class ArrayVolumePlot(qt.QWidget):
"""
Widget for plotting a n-D array (n >= 3) as a 3D scalar field.
Three axis arrays can be provided to calibrate the axes.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last 3 dimensions must have the same length as
the axes arrays.
Sliders are provided to select indices on the first (n - 3) dimensions of
the signal array, and the plot is updated to load the stack corresponding
to the selection.
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayVolumePlot, self).__init__(parent)
self.__signal = None
self.__signal_name = None
# the Z, Y, X axes apply to the last three dimensions of the signal
# (in that order)
self.__z_axis = None
self.__z_axis_name = None
self.__y_axis = None
self.__y_axis_name = None
self.__x_axis = None
self.__x_axis_name = None
from ._VolumeWindow import VolumeWindow
self._view = VolumeWindow(self)
self._hline = qt.QFrame(self)
self._hline.setFrameStyle(qt.QFrame.HLine)
self._hline.setFrameShadow(qt.QFrame.Sunken)
self._legend = qt.QLabel(self)
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self.__selector_is_connected = False
layout = qt.QVBoxLayout()
layout.addWidget(self._view)
layout.addWidget(self._hline)
layout.addWidget(self._legend)
layout.addWidget(self._selector)
self.setLayout(layout)
def getVolumeView(self):
"""Returns the plot used for the display
:rtype: SceneWindow
"""
return self._view
def setData(self,
signal,
x_axis=None,
y_axis=None,
z_axis=None,
signal_name=None,
xlabel=None,
ylabel=None,
zlabel=None,
title=None):
"""
:param signal: n-D dataset, whose last 3 dimensions are used as the
3D stack values.
:param x_axis: 1-D dataset used as the image's x coordinates. If
provided, its lengths must be equal to the length of the last
dimension of ``signal``.
:param y_axis: 1-D dataset used as the image's y. If provided,
its lengths must be equal to the length of the 2nd to last
dimension of ``signal``.
:param z_axis: 1-D dataset used as the image's z. If provided,
its lengths must be equal to the length of the 3rd to last
dimension of ``signal``.
:param signal_name: Label used in the legend
:param xlabel: Label for X axis
:param ylabel: Label for Y axis
:param zlabel: Label for Z axis
:param title: Graph title
"""
if self.__selector_is_connected:
self._selector.selectionChanged.disconnect(self._updateVolume)
self.__selector_is_connected = False
self.__signal = signal
self.__signal_name = signal_name or ""
self.__x_axis = x_axis
self.__x_axis_name = xlabel
self.__y_axis = y_axis
self.__y_axis_name = ylabel
self.__z_axis = z_axis
self.__z_axis_name = zlabel
self._selector.setData(signal)
self._selector.setAxisNames(["Y", "X", "Z"])
self._updateVolume()
# the legend label shows the selection slice producing the volume
# (only interesting for ndim > 3)
if signal.ndim > 3:
self._selector.setVisible(True)
self._legend.setVisible(True)
self._hline.setVisible(True)
else:
self._selector.setVisible(False)
self._legend.setVisible(False)
self._hline.setVisible(False)
if not self.__selector_is_connected:
self._selector.selectionChanged.connect(self._updateVolume)
self.__selector_is_connected = True
def _updateVolume(self):
"""Update displayed stack according to the current axes selector
data."""
x_axis = self.__x_axis
y_axis = self.__y_axis
z_axis = self.__z_axis
offset = []
scale = []
for axis in [x_axis, y_axis, z_axis]:
if axis is None:
calibration = NoCalibration()
elif len(axis) == 2:
calibration = LinearCalibration(y_intercept=axis[0],
slope=axis[1])
else:
calibration = ArrayCalibration(axis)
if not calibration.is_affine():
_logger.warning("Axis has not linear values, ignored")
offset.append(0.)
scale.append(1.)
else:
offset.append(calibration(0))
scale.append(calibration.get_slope())
legend = self.__signal_name + "["
for sl in self._selector.selection():
if sl == slice(None):
legend += ":, "
else:
legend += str(sl) + ", "
legend = legend[:-2] + "]"
self._legend.setText("Displayed data: " + legend)
# Update SceneWidget
data = self._selector.selectedData()
volumeView = self.getVolumeView()
volumeView.setData(data, offset=offset, scale=scale)
volumeView.setAxesLabels(self.__x_axis_name, self.__y_axis_name,
self.__z_axis_name)
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self.getVolumeView().clear()
class ArrayComplexImagePlot(qt.QWidget):
"""
Widget for plotting an image of complex from a multi-dimensional signal array
and two 1D axes array.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last two dimensions must have the same length as
the axes arrays.
