Skip to content

Commit

Permalink
Distribution: Friendlier bin boundaries
Browse files Browse the repository at this point in the history
  • Loading branch information
@janezd
janezd committed Jun 22, 2019
1 parent 11f2442 commit 6d08361
Showing 1 changed file with 91 additions and 40 deletions.
131 changes: 91 additions & 40 deletions Orange/widgets/visualize/owdistributions.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@
import pyqtgraph as pg

from AnyQt.QtWidgets import \
QSizePolicy, QListView, QToolTip, QGraphicsItem, QGraphicsRectItem
QSizePolicy, QListView, QGraphicsItem, QGraphicsRectItem
from AnyQt.QtGui import QColor, QPen, QBrush, QPainter, QPalette, QPolygonF
from AnyQt.QtCore import Qt, QRectF, QPointF, pyqtSignal as Signal
from scipy.stats import norm, rayleigh, beta, gamma, pareto, expon
Expand All @@ -22,7 +22,6 @@

from Orange.widgets.visualize.owscatterplotgraph import \
LegendItem as SPGLegendItem
from Orange.widgets.visualize.utils.plotutils import HelpEventDelegate


class ScatterPlotItem(pg.ScatterPlotItem):
Expand Down Expand Up @@ -233,8 +232,6 @@ class Outputs:

graph_name = "plot"

Bins = [2, 3, 4, 5, 8, 10, 12, 15, 20, 30, 50]

Fitters = (
("None", None, (), ()),
("Normal", norm, ("loc", "scale"), ("μ", "σ²")),
Expand All @@ -255,9 +252,11 @@ def __init__(self):
self.selection = set()
self.bar_items = []
self.curve_descriptions = None
self.possible_bins = []

self.last_click_idx = None
self.drag_operation = self.DragNone
self._user_var_bins = {}

gui.listView(
self.controlArea, self, "var", box="Variable",
Expand All @@ -271,8 +270,9 @@ def __init__(self):
slider = gui.hSlider(
box, self, "number_of_bins",
label="Number of bins", orientation=Qt.Horizontal,
minValue=0, maxValue=len(self.Bins) - 1, createLabel=False,
callback=self._on_bins_changed)
minValue=0, maxValue=max(1, len(self.possible_bins) - 1),
createLabel=False, callback=self._on_bins_changed)
slider.sliderReleased.connect(self._on_bin_slider_released)
gui.comboBox(
box, self, "fitted_distribution", label="Fitted distribution",
orientation=Qt.Horizontal, items=(name[0] for name in self.Fitters),
Expand All @@ -298,8 +298,6 @@ def __init__(self):
gui.auto_commit(
self.controlArea, self, "auto_apply", "&Apply", commit=self.apply)

slider.sliderReleased.connect(self.apply)

self.plotview = DistributionWidget(background=None)
self.plotview.item_clicked.connect(self._on_item_clicked)
self.plotview.mouse_released.connect(self._on_end_selecting)
Expand Down Expand Up @@ -367,12 +365,15 @@ def set_data(self, data):
if domain is not None and domain.has_discrete_class:
self.cvar = domain.class_var
self.selection.clear()
self._user_var_bins.clear()
self.openContext(domain)
self._recompute_binnings()
self._replot()
self.apply()

def _on_var_changed(self):
self.selection.clear()
self._recompute_binnings()
self._replot()
self.apply()

Expand All @@ -382,6 +383,11 @@ def _on_bins_changed(self):
# this is triggered when dragging, so don't call apply here;
# apply is called on sliderReleased

def _on_bin_slider_released(self):
self._user_var_bins[self.var] = self.number_of_bins
self.apply()


def _replot(self):
self._clear_plot()
self._set_axis_names()
Expand Down Expand Up @@ -487,7 +493,7 @@ def _disc_split_plot(self):
def _cont_plot(self):
self.ploti.getAxis("bottom").setTicks(None)
data = self.valid_data
y, x = np.histogram(data, bins=self.Bins[self.number_of_bins])
y, x = np.histogram(data, bins=self._get_bins())
total = len(data)
colors = [QColor(0, 128, 255)]
if self.fitted_distribution:
Expand All @@ -510,7 +516,7 @@ def _cont_plot(self):
def _cont_split_plot(self):
self.ploti.getAxis("bottom").setTicks(None)
data = self.valid_data
_, bins = np.histogram(data, bins=self.Bins[self.number_of_bins])
_, bins = np.histogram(data, bins=self._get_bins())
gvalues = self.cvar.values
varcolors = [QColor(*col) for col in self.cvar.colors]
if self.fitted_distribution:
Expand Down Expand Up @@ -552,29 +558,41 @@ def _str_int(self, x0, x1):
else:
return f"{var.name} in {var.repr_val(x0)} - {var.repr_val(x1)}"

