def triangles(im,
points=[],
contour=None,
fib_step=None,
isblank=False,
queue=None):
'''
Divides the image into triangular regions.
Arguments:
im: the image to work with.
points: a set of prefered points to use instead of provided option. Usually cached points
from previous itterations.
countour: the color of the contour to use. By default the contour is set to 30 intensity
darker than the color in the current region for contrast.
mesh: if not none, denotes the chunck size that the finonacci mesh will be used, but only if points is []
lloyd: when True, a lloyd mesh will be used, but only if points is []
isblank: when True, the image will just have the contour of the mesh overlayed when
filled. This flag treats the image as blank. if contour is set, the overlay will have the
contour=contour, else, 0 (black)
queue: when not None, the image is sent to the message queue instead of being returned.
This is particulal handy when running concurently.
'''
# copy the image, to prevent inplace operations
img = im.copy()
if len(points) != 0:
# we got free points to work with
pass
elif fib_step:
# use fibonacci mesh
points = fib_mesh(img, fib_step)
else:
# use random points
points = random_pts(img)
# create Delaunay triangles
tri = Delaunay(points)
# get all the vertices of the triangles to sent them to be filled
v = [points[x] for x in tri.simplices]
# fill all the vertices.
fill(img, v, contour=contour, isblank=isblank)
# if running concurently send the items to the queue
if queue:
label = 'Triangles: '
label += 'Fibonacci-Steps: {} '.format(
fib_step) if fib_step is not None else ''
label += 'Random-Steps: {} '.format(True) if fib_step is None else ''
queue.put((label, img))
else:
# return return the image instead.
return img
def pydoc(op, children=()):
doc = "\n\n".join([fill(p.text(), 0) for p in op.query["doc"]])
for ch in children:
doc += "\n\n " + pythonize(ch["@name"]) + " -- " + str(ch["@label"])
ch_descs ="\n\n".join([fill(p.text(), 4) for p in ch.query["doc"]])
if ch_descs:
doc += "\n\n" + ch_descs
return doc
def pydoc(op, children=()):
doc = "\n\n".join([fill(p.text(), 0) for p in op.query["doc"]])
for ch in children:
doc += "\n\n " + pythonize(ch["@name"]) + " -- " + str(ch["@label"])
ch_descs = "\n\n".join([fill(p.text(), 4) for p in ch.query["doc"]])
if ch_descs:
doc += "\n\n" + ch_descs
return doc
def docstring(self):
s = "\n\n".join([fill(d, 2) for d in [self.description] + self.docs])
for f in self.fields:
if f.docs:
s += "\n\n" + "\n\n".join([fill(f.docs[0], 4, f.name)] +
[fill(d, 4) for d in f.docs[1:]])
if self.responses:
s += "\n\nValid responses: "
for r in self.responses:
s += r.name + " "
return s
def docstring(self):
s = "\n\n".join([fill(d, 2) for d in [self.description] + self.docs])
for f in self.fields:
if f.docs:
s += "\n\n" + "\n\n".join([fill(f.docs[0], 4, f.name)] +
[fill(d, 4) for d in f.docs[1:]])
if self.responses:
s += "\n\nValid responses: "
for r in self.responses:
s += r.name + " "
return s
def pentagons(im,
points=[],
contour=None,
fib_step=None,
lloyd_cells=None,
isblank=False,
queue=None):
'''
Divides the image into hexagon/pentagon regions
Arguments:
im: the image to work with.
points: a set of prefered points to use instead of provided option. Usually cached points
from previous itterations.
countour: the color of the contour to use. By default the contour is set to 30 intensity
darker than the color in the current region for contrast.
mesh: when True, a finonacci mesh will be used, but only if points is []
lloyd: when True, a lloyd mesh will be used, but only if points is []
isblank: when True, the image will just have the contour of the mesh overlayed when
filled. This flag treats the image as blank. if contour is set, the overlay will have the
contour=contour, else, 0 (black)
queue: when not None, the image is sent to the message queue instead of being returned.
This is particulal handy when running concurently.
'''
# copy the image, to prevent inplace operations
img = im.copy()
if len(points) != 0:
# we got free points to work with
pass
elif fib_step:
# use fibonacci mesh
points = fib_mesh(img, fib_step)
elif lloyd_cells:
# use lloyd mesh
points = lloyd_mesh(img, lloyd_cells)
else:
# use random points
points = random_pts(img)
# create the voronoi dragrams
vor = Voronoi(points)
# get all the vertices ready and sent them to be filled or contoured
v = []
# get all the vertices of this region and
# check if the current region isn't reaching towards infinity
for region in vor.regions:
if not -1 in region and len(region) > 0:
v.append(vor.vertices[region])
# fill all regions at once
fill(img, v, contour=contour, isblank=isblank)
# if running concurently send the items to the queue
if queue:
label = 'Pentagons: '
label += 'Fibonacci-Steps: {} '.format(
fib_step) if fib_step is not None else ''
label += 'Lloyd-Cell-Size: {} '.format(
lloyd_cells) if lloyd_cells is not None else ''
label += 'Random-Steps: {} '.format(
True) if fib_step is None and lloyd_cells is None else ''
queue.put((label, img))
else:
# return them instead.
return img
parL = np.empty(numBins, np.dtype('float64'))
parLerr = np.empty(numBins, np.dtype('float64'))
for ibin in range(0, numBins):
hists.append(
r.TH1D("tmp_" + str(ibin), "tmp_" + str(ibin), 400, 0., 2.))
energies.append(int(ibin * energyBin))
energies.append(maxEnergy)
if numBins > len(ut.colors()):
for i in range(0, numBins - len(ut.colors())):
ut.colors().append(ut.colors()[i])
ut.alpha().append(0.3)
ut.lines().append(ut.lines()[i])
ut.width().append(ut.width()[i])
ut.fill().append(3001)
leg = r.TLegend(0.25, 0.55, 0.3, 0.85)
leg.SetTextFont(132)
leg.SetTextSize(0.05)
leg.SetFillColor(0)
leg.SetFillStyle(0)
ifile = "/eos/user/c/cneubuse/miniCalo2/pred/stage" + str(st) + "/out.root"
miX = 0.8
maX = 1.18
# if args.stage>1:
# miX=0.8
# maX=1.2
