class TemplatePicker(HasTraits):
template = Array
CC = Array
peaks = List
zero=Int(0)
tmp_size = Range(low=2, high=512, value=64, cols=4)
max_pos_x=Int(1023)
max_pos_y=Int(1023)
top = Range(low='zero',high='max_pos_x', value=20, cols=4)
left = Range(low='zero',high='max_pos_y', value=20, cols=4)
is_square = Bool
img_plot = Instance(Plot)
tmp_plot = Instance(Plot)
findpeaks = Button
peak_width = Range(low=2, high=200, value=10)
tab_selected = Event
ShowCC = Bool
img_container = Instance(Component)
container = Instance(Component)
colorbar= Instance(Component)
numpeaks_total = Int(0)
numpeaks_img = Int(0)
OK_custom=OK_custom_handler
cbar_selection = Instance(RangeSelection)
cbar_selected = Event
thresh=Trait(None,None,List,Tuple,Array)
thresh_upper=Float(1.0)
thresh_lower=Float(0.0)
numfiles=Int(1)
img_idx=Int(0)
tmp_img_idx=Int(0)
csr=Instance(BaseCursorTool)
traits_view = View(
HFlow(
VGroup(
Item("img_container",editor=ComponentEditor(), show_label=False),
Group(
Spring(),
Item("ShowCC", editor=BooleanEditor(), label="Show cross correlation image")),
label="Original image", show_border=True, trait_modified="tab_selected"
),
VGroup(
Group(
HGroup(
Item("left", label="Left coordinate", style="custom"),
Item("top", label="Top coordinate", style="custom"),
),
Item("tmp_size", label="Template size", style="custom"),
Item("tmp_plot",editor=ComponentEditor(height=256, width=256), show_label=False, resizable=True),
label="Template", show_border=True),
Group(
Item("peak_width", label="Peak width", style="custom"),
Group(
Spring(),
Item("findpeaks",editor=ButtonEditor(label="Find Peaks"),show_label=False),
Spring(),
),
HGroup(
Item("thresh_lower",label="Threshold Lower Value", editor=TextEditor(evaluate=float,
format_str='%1.4f')),
Item("thresh_upper",label="Threshold Upper Value", editor=TextEditor(evaluate=float,
format_str='%1.4f')),
),
HGroup(
Item("numpeaks_img",label="Number of Cells selected (this image)",style='readonly'),
Spring(),
Item("numpeaks_total",label="Total",style='readonly'), Spring(),
),
label="Peak parameters", show_border=True),
)
),
buttons = [ Action(name='OK', enabled_when = 'numpeaks_total > 0' ),
CancelButton ],
title="Template Picker",
handler=OK_custom, kind='livemodal',
key_bindings = key_bindings,
width=960, height=600)
def __init__(self, signal_instance):
super(TemplatePicker, self).__init__()
try:
import cv
except:
print "OpenCV unavailable. Can't do cross correlation without it. Aborting."
return None
self.OK_custom=OK_custom_handler()
self.sig=signal_instance
if not hasattr(self.sig.mapped_parameters,"original_files"):
self.sig.data=np.atleast_3d(self.sig.data)
self.titles=[self.sig.mapped_parameters.name]
else:
self.numfiles=len(self.sig.mapped_parameters.original_files.keys())
self.titles=self.sig.mapped_parameters.original_files.keys()
tmp_plot_data=ArrayPlotData(imagedata=self.sig.data[self.top:self.top+self.tmp_size,self.left:self.left+self.tmp_size,self.img_idx])
tmp_plot=Plot(tmp_plot_data,default_origin="top left")
tmp_plot.img_plot("imagedata", colormap=jet)
tmp_plot.aspect_ratio=1.0
self.tmp_plot=tmp_plot
self.tmp_plotdata=tmp_plot_data
self.img_plotdata=ArrayPlotData(imagedata=self.sig.data[:,:,self.img_idx])
self.img_container=self._image_plot_container()
self.crop_sig=None
def render_image(self):
plot = Plot(self.img_plotdata,default_origin="top left")
img=plot.img_plot("imagedata", colormap=gray)[0]
plot.title="%s of %s: "%(self.img_idx+1,self.numfiles)+self.titles[self.img_idx]
