class RegionSlicer(QtWidgets.QWidget):
'''
**Bases:** :class: `QWidget <pyqt.QtWidgets.QWidget>`
Adds a movable <pyqtgraph.LinearRegionItem> to a plot.
Provides a numpy slice and mask that would slices the x_array according to the region.
=============================== =============================================================================
**Signals:**
region_changed_signal() Emitted when region is changed or activated by user
=============================== =============================================================================
=============================== =============================================================================
**Properties:**
slice: numpy.slice[start:stop]
s_: :property:slice if activated else *s_return_on_deactivated*
mask: numpy boolean array of size 'x_array': True for where x is within region
false otherwise (ignores step). Useful if x_array is not ordered.
m_: mask if activated, else a mask equivalent to
*s_return_on_deactivated*
=============================== =============================================================================
'''
region_changed_signal = QtCore.Signal()
def __init__(self,
plot_item,
x_array=None,
name='array_slicer_name',
slicer_updated_func=lambda: None,
initial=[0, 100],
s_return_on_deactivated=np.s_[:],
brush=QtGui.QColor(0, 255, 0, 70),
ZValue=10,
font=QtGui.QFont("Times", 12),
label_line=1,
activated=False):
"""Create a new LinearRegionItem on plot_item.
====================== ==============================================================
**Arguments:**
plot_item <pyqtgraph.PlotDataItem> (recommended)
or <pyqtgraph.PlotItem> (does not grab x_array data from plot
item: initialize x_array manually
or use :func:`set_x_array <self.set_x_array>`)
x_array use to initialize x_array if plot_item == <pyqtgraph.PlotItem>.
name <str>
slicer_updated_func gets called when region is updated, alternatively use
:sig:region_changed_signal().
initial initial index values [start,stop,step] or [start,stop]
s_return_on_deactivated Object returned if RegionSlicer is not activated.
Slicing with np.s_[:] (default) and np.s_[0] gives the full and
an empty array respectively. Note, <RegionSlicer.slice> always
returns the last determined slice even if slice is deactivated.
brush,ZValue, are passed to the LinearRegionItem
font, passed to the label
label_line '0' or '1' for placement of label onto 'left' or 'right'
bounding line.
activated <bool> state at initialization
====================== ===============================================================
"""
QtWidgets.QWidget.__init__(self)
self.name = name
self.slicer_updated = slicer_updated_func
from ScopeFoundry.logged_quantity import LQCollection
self.settings = LQCollection()
if len(initial) == 2: initial.append(1)
start, stop, step = initial
self.start = self.settings.New('start', int, initial=start, vmin=0)
self.stop = self.settings.New('stop', int, initial=stop, vmin=0)
self.step = self.settings.New('step', int, initial=step)
self.activated = self.settings.New('activated',
bool,
initial=activated)
self.start.add_listener(self.on_change_start_stop)
self.stop.add_listener(self.on_change_start_stop)
self.activated.add_listener(self.on_change_activated)
self.s_return_on_deactivated = s_return_on_deactivated
if type(plot_item) == pg.PlotDataItem:
self.plot_data_item = plot_item
self.plot_data_item.sigPlotChanged.connect(
self.set_x_array_from_data_item)
self.parent_plot_item = plot_item.parentItem()
elif type(plot_item) == pg.PlotItem:
self.plot_data_item = None
self.parent_plot_item = plot_item
self.linear_region_item = pg.LinearRegionItem(brush=brush)
self.linear_region_item.setZValue(ZValue)
self.parent_plot_item.addItem(self.linear_region_item)
self.linear_region_item.sigRegionChangeFinished.connect(
self.on_change_region)
self.inf_line_label = pg.InfLineLabel(
self.linear_region_item.lines[label_line],
self.name,
position=0.78,
anchor=(0.5, 0.5))
#self.inf_line_label = pg.TextItem(self.name, anchor=(0.5, 0.5))
self.inf_line_label.setFont(font)
self.set_label('')
if x_array == None: #give it something to work with.
