Skip to content

Commit

Permalink
Merge pull request #931 from kernc/paintdata-fixups
Browse files Browse the repository at this point in the history 
OWPaintData: import numpy as np
  • Loading branch information
@lanzagar
lanzagar committed Dec 16, 2015
2 parents 238edaf + aed1e2c commit dc0c44b
Showing 1 changed file with 34 additions and 34 deletions.
68 changes: 34 additions & 34 deletions Orange/widgets/data/owpaintdata.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@
import itertools
from functools import partial

import numpy
import numpy as np

from PyQt4 import QtCore
from PyQt4 import QtGui
Expand Down Expand Up @@ -38,7 +38,7 @@ def indices_to_mask(indices, size):
:param int size: Size of the resulting mask.
"""
mask = numpy.zeros(size, dtype=bool)
mask = np.zeros(size, dtype=bool)
mask[indices] = True
return mask

Expand All @@ -59,7 +59,7 @@ def stack_on_condition(a, b, condition):
shape = list(a.shape)
shape[axis] = N
shape = tuple(shape)
arr = numpy.empty(shape, dtype=a.dtype)
arr = np.empty(shape, dtype=a.dtype)
arr[condition] = a
arr[~condition] = b
return arr
Expand Down Expand Up @@ -113,8 +113,8 @@ def transform(command, data):

@transform.register(Append)
def append(command, data):
numpy.clip(command.points[:, :2], 0, 1, out=command.points[:, :2])
return (numpy.vstack([data, command.points]),
np.clip(command.points[:, :2], 0, 1, out=command.points[:, :2])
return (np.vstack([data, command.points]),
DeleteIndices(slice(len(data),
len(data) + len(command.points))))

Expand All @@ -131,8 +131,8 @@ def delete(command, data, ):
if isinstance(command.indices, slice):
condition = indices_to_mask(command.indices, len(data))
else:
indices = numpy.asarray(command.indices)
if indices.dtype == numpy.bool:
indices = np.asarray(command.indices)
if indices.dtype == np.bool:
condition = indices
else:
condition = indices_to_mask(indices, len(data))
Expand Down Expand Up @@ -351,17 +351,17 @@ def __timout(self):


def random_state(rstate):
if isinstance(rstate, numpy.random.RandomState):
if isinstance(rstate, np.random.RandomState):
return rstate
else:
return numpy.random.RandomState(rstate)
return np.random.RandomState(rstate)


def create_data(x, y, radius, size, rstate):
random = random_state(rstate)
x = random.normal(x, radius / 2, size=size)
y = random.normal(y, radius / 2, size=size)
return numpy.c_[x, y]
return np.c_[x, y]


class MagnetTool(DataTool):
Expand Down Expand Up @@ -651,7 +651,7 @@ def indices_eq(ind1, ind2):
elif ind1 is ... and ind2 is ...:
return True

ind1, ind1 = numpy.array(ind1), numpy.array(ind2)
ind1, ind1 = np.array(ind1), np.array(ind2)

if ind1.shape != ind2.shape or ind1.dtype != ind2.dtype:
return False
Expand All @@ -667,37 +667,37 @@ def merge_cmd(composit):
g = merge_cmd(g)

if isinstance(f, Append) and isinstance(g, Append):
return Append(numpy.vstack((f.points, g.points)))
return Append(np.vstack((f.points, g.points)))
elif isinstance(f, Move) and isinstance(g, Move):
if indices_eq(f.indices, g.indices):
return Move(f.indices, f.delta + g.delta)
else:
# TODO: union of indices, ...
return composit
# elif isinstance(f, DeleteIndices) and isinstance(g, DeleteIndices):
# indices = numpy.array(g.indices)
# indices = np.array(g.indices)
# return DeleteIndices(indices)
else:
return composit


def apply_attractor(data, point, density, radius):
delta = data - point
dist_sq = numpy.sum(delta ** 2, axis=1)
dist = numpy.sqrt(dist_sq)
dist_sq = np.sum(delta ** 2, axis=1)
dist = np.sqrt(dist_sq)

dist[dist < radius] = 0
dist_sq = dist ** 2
valid = (dist_sq > 100 * numpy.finfo(dist.dtype).eps)
valid = (dist_sq > 100 * np.finfo(dist.dtype).eps)
assert valid.shape == (dist.shape[0],)

df = 0.05 * density / dist_sq[valid]

df_bound = 1 - radius / dist[valid]

df = numpy.clip(df, 0, df_bound)
df = np.clip(df, 0, df_bound)

dx = numpy.zeros_like(delta)
dx = np.zeros_like(delta)
dx[valid] = df.reshape(-1, 1) * delta[valid]
return dx