plot.aspect_ratio=float(self.sig.data.shape[1])/float(self.sig.data.shape[0])
#if not self.ShowCC:
csr = CursorTool(img, drag_button='left', color='white',
line_width=2.0)
self.csr=csr
csr.current_position=self.left, self.top
img.overlays.append(csr)
# attach the rectangle tool
plot.tools.append(PanTool(plot,drag_button="right"))
zoom = ZoomTool(plot, tool_mode="box", always_on=False, aspect_ratio=plot.aspect_ratio)
plot.overlays.append(zoom)
self.img_plot=plot
return plot
def render_scatplot(self):
peakdata=ArrayPlotData()
peakdata.set_data("index",self.peaks[self.img_idx][:,0])
peakdata.set_data("value",self.peaks[self.img_idx][:,1])
peakdata.set_data("color",self.peaks[self.img_idx][:,2])
scatplot=Plot(peakdata,aspect_ratio=self.img_plot.aspect_ratio,default_origin="top left")
scatplot.plot(("index", "value", "color"),
type="cmap_scatter",
name="my_plot",
color_mapper=jet(DataRange1D(low = 0.0,
high = 1.0)),
marker = "circle",
fill_alpha = 0.5,
marker_size = 6,
)
scatplot.x_grid.visible = False
scatplot.y_grid.visible = False
scatplot.range2d=self.img_plot.range2d
self.scatplot=scatplot
self.peakdata=peakdata
return scatplot
def _image_plot_container(self):
plot = self.render_image()
# Create a container to position the plot and the colorbar side-by-side
self.container=OverlayPlotContainer()
self.container.add(plot)
self.img_container = HPlotContainer(use_backbuffer = False)
self.img_container.add(self.container)
self.img_container.bgcolor = "white"
if self.numpeaks_img>0:
scatplot = self.render_scatplot()
self.container.add(scatplot)
colorbar = self.draw_colorbar()
self.img_container.add(colorbar)
return self.img_container
def draw_colorbar(self):
scatplot=self.scatplot
cmap_renderer = scatplot.plots["my_plot"][0]
selection = ColormappedSelectionOverlay(cmap_renderer, fade_alpha=0.35,
selection_type="range")
cmap_renderer.overlays.append(selection)
if self.thresh is not None:
cmap_renderer.color_data.metadata['selections']=self.thresh
cmap_renderer.color_data.metadata_changed={'selections':self.thresh}
# Create the colorbar, handing in the appropriate range and colormap
colormap=scatplot.color_mapper
colorbar = ColorBar(index_mapper=LinearMapper(range=DataRange1D(low = 0.0,
high = 1.0)),
orientation='v',
resizable='v',
width=30,
padding=20)
colorbar_selection=RangeSelection(component=colorbar)
colorbar.tools.append(colorbar_selection)
ovr=colorbar.overlays.append(RangeSelectionOverlay(component=colorbar,
border_color="white",
alpha=0.8,
fill_color="lightgray",
metadata_name='selections'))
#ipshell('colorbar, colorbar_selection and ovr available:')
self.cbar_selection=colorbar_selection
self.cmap_renderer=cmap_renderer
colorbar.plot = cmap_renderer
colorbar.padding_top = scatplot.padding_top
colorbar.padding_bottom = scatplot.padding_bottom
self.colorbar=colorbar
return colorbar
@on_trait_change('ShowCC')
def toggle_cc_view(self):
if self.ShowCC:
self.CC = cv_funcs.xcorr(self.sig.data[self.top:self.top+self.tmp_size,
self.left:self.left+self.tmp_size,self.img_idx],
self.sig.data[:,:,self.img_idx])
self.img_plotdata.set_data("imagedata",self.CC)
else:
self.img_plotdata.set_data("imagedata",self.sig.data[:,:,self.img_idx])
self.redraw_plots()
@on_trait_change("img_idx")
def update_img_depth(self):
if self.ShowCC:
self.CC = cv_funcs.xcorr(self.sig.data[self.top:self.top+self.tmp_size,
self.left:self.left+self.tmp_size,self.img_idx],
self.sig.data[:,:,self.img_idx])
self.img_plotdata.set_data("imagedata",self.CC)
else:
self.img_plotdata.set_data("imagedata",self.sig.data[:,:,self.img_idx])
self.img_plot.title="%s of %s: "%(self.img_idx+1,self.numfiles)+self.titles[self.img_idx]
self.redraw_plots()
@on_trait_change('tmp_size')
def update_max_pos(self):
max_pos_x=self.sig.data.shape[0]-self.tmp_size-1
if self.left>max_pos_x: self.left=max_pos_x
self.max_pos_x=max_pos_x
max_pos_y=self.sig.data.shape[1]-self.tmp_size-1
if self.top>max_pos_y: self.top=max_pos_y
self.max_pos_y=max_pos_y
return
def increase_img_idx(self,info):
if self.img_idx==(self.numfiles-1):
self.img_idx=0
else:
self.img_idx+=1
def decrease_img_idx(self,info):
if self.img_idx==0:
self.img_idx=self.numfiles-1
else:
self.img_idx-=1
@on_trait_change('left, top')
def update_csr_position(self):
self.csr.current_position=self.left,self.top
@on_trait_change('csr:current_position')
def update_top_left(self):
self.left,self.top=self.csr.current_position
@on_trait_change('left, top, tmp_size')
def update_tmp_plot(self):
self.tmp_plotdata.set_data("imagedata",
self.sig.data[self.top:self.top+self.tmp_size,self.left:self.left+self.tmp_size,self.img_idx])
grid_data_source = self.tmp_plot.range2d.sources[0]
grid_data_source.set_data(np.arange(self.tmp_size), np.arange(self.tmp_size))
self.tmp_img_idx=self.img_idx
return
@on_trait_change('left, top, tmp_size')
def update_CC(self):
if self.ShowCC:
self.CC = cv_funcs.xcorr(self.sig.data[self.top:self.top+self.tmp_size,
self.left:self.left+self.tmp_size,self.tmp_img_idx],
self.sig.data[:,:,self.img_idx])
self.img_plotdata.set_data("imagedata",self.CC)
grid_data_source = self.img_plot.range2d.sources[0]
grid_data_source.set_data(np.arange(self.CC.shape[1]),
np.arange(self.CC.shape[0]))
if self.numpeaks_total>0:
self.peaks=[np.array([[0,0,-1]])]
@on_trait_change('cbar_selection:selection')
def update_thresh(self):
try:
thresh=self.cbar_selection.selection
self.thresh=thresh
self.cmap_renderer.color_data.metadata['selections']=thresh
self.thresh_lower=thresh[0]
self.thresh_upper=thresh[1]
#cmap_renderer.color_data.metadata['selection_masks']=self.thresh
self.cmap_renderer.color_data.metadata_changed={'selections':thresh}
self.container.request_redraw()
self.img_container.request_redraw()
except:
pass
@on_trait_change('thresh_upper,thresh_lower')
def manual_thresh_update(self):
self.thresh=[self.thresh_lower,self.thresh_upper]
self.cmap_renderer.color_data.metadata['selections']=self.thresh
self.cmap_renderer.color_data.metadata_changed={'selections':self.thresh}
self.container.request_redraw()
self.img_container.request_redraw()
@on_trait_change('peaks,cbar_selection:selection,img_idx')
def calc_numpeaks(self):
try:
thresh=self.cbar_selection.selection
self.thresh=thresh
except:
thresh=[]
if thresh==[] or thresh==() or thresh==None:
thresh=(0,1)
self.numpeaks_total=int(np.sum([np.sum(np.ma.masked_inside(self.peaks[i][:,2],thresh[0],thresh[1]).mask) for i in xrange(len(self.peaks))]))
try:
self.numpeaks_img=int(np.sum(np.ma.masked_inside(self.peaks[self.img_idx][:,2],thresh[0],thresh[1]).mask))
except:
self.numpeaks_img=0
@on_trait_change('findpeaks')
def locate_peaks(self):
from hyperspy import peak_char as pc
peaks=[]
for idx in xrange(self.numfiles):
self.CC = cv_funcs.xcorr(self.sig.data[self.top:self.top+self.tmp_size,
self.left:self.left+self.tmp_size,self.tmp_img_idx],
self.sig.data[:,:,idx])
# peak finder needs peaks greater than 1. Multiply by 255 to scale them.
pks=pc.two_dim_findpeaks(self.CC*255, peak_width=self.peak_width, medfilt_radius=None)
pks[:,2]=pks[:,2]/255.