x_array = np.arange(512)
self.set_x_array(x_array)
@QtCore.Slot(object)
def set_x_array_from_data_item(self):
x_array, _ = self.plot_data_item.getData()
if not np.array_equal(x_array, self.x_array):
self.apply_new_x_array(x_array)
#print('set_x_array_from_data_item: new x_array()')
else:
pass
def set_x_array(self, x_array):
'''
use this function to update the x_array
not required to use if type(plot_item) == pg.PlotDataItem
'''
self.apply_new_x_array(x_array)
def apply_new_x_array(self, x_array):
mn, mx = self.linear_region_item.getRegion()
self.x_array = x_array
self.linear_region_item.setBounds(
[self.x_array.min(), self.x_array.max()])
self.settings['start'] = np.argmin((x_array - mn)**2)
self.settings['stop'] = np.argmin((x_array - mx)**2)
@property
def slice(self):
S = self.settings
return np.s_[S['start']:S['stop']:S['step']]
@property
def s_(self):
'''returns slice based on region if activated else `s_return_on_deactivated`'''
if self.activated.val:
return self.slice
else:
return self.s_return_on_deactivated
@property
def mask(self):
X = self.x_array[self.slice]
return (self.x_array >= X.min()) * (self.x_array <= X.max())
@property
def m_(self):
X = self.x_array[np.s_]
return (self.x_array >= X.min()) * (self.x_array <= X.max())
@QtCore.Slot(object)
def on_change_region(self):
'''
updates settings based on region
'''
print(self.name, 'on_change_region')
self.region_min, self.region_max = mn, mx = self.linear_region_item.getRegion(
)
self.settings['start'] = np.argmin((self.x_array - mn)**2)
self.settings['stop'] = np.argmin((self.x_array - mx)**2)
self.region_changed_signal.emit()
self.slicer_updated()
def on_change_start_stop(self):
'''
updates linear_region_item based on settings
'''
rgn = [0, 0]
rgn[0] = self.x_array[self.settings['start']]
try:
rgn[1] = self.x_array[self.settings['stop']]
except IndexError:
rgn[1] = self.x_array[self.settings['start']]
self.linear_region_item.setRegion(rgn)
def on_change_activated(self):
activated = self.activated.val
self.linear_region_item.setEnabled(activated)
self.linear_region_item.setAcceptHoverEvents(activated)
self.linear_region_item.setAcceptTouchEvents(activated)
self.linear_region_item.setVisible(activated)
self.region_changed_signal.emit()
self.slicer_updated()
def New_UI(self, include=None, exclude=[]):
ui_widget = self.settings.New_UI(include, exclude)
ui_widget.layout().insertRow(
0, QtWidgets.QLabel("<b>{}</b>".format(self.name)))
return ui_widget
def set_label(self, text='', title=None, color=(200, 200, 200)):
if title == None:
title = self.name
label = '<h3>{}</h3>{}'.format(title, text)
self.inf_line_label.setHtml(label)
self.inf_line_label.setColor(color)
def get_sliced_xar(self, ar=None):
if ar == None:
if hasattr(self, 'plot_data_item'):
_, y = self.plot_data_item.getData()
return (self.x_array[self.s_], y[self.s_])
else:
return None
else:
return (self.x_array[self.s_], ar[self.s_])
class HyperSpectralBaseView(DataBrowserView):
name = 'HyperSpectralBaseView'
def setup(self):
## Dummy data Structures (override in func:self.load_data())
self.hyperspec_data = np.arange(10 * 10 * 34).reshape((10, 10, 34))
self.display_image = self.hyperspec_data.sum(
-1) # np.random.rand(10,10)
self.spec_x_array = np.arange(34)