@staticmethod
def _split_tooltip(valname, tot_group, total, gvalues, freqs):
div_group = tot_group or 1
cs = "white-space:pre; text-align: right;"
b = "border-top: 1px solid black;"
s = f"style='{cs} padding-left: 1em'"
snp = f"style='{cs}'"
return f"<table style='border-collapse: collapse'>" \
f"<tr><th {s}>{valname}:</th>" \
f"<td {snp}><b>{int(tot_group)}</b></td>" \
"<td/>" \
f"<td {s}><b>{100 * tot_group / total:.2f} %</b></td></tr>" + \
f"<tr><td/><td/><td {s}>(in group)</td><td {s}>(overall)</td>" \
"</tr>" + \
"".join(
"<tr>"
f"<th {s}>{value}:</th>"
f"<td {snp}><b>{int(freq)}</b><td`>"
f"<td {s}>{100 * freq / div_group:.2f} %</td>"
f"<td {s}>{100 * freq / total:.2f} %</td>"
"</tr>"
for value, freq in zip(gvalues, freqs)) + \
"</table>"
def _recompute_binnings(self):
if self.var is None:
self.possible_bins = []
return

column = self.data.get_column_view(self.var)[0]
# TODO: do something if all are nan; perhaps in on_var_changed
mn, mx = np.nanmin(column), np.nanmax(column)
diff = mx - mn
f10 = 10 ** -np.floor(np.log10(diff))
self.possible_bins = bins = []
for f in (50, 25, 20, 10, 5, 2, 1,
0.5, 0.25, 0.2, 0.1, 0.05, 0.025, 0.02, 0.01):
width = f / f10
mn_ = np.floor(mn / width) * width
mx_ = np.ceil(mx / width) * width
nbins = np.round((mx_ - mn_) / width)
if 1 < nbins <= 50 and (not bins or bins[-1][1] != nbins):
bins.append((mn_, nbins, width))

user_bins = self._user_var_bins.get(self.var)
self.controls.number_of_bins.setMaximum(len(bins) - 1)
if user_bins is None:
target = len(self.bar_items) or 5
for user_bins, (_, nbins, _2) in enumerate(bins):
if nbins > target:
if user_bins > 0 and \
target - bins[user_bins - 1][1] < nbins - target:
user_bins -= 1
break
self.number_of_bins = user_bins

def _get_bins(self):
mx, n, step = self.possible_bins[self.number_of_bins]
return mx + step * np.arange(n + 1)

def _fit_approximation(self, y):
def join_pars(pairs):
Expand Down Expand Up @@ -622,6 +640,30 @@ def _plot_approximations(self, x0, x1, fitters, colors, prior_probs=0):
if not show_probs:
self.plot_pdf.autoRange()

@staticmethod
def _split_tooltip(valname, tot_group, total, gvalues, freqs):
div_group = tot_group or 1
cs = "white-space:pre; text-align: right;"
b = "border-top: 1px solid black;"
s = f"style='{cs} padding-left: 1em'"
snp = f"style='{cs}'"
return f"<table style='border-collapse: collapse'>" \
f"<tr><th {s}>{valname}:</th>" \
f"<td {snp}><b>{int(tot_group)}</b></td>" \
"<td/>" \
f"<td {s}><b>{100 * tot_group / total:.2f} %</b></td></tr>" + \
f"<tr><td/><td/><td {s}>(in group)</td><td {s}>(overall)</td>" \
"</tr>" + \
"".join(
"<tr>"
f"<th {s}>{value}:</th>"
f"<td {snp}><b>{int(freq)}</b></td>"
f"<td {s}>{100 * freq / div_group:.2f} %</td>"
f"<td {s}>{100 * freq / total:.2f} %</td>"
"</tr>"
for value, freq in zip(gvalues, freqs)) + \
"</table>"

def _on_item_clicked(self, item, modifiers, drag):
def add_or_remove(idx, add):
self.drag_operation = [self.DragRemove, self.DragAdd][add]
Expand Down Expand Up @@ -724,8 +766,9 @@ def apply(self):
histogram_data = None
else:
group_indices, values = self._get_output_indices_cont()
histogram_indices = self._get_histogram_indices()
histogram_data = create_groups_table(data, histogram_indices)
hist_indices, hist_values = self._get_histogram_indices()
histogram_data = create_groups_table(
data, hist_indices, values=hist_values)
selected_data = create_groups_table(
data, group_indices, include_unselected=False, values=values)
annotated_data = create_annotated_table(
Expand All @@ -746,25 +789,33 @@ def _get_output_indices_disc(self):
def _get_output_indices_cont(self):
group_indices = np.zeros(len(self.data), dtype=np.int32)
col = self.data.get_column_view(self.var)[0]
values = []
for group_idx, group in enumerate(self._selection_groups(), start=1):
x0 = x1 = None
for bar_idx in group:
mask = self._get_cont_baritem_indices(col, bar_idx)
minx, maxx, mask = self._get_cont_baritem_indices(col, bar_idx)
if x0 is None:
x0 = minx
x1 = maxx
group_indices[mask] = group_idx
return group_indices, None
values.append(self._str_int(x0, x1))
return group_indices, values

def _get_histogram_indices(self):
group_indices = np.zeros(len(self.data), dtype=np.int32)
col = self.data.get_column_view(self.var)[0]
values = []
for bar_idx in range(len(self.bar_items)):
mask = self._get_cont_baritem_indices(col, bar_idx)
x0, x1, mask = self._get_cont_baritem_indices(col, bar_idx)
group_indices[mask] = bar_idx + 1
return group_indices
values.append(self._str_int(x0, x1))
return group_indices, values

def _get_cont_baritem_indices(self, col, bar_idx):
rect = self.bar_items[bar_idx].boundingRect()
minx = rect.left()
maxx = rect.right() + (bar_idx == len(self.bar_items) - 1)
return (col >= minx) * (col < maxx)
return minx, maxx, (col >= minx) * (col < maxx)

def display_legend(self):
if self.cvar is None:
Expand Down

0 comments on commit 6d08361

Please sign in to comment.