peaks.append(pks)
self.peaks=peaks
def mask_peaks(self,idx):
thresh=self.cbar_selection.selection
if thresh==[]:
thresh=(0,1)
mpeaks=np.ma.asarray(self.peaks[idx])
mpeaks[:,2]=np.ma.masked_outside(mpeaks[:,2],thresh[0],thresh[1])
return mpeaks
@on_trait_change("peaks")
def redraw_plots(self):
oldplot=self.img_plot
self.container.remove(oldplot)
newplot=self.render_image()
self.container.add(newplot)
self.img_plot=newplot
try:
# if these haven't been created before, this will fail. wrap in try to prevent that.
oldscat=self.scatplot
self.container.remove(oldscat)
oldcolorbar = self.colorbar
self.img_container.remove(oldcolorbar)
except:
pass
if self.numpeaks_img>0:
newscat=self.render_scatplot()
self.container.add(newscat)
self.scatplot=newscat
colorbar = self.draw_colorbar()
self.img_container.add(colorbar)
self.colorbar=colorbar
self.container.request_redraw()
self.img_container.request_redraw()
def crop_cells_stack(self):
from eelslab.signals.aggregate import AggregateCells
if self.numfiles==1:
self.crop_sig=self.crop_cells()
return
else:
crop_agg=[]
for idx in xrange(self.numfiles):
crop_agg.append(self.crop_cells(idx))
self.crop_sig=AggregateCells(*crop_agg)
return
def crop_cells(self,idx=0):
print "cropping cells..."
from hyperspy.signals.image import Image
# filter the peaks that are outside the selected threshold
peaks=np.ma.compress_rows(self.mask_peaks(idx))
tmp_sz=self.tmp_size
data=np.zeros((tmp_sz,tmp_sz,peaks.shape[0]))
if not hasattr(self.sig.mapped_parameters,"original_files"):
parent=self.sig
else:
parent=self.sig.mapped_parameters.original_files[self.titles[idx]]
for i in xrange(peaks.shape[0]):
# crop the cells from the given locations
data[:,:,i]=self.sig.data[peaks[i,1]:peaks[i,1]+tmp_sz,peaks[i,0]:peaks[i,0]+tmp_sz,idx]
crop_sig=Image({'data':data,
'mapped_parameters':{
'name':'Cropped cells from %s'%self.titles[idx],
'record_by':'image',
'locations':peaks,
'parent':parent,
}
})
return crop_sig
# attach a class member that has the locations from which the images were cropped
print "Complete. "
class TemplatePicker(HasTraits):
template = Array
CC = Array
peaks = List
zero = Int(0)
tmp_size = Range(low=2, high=512, value=64, cols=4)
max_pos_x = Int(1023)
max_pos_y = Int(1023)
top = Range(low='zero', high='max_pos_x', value=20, cols=4)
left = Range(low='zero', high='max_pos_y', value=20, cols=4)
is_square = Bool
img_plot = Instance(Plot)
tmp_plot = Instance(Plot)
findpeaks = Button
peak_width = Range(low=2, high=200, value=10)
tab_selected = Event
ShowCC = Bool
img_container = Instance(Component)
container = Instance(Component)
colorbar = Instance(Component)
numpeaks_total = Int(0)
numpeaks_img = Int(0)
OK_custom = OK_custom_handler
cbar_selection = Instance(RangeSelection)
cbar_selected = Event
thresh = Trait(None, None, List, Tuple, Array)
thresh_upper = Float(1.0)
thresh_lower = Float(0.0)
numfiles = Int(1)
img_idx = Int(0)
tmp_img_idx = Int(0)
csr = Instance(BaseCursorTool)
traits_view = View(HFlow(
VGroup(Item("img_container",
editor=ComponentEditor(),
show_label=False),
Group(
Spring(),
Item("ShowCC",
editor=BooleanEditor(),
label="Show cross correlation image")),
label="Original image",
show_border=True,
trait_modified="tab_selected"),
VGroup(
Group(HGroup(
Item("left", label="Left coordinate", style="custom"),
Item("top", label="Top coordinate", style="custom"),
),
Item("tmp_size", label="Template size", style="custom"),
Item("tmp_plot",
editor=ComponentEditor(height=256, width=256),
show_label=False,
resizable=True),
label="Template",
show_border=True),
Group(Item("peak_width", label="Peak width", style="custom"),
Group(
Spring(),
Item("findpeaks",
editor=ButtonEditor(label="Find Peaks"),
show_label=False),
Spring(),
),
HGroup(
Item("thresh_lower",
label="Threshold Lower Value",
editor=TextEditor(evaluate=float,
format_str='%1.4f')),
Item("thresh_upper",
label="Threshold Upper Value",
editor=TextEditor(evaluate=float,
format_str='%1.4f')),
),
HGroup(
Item("numpeaks_img",
label="Number of Cells selected (this image)",
style='readonly'),
Spring(),
Item("numpeaks_total", label="Total", style='readonly'),
Spring(),
),
label="Peak parameters",
show_border=True),
)),
buttons=[
Action(name='OK',
enabled_when='numpeaks_total > 0'),
CancelButton
],
title="Template Picker",
handler=OK_custom,
kind='livemodal',
key_bindings=key_bindings,
width=960,
height=600)
def __init__(self, signal_instance):
super(TemplatePicker, self).__init__()
try:
import cv
except:
print "OpenCV unavailable. Can't do cross correlation without it. Aborting."