# Call :func:set_scalebar_params() during self.load_data() to add a scalebar!
self.scalebar_type = None
# Will be filled derived maps and images
self.display_images = dict()
self.spec_x_arrays = dict()
## Graphs and Interface
self.ui = self.dockarea = dockarea.DockArea()
self.imview = pg.ImageView()
self.imview.getView().invertY(False) # lower left origin
self.image_dock = self.dockarea.addDock(name='Image',
widget=self.imview)
self.graph_layout = pg.GraphicsLayoutWidget()
self.spec_dock = self.dockarea.addDock(name='Spec Plot',
widget=self.graph_layout)
self.line_colors = ['w', 'r', 'g', 'b', 'y', 'm', 'c']
self.spec_plot = self.graph_layout.addPlot()
self.spec_plot.setLabel('left', 'Intensity', units='counts')
self.rect_plotdata = self.spec_plot.plot(y=[0, 2, 1, 3, 2],
pen=self.line_colors[0])
self.point_plotdata = self.spec_plot.plot(y=[0, 2, 1, 3, 2],
pen=self.line_colors[1])
self.point_plotdata.setZValue(-1)
#correlation plot
self.corr_layout = pg.GraphicsLayoutWidget()
self.corr_plot = self.corr_layout.addPlot()
self.corr_plotdata = pg.ScatterPlotItem(x=[0, 1, 2, 3, 4],
y=[0, 2, 1, 3, 2],
size=17,
pen=pg.mkPen(None),
brush=pg.mkBrush(
255, 255, 255, 60))
self.corr_plot.addItem(self.corr_plotdata)
self.corr_dock = self.dockarea.addDock(name='correlation',
widget=self.corr_layout,
position='below',
relativeTo=self.spec_dock)
self.corr_plotdata.sigClicked.connect(self.corr_plot_clicked)
self.spec_dock.raiseDock()
# Rectangle ROI
self.rect_roi = pg.RectROI([20, 20], [20, 20], pen=self.line_colors[0])
self.rect_roi.addTranslateHandle((0.5, 0.5))
self.imview.getView().addItem(self.rect_roi)
self.rect_roi.sigRegionChanged[object].connect(self.on_change_rect_roi)
# Point ROI
self.circ_roi = pg.CircleROI((0, 0), (2, 2),
movable=True,
pen=self.line_colors[1])
#self.circ_roi.removeHandle(self.circ_roi.getHandles()[0])
h = self.circ_roi.addTranslateHandle((0.5, .5))
h.pen = pg.mkPen(pen=self.line_colors[1])
h.update()
self.imview.getView().addItem(self.circ_roi)
self.circ_roi.removeHandle(0)
self.circ_roi_plotline = pg.PlotCurveItem([0], pen=self.line_colors[1])
self.imview.getView().addItem(self.circ_roi_plotline)
self.circ_roi.sigRegionChanged[object].connect(self.on_update_circ_roi)
#font
font = QtGui.QFont("Times", 12)
font.setBold(True)
#settings
self.default_display_image_choices = ['default', 'sum']
self.settings.New('display_image',
str,
choices=self.default_display_image_choices,
initial='default')
self.settings.display_image.add_listener(self.on_change_display_image)
self.default_x_axis_choices = ['default', 'index']
self.x_axis = self.settings.New('x_axis',
str,
initial='default',
choices=self.default_x_axis_choices)
self.x_axis.add_listener(self.on_change_x_axis)
self.norm_data = self.settings.New('norm_data', bool, initial=False)
self.norm_data.add_listener(self.update_display)
self.bg_subtract = self.settings.New('bg_subtract',
str,
initial='None',
choices=('None', 'bg_slice',
'costum_const'))
self.norm_data.add_listener(self.update_display)
self.bg_counts = self.settings.New('bg_value',
initial=0,
unit='cts/bin')
self.bg_counts.add_listener(self.update_display)
self.settings.New('default_view_on_load', bool, initial=True)
self.binning = self.settings.New('binning', int, initial=1, vmin=1)
self.binning.add_listener(self.update_display)
self.spatial_binning = self.settings.New('spatial_binning',
int,
initial=1,
vmin=1)
self.spatial_binning.add_listener(self.bin_spatially)
self.cor_X_data = self.settings.New(
'cor_X_data',
str,
choices=self.default_display_image_choices,
initial='default')
self.cor_Y_data = self.settings.New(
'cor_Y_data',
str,
choices=self.default_display_image_choices,
initial='sum')
self.cor_X_data.add_listener(self.on_change_corr_settings)
self.cor_Y_data.add_listener(self.on_change_corr_settings)
# data slicers
self.x_slicer = RegionSlicer(self.spec_plot,
name='x_slice',
brush=QtGui.QColor(0, 255, 0, 70),
ZValue=10,
font=font,
initial=[100, 511])
self.bg_slicer = RegionSlicer(self.spec_plot,
name='bg_slice',
slicer_updated_func=self.update_display,
brush=QtGui.QColor(255, 255, 255, 70),
ZValue=11,
font=font,
initial=[0, 80],
label_line=0)
self.bg_slicer.activated.add_listener(
lambda: self.bg_subtract.update_value('bg_slice')
if self.bg_slicer.activated.val else None)
self.show_lines = ['show_circ_line', 'show_rect_line']
for x in self.show_lines:
lq = self.settings.New(x, bool, initial=True)
lq.add_listener(self.on_change_show_lines)
# peakutils
self.peakutils_settings = LQCollection()
self.show_peak_line = self.peakutils_settings.New('show_peak_line',
bool,
initial=False)
self.show_peak_line.add_listener(self.update_display)
self.baseline_deg = self.peakutils_settings.New('baseline_deg',
int,
initial=0,
vmin=-1,
vmax=100)
self.baseline_deg.add_listener(self.update_display)
self.thres = self.peakutils_settings.New('thres',