Sliders are provided to select indices on the first (n - 2) dimensions of
the signal array, and the plot is updated to show the image corresponding
to the selection.
If one or both of the axes does not have regularly spaced values, the
the image is plotted as a coloured scatter plot.
"""
def __init__(self, parent=None, colormap=None):
"""
:param parent: Parent QWidget
"""
super(ArrayComplexImagePlot, self).__init__(parent)
self.__signals = None
self.__signals_names = None
self.__x_axis = None
self.__x_axis_name = None
self.__y_axis = None
self.__y_axis_name = None
self._plot = ComplexImageView(self)
if colormap is not None:
for mode in (ComplexImageView.ComplexMode.ABSOLUTE,
ComplexImageView.ComplexMode.SQUARE_AMPLITUDE,
ComplexImageView.ComplexMode.REAL,
ComplexImageView.ComplexMode.IMAGINARY):
self._plot.setColormap(colormap, mode)
self._plot.getPlot().getIntensityHistogramAction().setVisible(True)
self._plot.setKeepDataAspectRatio(True)
# not closable
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self._selector.selectionChanged.connect(self._updateImage)
self._auxSigSlider = HorizontalSliderWithBrowser(parent=self)
self._auxSigSlider.setMinimum(0)
self._auxSigSlider.setValue(0)
self._auxSigSlider.valueChanged[int].connect(self._sliderIdxChanged)
self._auxSigSlider.setToolTip("Select auxiliary signals")
layout = qt.QVBoxLayout()
layout.addWidget(self._plot)
layout.addWidget(self._selector)
layout.addWidget(self._auxSigSlider)
self.setLayout(layout)
def _sliderIdxChanged(self, value):
self._updateImage()
def getPlot(self):
"""Returns the plot used for the display
:rtype: PlotWidget
"""
return self._plot.getPlot()
def setImageData(self,
signals,
x_axis=None,
y_axis=None,
signals_names=None,
xlabel=None,
ylabel=None,
title=None):
"""
:param signals: list of n-D datasets, whose last 2 dimensions are used as the
image's values, or list of 3D datasets interpreted as RGBA image.
:param x_axis: 1-D dataset used as the image's x coordinates. If
provided, its lengths must be equal to the length of the last
dimension of ``signal``.
:param y_axis: 1-D dataset used as the image's y. If provided,
its lengths must be equal to the length of the 2nd to last
dimension of ``signal``.
:param signals_names: Names for each image, used as subtitle and legend.