return None
self.OK_custom = OK_custom_handler()
self.sig = signal_instance
if not hasattr(self.sig.mapped_parameters, "original_files"):
self.sig.data = np.atleast_3d(self.sig.data)
self.titles = [self.sig.mapped_parameters.name]
else:
self.numfiles = len(
self.sig.mapped_parameters.original_files.keys())
self.titles = self.sig.mapped_parameters.original_files.keys()
tmp_plot_data = ArrayPlotData(
imagedata=self.sig.data[self.top:self.top + self.tmp_size,
self.left:self.left + self.tmp_size,
self.img_idx])
tmp_plot = Plot(tmp_plot_data, default_origin="top left")
tmp_plot.img_plot("imagedata", colormap=jet)
tmp_plot.aspect_ratio = 1.0
self.tmp_plot = tmp_plot
self.tmp_plotdata = tmp_plot_data
self.img_plotdata = ArrayPlotData(
imagedata=self.sig.data[:, :, self.img_idx])
self.img_container = self._image_plot_container()
self.crop_sig = None
def render_image(self):
plot = Plot(self.img_plotdata, default_origin="top left")
img = plot.img_plot("imagedata", colormap=gray)[0]
plot.title = "%s of %s: " % (self.img_idx + 1,
self.numfiles) + self.titles[self.img_idx]
plot.aspect_ratio = float(self.sig.data.shape[1]) / float(
self.sig.data.shape[0])
#if not self.ShowCC:
csr = CursorTool(img,
drag_button='left',
color='white',
line_width=2.0)
self.csr = csr
csr.current_position = self.left, self.top
img.overlays.append(csr)
# attach the rectangle tool
plot.tools.append(PanTool(plot, drag_button="right"))
zoom = ZoomTool(plot,
tool_mode="box",
always_on=False,
aspect_ratio=plot.aspect_ratio)
plot.overlays.append(zoom)
self.img_plot = plot
return plot
def render_scatplot(self):
peakdata = ArrayPlotData()
peakdata.set_data("index", self.peaks[self.img_idx][:, 0])
peakdata.set_data("value", self.peaks[self.img_idx][:, 1])
peakdata.set_data("color", self.peaks[self.img_idx][:, 2])
scatplot = Plot(peakdata,
aspect_ratio=self.img_plot.aspect_ratio,
default_origin="top left")
scatplot.plot(
("index", "value", "color"),
type="cmap_scatter",
name="my_plot",
color_mapper=jet(DataRange1D(low=0.0, high=1.0)),
marker="circle",
fill_alpha=0.5,
marker_size=6,
)
scatplot.x_grid.visible = False
scatplot.y_grid.visible = False
scatplot.range2d = self.img_plot.range2d
self.scatplot = scatplot
self.peakdata = peakdata
return scatplot
def _image_plot_container(self):
plot = self.render_image()
# Create a container to position the plot and the colorbar side-by-side
self.container = OverlayPlotContainer()
self.container.add(plot)
self.img_container = HPlotContainer(use_backbuffer=False)
self.img_container.add(self.container)
self.img_container.bgcolor = "white"
if self.numpeaks_img > 0:
scatplot = self.render_scatplot()
self.container.add(scatplot)
colorbar = self.draw_colorbar()
self.img_container.add(colorbar)
return self.img_container
def draw_colorbar(self):
scatplot = self.scatplot
cmap_renderer = scatplot.plots["my_plot"][0]
selection = ColormappedSelectionOverlay(cmap_renderer,
fade_alpha=0.35,
selection_type="range")
cmap_renderer.overlays.append(selection)
if self.thresh is not None:
cmap_renderer.color_data.metadata['selections'] = self.thresh
cmap_renderer.color_data.metadata_changed = {
'selections': self.thresh
}
# Create the colorbar, handing in the appropriate range and colormap
colormap = scatplot.color_mapper
colorbar = ColorBar(
index_mapper=LinearMapper(range=DataRange1D(low=0.0, high=1.0)),
orientation='v',
resizable='v',
width=30,
padding=20)
colorbar_selection = RangeSelection(component=colorbar)
colorbar.tools.append(colorbar_selection)