float,
initial=0.5,
vmin=0,
vmax=1)
self.thres.add_listener(self.update_display)
self.peakutils_settings.New('unique_solution', bool, initial=False)
self.peakutils_settings.New('min_dist', int, initial=-1)
self.peakutils_settings.New('gaus_fit_refinement', bool, initial=True)
self.peakutils_settings.New('ignore_phony_refinements',
bool,
initial=True)
self.base_plotdata = self.spec_plot.plot(y=[0, 2, 1, 3, 2],
pen=self.line_colors[2])
self.peak_lines = []
# Settings Docks
self.settings_dock = self.dockarea.addDock(name='settings',
position='left',
relativeTo=self.image_dock)
self.settings_widgets = [
] # Hack part 1/2: allows to use settings.New_UI() and have settings defined in scan_specific_setup()
self.settings_widgets.append(self.x_slicer.New_UI())
self.settings_widgets.append(self.bg_slicer.New_UI())
self.scan_specific_setup()
self.generate_settings_ui(
) # Hack part 2/2: Need to generate settings after scan_specific_setup()
self.peakutils_dock = self.dockarea.addDock(
name='PeakUtils', position='below', relativeTo=self.settings_dock)
self.peakutils_dock.addWidget(self.peakutils_settings.New_UI())
self.settings_dock.raiseDock()
self.settings_dock.setStretch(x=0, y=0)
self.peakutils_dock.setStretch(x=0, y=0)
self.image_dock.setStretch(x=100, y=1)
#self.settings_dock.widgetArea.setStyleSheet('Dock > QWidget {border:0px; border-radius:0px}')
#self.peakutils_dock.widgetArea.setStyleSheet('Dock > QWidget {border:0px; border-radius:0px}')
def generate_settings_ui(self):
self.settings_ui = self.settings.New_UI()
self.settings_dock.addWidget(self.settings_ui)
#some more self.settings_widgets[]
#self.update_display_pushButton = QtWidgets.QPushButton(text = 'update display')
#self.settings_widgets.append(self.update_display_pushButton)
#self.update_display_pushButton.clicked.connect(self.update_display)
self.default_view_pushButton = QtWidgets.QPushButton(
text='default img view')
self.settings_widgets.append(self.default_view_pushButton)
self.default_view_pushButton.clicked.connect(self.default_image_view)
self.recalc_median_pushButton = QtWidgets.QPushButton(
text='recalc median map')
self.settings_widgets.append(self.recalc_median_pushButton)
self.recalc_median_pushButton.clicked.connect(self.recalc_median_map)
self.recalc_sum_pushButton = QtWidgets.QPushButton(
text='recalc sum map')
self.settings_widgets.append(self.recalc_sum_pushButton)
self.recalc_sum_pushButton.clicked.connect(self.recalc_sum_map)
self.recalc_peak_map_pushButton = QtWidgets.QPushButton(
text='recalc PeakUtils map')
self.settings_widgets.append(self.recalc_peak_map_pushButton)
self.recalc_peak_map_pushButton.clicked.connect(self.recalc_peak_map)
self.save_state_pushButton = QtWidgets.QPushButton(text='save_state')
self.settings_widgets.append(self.save_state_pushButton)
self.save_state_pushButton.clicked.connect(self.save_state)
self.delete_current_display_image_pushButton = QtWidgets.QPushButton(
text='remove image')
self.settings_widgets.append(
self.delete_current_display_image_pushButton)
self.delete_current_display_image_pushButton.clicked.connect(
self.delete_current_display_image)
# Place the self.settings_widgets[] on a grid
ui_widget = QtWidgets.QWidget()
gridLayout = QtWidgets.QGridLayout()
gridLayout.setSpacing(0)
gridLayout.setContentsMargins(0, 0, 0, 0)
ui_widget.setLayout(gridLayout)
for i, w in enumerate(self.settings_widgets):
gridLayout.addWidget(w, int(i / 2), i % 2)
self.settings_dock.addWidget(ui_widget)
self.settings_dock.layout.setSpacing(0)
def add_spec_x_array(self, key, array):
self.spec_x_arrays[key] = array
self.settings.x_axis.add_choices(key, allow_duplicates=False)
def add_display_image(self, key, image):
key = self.add_descriptor_suffixes(key)
self.display_images[key] = image
self.settings.display_image.add_choices(key, allow_duplicates=False)
self.cor_X_data.change_choice_list(self.display_images.keys())
self.cor_Y_data.change_choice_list(self.display_images.keys())
self.cor_X_data.update_value(self.cor_Y_data.val)
self.cor_Y_data.update_value(key)
self.on_change_corr_settings()
def add_descriptor_suffixes(self, key):
if self.x_slicer.activated.val:
key += '_x{}-{}'.format(self.x_slicer.start.val,
self.x_slicer.stop.val)
if self.settings[
'bg_subtract'] == 'bg_slice' and self.bg_slicer.activated.val:
key += '_bg{}-{}'.format(self.bg_slicer.start.val,
self.bg_slicer.stop.val)
if self.settings['bg_subtract'] == 'costum_count':
key += '_bg{1.2f}'.format(self.bg_counts.val)
return key
def delete_current_display_image(self):
key = self.settings.display_image.val
del self.display_images[key]
self.settings.display_image.remove_choices(key)
self.cor_X_data.remove_choices(key)
self.cor_Y_data.remove_choices(key)
def get_xy(self, ji_slice, apply_use_x_slice=False):
'''
returns processed hyperspec_data averaged over a given spatial slice.