Expand All @@ -706,15 +706,15 @@ def apply_jitter(data, point, density, radius, rstate=None):
random = random_state(rstate)

delta = data - point
dist_sq = numpy.sum(delta ** 2, axis=1)
dist = numpy.sqrt(dist_sq)
valid = dist_sq > 100 * numpy.finfo(dist_sq.dtype).eps
dist_sq = np.sum(delta ** 2, axis=1)
dist = np.sqrt(dist_sq)
valid = dist_sq > 100 * np.finfo(dist_sq.dtype).eps

df = 0.05 * density / dist_sq[valid]
df_bound = 1 - radius / dist[valid]
df = numpy.clip(df, 0, df_bound)
df = np.clip(df, 0, df_bound)

dx = numpy.zeros_like(delta)
dx = np.zeros_like(delta)
jitter = random.normal(0, 0.1, size=(df.size, data.shape[1]))

dx[valid, :] = df.reshape(-1, 1) * jitter
Expand Down Expand Up @@ -798,7 +798,7 @@ def __init__(self):
self.class_model.rowsInserted.connect(self._class_count_changed)
self.class_model.rowsRemoved.connect(self._class_count_changed)

self.data = numpy.zeros((0, 3))
self.data = np.zeros((0, 3))
self.colors = colorpalette.ColorPaletteGenerator(
len(colorpalette.DefaultRGBColors))
self.tools_cache = {}
Expand Down Expand Up @@ -975,11 +975,11 @@ def set_data(self, data):
y = np.ones(X.shape[0])
else:
y = data[:, y].Y
while len(self.class_model) < numpy.unique(y).size:
while len(self.class_model) < np.unique(y).size:
self.add_new_class_label(undoable=False)

X = numpy.array([scale(vals) for vals in data.X[:, :2].T]).T
self.data = numpy.column_stack((X, y))
X = np.array([scale(vals) for vals in data.X[:, :2].T]).T
self.data = np.column_stack((X, y))
self._replot()

def add_new_class_label(self, undoable=True):
Expand Down Expand Up @@ -1102,15 +1102,15 @@ def _add_command(self, cmd):

if isinstance(cmd, Append):
cls = self.selected_class_label()
points = numpy.array([(p.x(), p.y() if self.hasAttr2 else 0, cls)
points = np.array([(p.x(), p.y() if self.hasAttr2 else 0, cls)
for p in cmd.points])
self.undo_stack.push(UndoCommand(Append(points), self, text=name))
elif isinstance(cmd, Move):
self.undo_stack.push(UndoCommand(cmd, self, text=name))
elif isinstance(cmd, SelectRegion):
indices = [i for i, (x, y) in enumerate(self.data[:, :2])
if cmd.region.contains(QPointF(x, y))]
indices = numpy.array(indices, dtype=int)
indices = np.array(indices, dtype=int)
self._selected_indices = indices
elif isinstance(cmd, DeleteSelection):
indices = self._selected_indices
Expand All @@ -1124,7 +1124,7 @@ def _add_command(self, cmd):
self.undo_stack.push(
UndoCommand(
Move((self._selected_indices, slice(0, 2)),
numpy.array([cmd.delta.x(), cmd.delta.y()])),
np.array([cmd.delta.x(), cmd.delta.y()])),
self, text="Move")
)
elif isinstance(cmd, DeleteIndices):
Expand All @@ -1137,12 +1137,12 @@ def _add_command(self, cmd):
1 + self.density / 20, cmd.rstate)
self._add_command(Append([QPointF(*p) for p in zip(*data.T)]))
elif isinstance(cmd, Jitter):
point = numpy.array([cmd.pos.x(), cmd.pos.y()])
point = np.array([cmd.pos.x(), cmd.pos.y()])
delta = - apply_jitter(self.data[:, :2], point,
self.density / 100.0, 0, cmd.rstate)
self._add_command(Move((..., slice(0, 2)), delta))
elif isinstance(cmd, Magnet):
point = numpy.array([cmd.pos.x(), cmd.pos.y()])
point = np.array([cmd.pos.x(), cmd.pos.y()])
delta = - apply_attractor(self.data[:, :2], point,
self.density / 100.0, 0)
self._add_command(Move((..., slice(0, 2)), delta))
Expand All @@ -1164,7 +1164,7 @@ def pen(color):

self._scatter_item = pg.ScatterPlotItem(
self.data[:, 0],
self.data[:, 1] if self.hasAttr2 else numpy.zeros(self.data.shape[0]),
self.data[:, 1] if self.hasAttr2 else np.zeros(self.data.shape[0]),
symbol="+",
pen=[pens[int(ci)] for ci in self.data[:, 2]]
)
Expand All @@ -1185,7 +1185,7 @@ def commit(self):
attrs = (Orange.data.ContinuousVariable(self.attr1),
Orange.data.ContinuousVariable(self.attr2))
else:
X, Y = self.data[:, numpy.newaxis, 0], self.data[:, 2]
X, Y = self.data[:, np.newaxis, 0], self.data[:, 2]
attrs = (Orange.data.ContinuousVariable(self.attr1),)
if len(self.class_model) > 1:
domain = Orange.data.Domain(
Expand Down

0 comments on commit dc0c44b

Please sign in to comment.