ovr = colorbar.overlays.append(
RangeSelectionOverlay(component=colorbar,
border_color="white",
alpha=0.8,
fill_color="lightgray",
metadata_name='selections'))
#ipshell('colorbar, colorbar_selection and ovr available:')
self.cbar_selection = colorbar_selection
self.cmap_renderer = cmap_renderer
colorbar.plot = cmap_renderer
colorbar.padding_top = scatplot.padding_top
colorbar.padding_bottom = scatplot.padding_bottom
self.colorbar = colorbar
return colorbar
@on_trait_change('ShowCC')
def toggle_cc_view(self):
if self.ShowCC:
self.CC = cv_funcs.xcorr(
self.sig.data[self.top:self.top + self.tmp_size,
self.left:self.left + self.tmp_size,
self.img_idx], self.sig.data[:, :, self.img_idx])
self.img_plotdata.set_data("imagedata", self.CC)
else:
self.img_plotdata.set_data("imagedata",
self.sig.data[:, :, self.img_idx])
self.redraw_plots()
@on_trait_change("img_idx")
def update_img_depth(self):
if self.ShowCC:
self.CC = cv_funcs.xcorr(
self.sig.data[self.top:self.top + self.tmp_size,
self.left:self.left + self.tmp_size,
self.img_idx], self.sig.data[:, :, self.img_idx])
self.img_plotdata.set_data("imagedata", self.CC)
else:
self.img_plotdata.set_data("imagedata",
self.sig.data[:, :, self.img_idx])
self.img_plot.title = "%s of %s: " % (
self.img_idx + 1, self.numfiles) + self.titles[self.img_idx]
self.redraw_plots()
@on_trait_change('tmp_size')
def update_max_pos(self):
max_pos_x = self.sig.data.shape[0] - self.tmp_size - 1
if self.left > max_pos_x: self.left = max_pos_x
self.max_pos_x = max_pos_x
max_pos_y = self.sig.data.shape[1] - self.tmp_size - 1
if self.top > max_pos_y: self.top = max_pos_y
self.max_pos_y = max_pos_y
return
def increase_img_idx(self, info):
if self.img_idx == (self.numfiles - 1):
self.img_idx = 0
else:
self.img_idx += 1
def decrease_img_idx(self, info):
if self.img_idx == 0:
self.img_idx = self.numfiles - 1
else:
self.img_idx -= 1
@on_trait_change('left, top')
def update_csr_position(self):
self.csr.current_position = self.left, self.top
@on_trait_change('csr:current_position')
def update_top_left(self):
self.left, self.top = self.csr.current_position
@on_trait_change('left, top, tmp_size')
def update_tmp_plot(self):
self.tmp_plotdata.set_data(
"imagedata",
self.sig.data[self.top:self.top + self.tmp_size,
self.left:self.left + self.tmp_size, self.img_idx])
grid_data_source = self.tmp_plot.range2d.sources[0]
grid_data_source.set_data(np.arange(self.tmp_size),
np.arange(self.tmp_size))
self.tmp_img_idx = self.img_idx
return
@on_trait_change('left, top, tmp_size')
def update_CC(self):
if self.ShowCC:
self.CC = cv_funcs.xcorr(
self.sig.data[self.top:self.top + self.tmp_size,
self.left:self.left + self.tmp_size,
self.tmp_img_idx], self.sig.data[:, :,
self.img_idx])
self.img_plotdata.set_data("imagedata", self.CC)
grid_data_source = self.img_plot.range2d.sources[0]
grid_data_source.set_data(np.arange(self.CC.shape[1]),
np.arange(self.CC.shape[0]))
if self.numpeaks_total > 0:
self.peaks = [np.array([[0, 0, -1]])]
@on_trait_change('cbar_selection:selection')
def update_thresh(self):
try:
thresh = self.cbar_selection.selection
self.thresh = thresh
self.cmap_renderer.color_data.metadata['selections'] = thresh
self.thresh_lower = thresh[0]
self.thresh_upper = thresh[1]
#cmap_renderer.color_data.metadata['selection_masks']=self.thresh
self.cmap_renderer.color_data.metadata_changed = {
'selections': thresh
}
self.container.request_redraw()
self.img_container.request_redraw()
except:
pass