'''
x, hyperspec_dat = self.get_xhyperspec_data(apply_use_x_slice)
y = hyperspec_dat[ji_slice].mean(axis=(0, 1))
#self.databrowser.ui.statusbar.showMessage('get_xy(), counts in slice: {}'.format( y.sum() ) )
if self.settings['norm_data']:
y = norm(y)
return (x, y)
def get_bg(self):
bg_subtract_mode = self.bg_subtract.val
if bg_subtract_mode == 'bg_slice':
if not self.bg_slicer.activated:
self.bg_slicer.activated.update_value(True)
bg_slice = self.bg_slicer.slice
bg = self.hyperspec_data[:, :, bg_slice].mean()
self.bg_slicer.set_label(title=bg_subtract_mode,
text='{:1.1f} cts<br>{} bins'.format(
bg, bg_slice.stop - bg_slice.start))
elif bg_subtract_mode == 'costum_const':
bg = self.bg_counts.val
self.bg_slicer.set_label('', title=bg_subtract_mode)
else:
bg = 0
self.bg_slicer.set_label('', title=bg_subtract_mode)
return bg
def get_xhyperspec_data(self, apply_use_x_slice=True):
'''
returns processed hyperspec_data
'''
bg = self.get_bg()
hyperspec_data = self.hyperspec_data
x = self.spec_x_array
if apply_use_x_slice and self.x_slicer.activated.val:
x = x[self.x_slicer.slice]
hyperspec_data = hyperspec_data[:, :, self.x_slicer.slice]
binning = self.settings['binning']
if binning != 1:
x, hyperspec_data = bin_y_average_x(x,
hyperspec_data,
binning,
-1,
datapoints_lost_warning=False)
bg *= binning
msg = 'effective subtracted bg value is binnging*bg ={:0.1f} which is up to {:2.1f}% of max value.'.format(
bg, bg / np.max(hyperspec_data) * 100)
self.databrowser.ui.statusbar.showMessage(msg)
if self.settings['norm_data']:
return (x, norm_map(hyperspec_data - bg))
else:
return (x, hyperspec_data - bg)
def on_change_x_axis(self):
key = self.settings['x_axis']
if key in self.spec_x_arrays:
self.spec_x_array = self.spec_x_arrays[key]
self.x_slicer.set_x_array(self.spec_x_array)
self.bg_slicer.set_x_array(self.spec_x_array)
self.spec_plot.setLabel('bottom', key)
self.update_display()
def on_change_display_image(self):
key = self.settings['display_image']
if key in self.display_images:
self.display_image = self.display_images[key]
self.update_display()
if self.display_image.shape == (1, 1):
self.databrowser.ui.statusbar.showMessage(
'Can not display single pixel image!')
def scan_specific_setup(self):
#override this!
