border, tuple):
assert len(border) == 4, 'Border tuple must be of size 4.'
self.border = border
else:
self.border = (border, border, border, border)
t, r, b, l = self.border
def _draw_border(point_1, point_2, border_width, parent):
pen = QPen(QColor(self.border_color))
pen.setCosmetic(True)
pen.setWidth(border_width)
line = QGraphicsLineItem(QLineF(point_1, point_2), parent)
line.setPen(pen)
return line
top_left = QPointF(0, 0)
bottom_left = QPointF(0, self.height)
top_right = QPointF(self.width, 0)
bottom_right = QPointF(self.width, self.height)
self.border_top = _draw_border(top_left, top_right, t,
self) if t else None
self.border_bottom = _draw_border(bottom_left, bottom_right, b,
self) if b else None
self.border_left = _draw_border(top_left, bottom_left, l,
self) if l else None
self.border_right = _draw_border(top_right, bottom_right, r,
self) if r else None
# _plot_`dim` accounts for all the paddings and spacings
self._plot_height = self.height
self._plot_height -= top_padding + bottom_padding
self._plot_height -= t / 4 + b / 4
self._plot_width = self.width
self._plot_width -= 2 * side_padding
self._plot_width -= (self.n_bins - 2) * bar_spacing
self._plot_width -= l / 4 + r / 4
self.__layout = QGraphicsLinearLayout(Qt.Horizontal, self)
self.__layout.setContentsMargins(side_padding + r / 2,
top_padding + t / 2,
side_padding + l / 2,
bottom_padding + b / 2)
self.__layout.setSpacing(bar_spacing)
# If the data contains any non-NaN values, we can draw a histogram
if self.x.size > 0:
self.edges, self.distributions = self._histogram()
self._draw_histogram()
def test_nanunique(self):
x = csr_matrix(np.array([0, 1, 1, np.nan]))
np.testing.assert_array_equal(nanunique(x), np.array([0, 1]))
def __init__(self, data, variable, parent=None, height=200,
width=300, side_padding=5, top_padding=20, bar_spacing=4,
border=0, border_color=None, color_attribute=None, n_bins=10):
super().__init__(parent)
self.height, self.width = height, width
self.padding = side_padding
self.bar_spacing = bar_spacing
self.data = data
self.attribute = data.domain[variable]
self.x = data.get_column_view(self.attribute)[0].astype(np.float64)
self.x_nans = np.isnan(self.x)
self.x = self.x[~self.x_nans]
if self.attribute.is_discrete:
self.n_bins = len(self.attribute.values)
elif self.attribute.is_continuous:
# If the attribute is continuous but contains fewer values than the
# bins, it is better to assign each their own bin. We will require
# at least 2 bins so that the histogram still visually makes sense
# except if there is only a single value, then we use 3 bins for
# symmetry
num_unique = ut.nanunique(self.x).shape[0]
if num_unique == 1:
self.n_bins = 3
else:
self.n_bins = min(max(2, num_unique), n_bins)
# Handle target variable index
self.color_attribute = color_attribute
if self.color_attribute is not None:
self.target_var = data.domain[color_attribute]
self.y = data.get_column_view(color_attribute)[0]
self.y = self.y[~self.x_nans]
if not np.issubdtype(self.y.dtype, np.number):
self.y = self.y.astype(np.float64)
else:
self.target_var, self.y = None, None
# Borders
self.border_color = border_color if border_color is not None else '#000'
if isinstance(border, tuple):
assert len(border) == 4, 'Border tuple must be of size 4.'
self.border = border
else:
self.border = (border, border, border, border)
t, r, b, l = self.border
def _draw_border(point_1, point_2, border_width, parent):
pen = QPen(QColor(self.border_color))
pen.setCosmetic(True)
pen.setWidth(border_width)
line = QGraphicsLineItem(QLineF(point_1, point_2), parent)
line.setPen(pen)
return line
top_left = QPointF(0, 0)
bottom_left = QPointF(0, self.height)
top_right = QPointF(self.width, 0)
bottom_right = QPointF(self.width, self.height)
self.border_top = _draw_border(top_left, top_right, t, self) if t else None
self.border_bottom = _draw_border(bottom_left, bottom_right, b, self) if b else None
self.border_left = _draw_border(top_left, bottom_left, l, self) if l else None
self.border_right = _draw_border(top_right, bottom_right, r, self) if r else None
# _plot_`dim` accounts for all the paddings and spacings
self._plot_height = self.height
self._plot_height -= top_padding
self._plot_height -= t / 4 + b / 4
self._plot_width = self.width
self._plot_width -= 2 * side_padding
self._plot_width -= (self.n_bins - 2) * bar_spacing
self._plot_width -= l / 4 + r / 4
self.__layout = QGraphicsLinearLayout(Qt.Horizontal, self)
self.__layout.setContentsMargins(
side_padding + r / 2,
top_padding + t / 2,
side_padding + l / 2,
b / 2
)
self.__layout.setSpacing(bar_spacing)
# If the data contains any non-NaN values, we can draw a histogram
if self.x.size > 0:
self.edges, self.distributions = self._histogram()
self._draw_histogram()
def test_nanunique(self):
x = csr_matrix(np.array([0, 1, 1, np.nan]))
np.testing.assert_array_equal(
nanunique(x),
np.array([0, 1])
)
def remove_unused_values(var, data):
unique = nanunique(data.get_column_view(var)[0].astype(float)).astype(int)
if len(unique) == len(var.values):
return var
used_values = [var.values[i] for i in unique]
return DiscreteVariable(var.name, values=used_values, sparse=var.sparse)