@on_trait_change('thresh_upper,thresh_lower')
def manual_thresh_update(self):
self.thresh = [self.thresh_lower, self.thresh_upper]
self.cmap_renderer.color_data.metadata['selections'] = self.thresh
self.cmap_renderer.color_data.metadata_changed = {
'selections': self.thresh
}
self.container.request_redraw()
self.img_container.request_redraw()
@on_trait_change('peaks,cbar_selection:selection,img_idx')
def calc_numpeaks(self):
try:
thresh = self.cbar_selection.selection
self.thresh = thresh
except:
thresh = []
if thresh == [] or thresh == () or thresh == None:
thresh = (0, 1)
self.numpeaks_total = int(
np.sum([
np.sum(
np.ma.masked_inside(self.peaks[i][:, 2], thresh[0],
thresh[1]).mask)
for i in xrange(len(self.peaks))
]))
try:
self.numpeaks_img = int(
np.sum(
np.ma.masked_inside(self.peaks[self.img_idx][:, 2],
thresh[0], thresh[1]).mask))
except:
self.numpeaks_img = 0
@on_trait_change('findpeaks')
def locate_peaks(self):
from hyperspy import peak_char as pc
peaks = []
for idx in xrange(self.numfiles):
self.CC = cv_funcs.xcorr(
self.sig.data[self.top:self.top + self.tmp_size,
self.left:self.left + self.tmp_size,
self.tmp_img_idx], self.sig.data[:, :, idx])
# peak finder needs peaks greater than 1. Multiply by 255 to scale them.
pks = pc.two_dim_findpeaks(self.CC * 255,
peak_width=self.peak_width,
medfilt_radius=None)
pks[:, 2] = pks[:, 2] / 255.
peaks.append(pks)
self.peaks = peaks
def mask_peaks(self, idx):
thresh = self.cbar_selection.selection
if thresh == []:
thresh = (0, 1)
mpeaks = np.ma.asarray(self.peaks[idx])
mpeaks[:, 2] = np.ma.masked_outside(mpeaks[:, 2], thresh[0], thresh[1])
return mpeaks
@on_trait_change("peaks")
def redraw_plots(self):
oldplot = self.img_plot
self.container.remove(oldplot)
newplot = self.render_image()
self.container.add(newplot)
self.img_plot = newplot
try:
# if these haven't been created before, this will fail. wrap in try to prevent that.
oldscat = self.scatplot
self.container.remove(oldscat)
oldcolorbar = self.colorbar
self.img_container.remove(oldcolorbar)
except:
pass
if self.numpeaks_img > 0:
newscat = self.render_scatplot()
self.container.add(newscat)
self.scatplot = newscat
colorbar = self.draw_colorbar()
self.img_container.add(colorbar)
self.colorbar = colorbar
self.container.request_redraw()
self.img_container.request_redraw()
def crop_cells_stack(self):
from eelslab.signals.aggregate import AggregateCells
if self.numfiles == 1:
self.crop_sig = self.crop_cells()
return
else:
crop_agg = []
for idx in xrange(self.numfiles):
crop_agg.append(self.crop_cells(idx))
self.crop_sig = AggregateCells(*crop_agg)
return
def crop_cells(self, idx=0):
print "cropping cells..."
from hyperspy.signals.image import Image
# filter the peaks that are outside the selected threshold
peaks = np.ma.compress_rows(self.mask_peaks(idx))
tmp_sz = self.tmp_size
data = np.zeros((tmp_sz, tmp_sz, peaks.shape[0]))
if not hasattr(self.sig.mapped_parameters, "original_files"):
parent = self.sig
else:
parent = self.sig.mapped_parameters.original_files[
self.titles[idx]]
for i in xrange(peaks.shape[0]):
# crop the cells from the given locations
data[:, :, i] = self.sig.data[peaks[i, 1]:peaks[i, 1] + tmp_sz,
peaks[i,
0]:peaks[i, 0] + tmp_sz, idx]
crop_sig = Image({
'data': data,
'mapped_parameters': {
'name': 'Cropped cells from %s' % self.titles[idx],
'record_by': 'image',
'locations': peaks,
'parent': parent,
}
})
return crop_sig
# attach a class member that has the locations from which the images were cropped
print "Complete. "