pass
def is_file_supported(self, fname):
# override this!
return False
def post_load(self):
# override this!
pass
def on_change_data_filename(self, fname):
self.reset()
try:
self.scalebar_type = None
self.load_data(fname)
if self.settings['spatial_binning'] != 1:
self.hyperspec_data = bin_2D(self.hyperspec_data,
self.settings['spatial_binning'])
self.display_image = bin_2D(self.display_image,
self.settings['spatial_binning'])
self.display_images['default'] = self.display_image
self.display_images['sum'] = self.hyperspec_data.sum(axis=-1)
self.spec_x_arrays['default'] = self.spec_x_array
self.spec_x_arrays['index'] = np.arange(
self.hyperspec_data.shape[-1])
self.databrowser.ui.statusbar.clearMessage()
self.post_load()
self.add_scalebar()
except Exception as err:
HyperSpectralBaseView.load_data(self,
fname) # load default dummy data
self.databrowser.ui.statusbar.showMessage(
"failed to load {}: {}".format(fname, err))
raise (err)
finally:
self.on_change_display_image()
self.on_change_corr_settings()
self.update_display()
self.on_change_x_axis()
if self.settings['default_view_on_load']:
self.default_image_view()
def add_scalebar(self):
''' not intended to use: Call set_scalebar_params() during load_data()'''
if hasattr(self, 'scalebar'):
self.imview.getView().removeItem(self.scalebar)
del self.scalebar
if self.scalebar_type == 'ConfocalScaleBar':
from viewers.scalebars import ConfocalScaleBar
num_px = self.display_image.shape[1] #horizontal dimension!
kwargs = self.scalebar_kwargs
self.scalebar = ConfocalScaleBar(num_px=num_px, **kwargs)
self.scalebar.setParentItem(self.imview.getView())
self.scalebar.anchor((1, 1), (1, 1), offset=kwargs['offset'])
elif self.scalebar_type == None:
self.scalebar = None
def set_scalebar_params(self,
h_span,
units='m',
scalebar_type='ConfocalScaleBar',
stroke_width=10,
brush='w',
pen='k',
offset=(-20, -20)):
'''
call this function during load_data() to add a scalebar!
*h_span* horizontal length of image in units of *units* if positive.
Else, scalebar is in units of pixels (*units* ignored).
*units* SI length unit of *h_span*.
*scalebar_type* is either `None` (no scalebar will be added)
or `"ConfocalScaleBar"` (default).
*stroke_width*, *brush*, *pen* and *offset* affect appearance and
positioning of the scalebar.
'''
assert scalebar_type in [None, 'ConfocalScaleBar']
self.scalebar_type = scalebar_type
span_meter = {
'm': 1,
'cm': 1e-2,
'mm': 1e-3,
'um': 1e-6,
'nm': 1e-9,
'pm': 1e-12,
'fm': 1e-15
}[units] * h_span
self.scalebar_kwargs = {
'span': span_meter,
'brush': brush,
'pen': pen,
'width': stroke_width,
'offset': offset
}
def update_display(self):
# pyqtgraph axes are (x,y), but display_images are in (y,x) so we need to transpose
if self.display_image is not None:
self.imview.setImage(self.display_image.T)
self.on_change_rect_roi()
self.on_update_circ_roi()
def reset(self):
'''
resets the dictionaries
'''
keys_to_delete = list(
set(self.display_images.keys()) -
set(self.default_display_image_choices))
for key in keys_to_delete:
del self.display_images[key]
keys_to_delete = list(
set(self.spec_x_arrays.keys()) - set(self.default_x_axis_choices))
for key in keys_to_delete:
del self.spec_x_arrays[key]
self.settings.display_image.change_choice_list(
self.default_display_image_choices)
self.settings.x_axis.change_choice_list(self.default_x_axis_choices)
def load_data(self, fname):
"""
override to set hyperspectral dataset and the display image
need to define:
* self.hyperspec_data (shape Ny, Nx, Nspec)
* self.display_image (shape Ny, Nx)
* self.spec_x_array (shape Nspec)
"""
self.hyperspec_data = np.arange(10 * 10 * 34).reshape((10, 10, 34))
self.display_image = self.hyperspec_data.sum(-1)
self.spec_x_array = np.arange(34)
@QtCore.Slot(object)
def on_change_rect_roi(self, roi=None):
# pyqtgraph axes are (x,y), but hyperspec is in (y,x,spec) hence axes=(1,0)
roi_slice, roi_tr = self.rect_roi.getArraySlice(
self.hyperspec_data, self.imview.getImageItem(), axes=(1, 0))
self.rect_roi_slice = roi_slice
x, y = self.get_xy(roi_slice, apply_use_x_slice=False)
self.rect_plotdata.setData(x, y)
self.on_change_corr_settings()
self.update_peaks(*self.get_xy(roi_slice, apply_use_x_slice=True),
pen=self.line_colors[0])
@QtCore.Slot(object)
def on_update_circ_roi(self, roi=None):
if roi is None:
roi = self.circ_roi
roi_state = roi.saveState()
x0, y0 = roi_state['pos']
xc = x0 + 1
yc = y0 + 1
Ny, Nx, Nspec = self.hyperspec_data.shape
i = max(0, min(int(xc), Nx - 1))
j = max(0, min(int(yc), Ny - 1))
self.circ_roi_plotline.setData([xc, i + 0.5], [yc, j + 0.5])
self.circ_roi_ji = (j, i)
self.circ_roi_slice = np.s_[j:j + 1, i:i + 1]
x, y = self.get_xy(self.circ_roi_slice, apply_use_x_slice=False)
self.point_plotdata.setData(x, y)
self.on_change_corr_settings()
self.update_peaks(*self.get_xy(self.circ_roi_slice,
apply_use_x_slice=True),
pen=self.line_colors[1])
def update_peaks(self, wls, spec, pen='g'):
self.base_plotdata.setVisible(
self.peakutils_settings['show_peak_line'])
for l in self.peak_lines:
self.spec_plot.removeItem(l)
l.deleteLater()
self.peak_lines = []
if self.peakutils_settings['show_peak_line']:
PS = self.peakutils_settings
import peakutils
base = 1.0 * peakutils.baseline(spec, PS['baseline_deg'])
self.base_plotdata.setData(wls, base)
self.base_plotdata.setPen(pen)
if PS['min_dist'] < 0:
min_dist = int(len(wls) / 2)
else:
min_dist = PS['min_dist']
peaks_ = peaks(spec - base, wls, PS['thres'],
PS['unique_solution'], min_dist,
PS['gaus_fit_refinement'],
PS['ignore_phony_refinements'])
for p in np.atleast_1d(peaks_):
l = pg.InfiniteLine(pos=(p, 0),
movable=False,
angle=90,
pen=pen,
label='PeakUtils {value:0.2f}',
labelOpts={
'color': pen,
'movable': True,
'fill': (200, 200, 200, 100)
})
self.spec_plot.addItem(l)
self.peak_lines.append(l)
def on_change_show_lines(self):
self.point_plotdata.setVisible(self.settings['show_circ_line'])
self.rect_plotdata.setVisible(self.settings['show_rect_line'])
def default_image_view(self):
'sets rect_roi congruent to imageItem and optimizes size of imageItem to fit the ViewBox'
iI = self.imview.imageItem
h, w = iI.height(), iI.width()
self.rect_roi.setSize((w, h))
self.rect_roi.setPos((0, 0))
self.imview.getView().enableAutoRange()
self.spec_plot.enableAutoRange()
def recalc_median_map(self):
x, hyperspec_data = self.get_xhyperspec_data(apply_use_x_slice=True)
median_map = spectral_median_map(hyperspec_data, x)
self.add_display_image('median_map', median_map)
def recalc_sum_map(self):
x, hyperspec_data = self.get_xhyperspec_data(apply_use_x_slice=True)
_sum = hyperspec_data.sum(-1)
self.add_display_image('sum', _sum)
def recalc_peak_map(self):
x, hyperspec_data = self.get_xhyperspec_data(apply_use_x_slice=True)
PS = self.peakutils_settings
_map = peak_map(hyperspec_data, x, PS['thres'], int(len(x) / 2),
PS['gaus_fit_refinement'],
PS['ignore_phony_refinements'])
map_name = 'peak_map'
if PS['gaus_fit_refinement']:
map_name += '_refined'
if PS['ignore_phony_refinements']:
map_name += '_ignored'
self.add_display_image(map_name, _map)
def on_change_corr_settings(self):
try:
xname = self.settings['cor_X_data']
yname = self.settings['cor_Y_data']
X = self.display_images[xname]
Y = self.display_images[yname]
#Note, the correlation plot is a dimensionality reduction
# (i,j,X,Y) --> (X,Y). To map the scatter points back to the image
# we need to associate every (X,Y) on the correlation plot with
# their indices (i,j); in particular
# indices = [(j0,i0), (j0,i1), ...]
indices = list(np.indices((X.shape)).reshape(2, -1).T)
self.corr_plotdata.setData(X.flat,
Y.flat,
brush=pg.mkBrush(255, 255, 255, 50),
pen=None,
data=indices)
# mark points within rect_roi
mask = np.zeros_like(X, dtype=bool)
mask[self.rect_roi_slice[0:2]] = True
cor_x = X[mask].flatten()
cor_y = Y[mask].flatten()
self.corr_plotdata.addPoints(cor_x,
cor_y,
brush=pg.mkBrush(255, 255, 204, 60),
pen=pg.mkPen(self.line_colors[0],
width=0.5))
# mark circ_roi point
j, i = self.circ_roi_ji
x_circ, y_circ = np.atleast_1d(X[j, i]), np.atleast_1d(Y[j, i])
self.corr_plotdata.addPoints(x=x_circ,
y=y_circ,
pen=pg.mkPen(self.line_colors[1],
width=3))
##some more plot details
#self.corr_plot.getViewBox().setRange(xRange=(cor_x.min(), cor_x.max()),
# yRange=(cor_y.min(), cor_y.max()))
self.corr_plot.autoRange()
self.corr_plot.setLabels(**{'bottom': xname, 'left': yname})
sm = spearmanr(cor_x, cor_y)
text = 'Pearson\'s corr: {:.3f}<br>Spearman\'s: corr={:.3f}, pvalue={:.3f}'.format(
np.corrcoef(cor_x, cor_y)[0, 1], sm.correlation, sm.pvalue)
self.corr_plot.setTitle(text)
except Exception as err:
self.databrowser.ui.statusbar.showMessage(
'Error in on_change_corr_settings: {}'.format(err))
def corr_plot_clicked(self, plotitem, points):
'''
call back function to locate a point on the correlation plot on the image.
*points* is a list of <pg.ScatterPlotItem.SpotItem> under the mouse
pointer during click event. For points within the rect_roi, there are
two items representing a pixel, but only most button one contains the
(j,i) as data.
'''
j, i = points[-1].data()
self.circ_roi.setPos(i - 0.5, j - 0.5)
def bin_spatially(self):
if not (self.settings['display_image']
in self.default_display_image_choices):
self.settings.display_image.update_value(
self.default_display_image_choices[0])
fname = self.databrowser.settings['data_filename']
self.on_change_data_filename(fname)
def save_state(self):
from ScopeFoundry import h5_io
fname = self.databrowser.settings['data_filename']
view_state_fname = '{fname}_state_view_{timestamp:%y%m%d_%H%M%S}.{ext}'.format(
fname=fname.strip('.h5'),
timestamp=datetime.fromtimestamp(time.time()),
ext='h5')
h5_file = h5py.File(name=view_state_fname)
with h5_file as h5_file:
h5_group_display_images = h5_file.create_group('display_images')
for k, v in self.display_images.items():
h5_group_display_images.create_dataset(k, data=v)
h5_group_spec_x_array = h5_file.create_group('spec_x_arrays')
for k, v in self.spec_x_arrays.items():
h5_group_spec_x_array.create_dataset(k, data=v)
h5_group_settings_group = h5_file.create_group('settings')
h5_io.h5_save_lqcoll_to_attrs(self.settings,
h5_group_settings_group)
h5_group_settings_group = h5_file.create_group('x_slicer_settings')
h5_io.h5_save_lqcoll_to_attrs(self.x_slicer.settings,
h5_group_settings_group)
h5_group_settings_group = h5_file.create_group(
'bg_slicer_settings')
h5_io.h5_save_lqcoll_to_attrs(self.bg_slicer.settings,
h5_group_settings_group)
self.view_specific_save_state_func(h5_file)
h5_file.close()
def view_specific_save_state_func(self, h5_file):
'''
you can override me, use 'h5_file' - it's already open
e.g: h5_file.create_group('scan_specific_settings')
...
'''
pass
def load_state(self, fname_idx=-1):
# does not work properly, maybe because the order the settings are set matters?
path = sibling_path(self.databrowser.settings['data_filename'], '')
pre_state_fname = self.databrowser.settings['data_filename'].strip(
path).strip('.h5')
state_files = []
for x in os.listdir(path):
if pre_state_fname in x:
if 'state_view' in x:
state_files.append(x)
print('state_files', state_files)
if len(state_files) != 0:
h5_file = h5py.File(path + state_files[fname_idx])
for k, v in h5_file['bg_slicer_settings'].attrs.items():
try:
self.bg_slicer.settings[k] = v
except:
pass
for k, v in h5_file['x_slicer_settings'].attrs.items():
try:
self.x_slicer.settings[k] = v
except:
pass
for k, v in h5_file['settings'].attrs.items():
try:
self.settings[k] = v
except:
pass
for k, v in h5_file['biexponential_settings'].attrs.items():
self.biexponential_settings[k] = v
for k, v in h5_file['export_settings'].attrs.items():
self.export_settings[k] = v
h5_file.close()
print('loaded', state_files[fname_idx])