:param xlabel: Label for X axis
:param ylabel: Label for Y axis
:param title: Graph title
"""
self._selector.selectionChanged.disconnect(self._updateImage)
self._auxSigSlider.valueChanged.disconnect(self._sliderIdxChanged)
self.__signals = signals
self.__signals_names = signals_names
self.__x_axis = x_axis
self.__x_axis_name = xlabel
self.__y_axis = y_axis
self.__y_axis_name = ylabel
self.__title = title
self._selector.clear()
self._selector.setAxisNames(["Y", "X"])
self._selector.setData(signals[0])
if len(signals[0].shape) <= 2:
self._selector.hide()
else:
self._selector.show()
self._auxSigSlider.setMaximum(len(signals) - 1)
if len(signals) > 1:
self._auxSigSlider.show()
else:
self._auxSigSlider.hide()
self._auxSigSlider.setValue(0)
self._updateImage()
self._plot.getPlot().resetZoom()
self._selector.selectionChanged.connect(self._updateImage)
self._auxSigSlider.valueChanged.connect(self._sliderIdxChanged)
def _updateImage(self):
selection = self._selector.selection()
auxSigIdx = self._auxSigSlider.value()
images = [img[selection] for img in self.__signals]
image = images[auxSigIdx]
x_axis = self.__x_axis
y_axis = self.__y_axis
if x_axis is None and y_axis is None:
xcalib = NoCalibration()
ycalib = NoCalibration()
else:
if x_axis is None:
# no calibration
x_axis = numpy.arange(image.shape[1])
elif numpy.isscalar(x_axis) or len(x_axis) == 1:
# constant axis
x_axis = x_axis * numpy.ones((image.shape[1], ))
elif len(x_axis) == 2:
# linear calibration
x_axis = x_axis[0] * numpy.arange(image.shape[1]) + x_axis[1]
if y_axis is None:
y_axis = numpy.arange(image.shape[0])
elif numpy.isscalar(y_axis) or len(y_axis) == 1:
y_axis = y_axis * numpy.ones((image.shape[0], ))
elif len(y_axis) == 2:
y_axis = y_axis[0] * numpy.arange(image.shape[0]) + y_axis[1]
xcalib = ArrayCalibration(x_axis)
ycalib = ArrayCalibration(y_axis)
self._plot.setData(image)
if xcalib.is_affine():
xorigin, xscale = xcalib(0), xcalib.get_slope()
else:
_logger.warning("Unsupported complex image X axis calibration")
xorigin, xscale = 0., 1.
if ycalib.is_affine():
yorigin, yscale = ycalib(0), ycalib.get_slope()
else:
_logger.warning("Unsupported complex image Y axis calibration")
yorigin, yscale = 0., 1.
self._plot.setOrigin((xorigin, yorigin))
self._plot.setScale((xscale, yscale))
if self.__title:
title = self.__title
if len(self.__signals_names) > 1:
# Append dataset name only when there is many datasets
title += '\n' + self.__signals_names[auxSigIdx]
else:
title = self.__signals_names[auxSigIdx]
self._plot.setGraphTitle(title)
self._plot.getXAxis().setLabel(self.__x_axis_name)
self._plot.getYAxis().setLabel(self.__y_axis_name)
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self._plot.setData(None)
class ArrayStackPlot(qt.QWidget):
"""
Widget for plotting a n-D array (n >= 3) as a stack of images.
Three axis arrays can be provided to calibrate the axes.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last 3 dimensions must have the same length as
the axes arrays.
Sliders are provided to select indices on the first (n - 3) dimensions of
the signal array, and the plot is updated to load the stack corresponding
to the selection.
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayStackPlot, self).__init__(parent)
self.__signal = None
self.__signal_name = None
# the Z, Y, X axes apply to the last three dimensions of the signal
# (in that order)
self.__z_axis = None
self.__z_axis_name = None
self.__y_axis = None
self.__y_axis_name = None
self.__x_axis = None
self.__x_axis_name = None
self._stack_view = StackView(self)
self._hline = qt.QFrame(self)
self._hline.setFrameStyle(qt.QFrame.HLine)
self._hline.setFrameShadow(qt.QFrame.Sunken)
self._legend = qt.QLabel(self)
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self.__selector_is_connected = False
layout = qt.QVBoxLayout()
layout.addWidget(self._stack_view)
layout.addWidget(self._hline)
layout.addWidget(self._legend)
layout.addWidget(self._selector)
self.setLayout(layout)
def getStackView(self):
"""Returns the plot used for the display
:rtype: StackView
"""
return self._stack_view
def setStackData(self,
signal,
x_axis=None,
y_axis=None,
z_axis=None,
signal_name=None,
xlabel=None,
ylabel=None,
zlabel=None,
title=None):
"""
:param signal: n-D dataset, whose last 3 dimensions are used as the
3D stack values.
:param x_axis: 1-D dataset used as the image's x coordinates. If
provided, its lengths must be equal to the length of the last
dimension of ``signal``.
:param y_axis: 1-D dataset used as the image's y. If provided,
its lengths must be equal to the length of the 2nd to last
dimension of ``signal``.
:param z_axis: 1-D dataset used as the image's z. If provided,
its lengths must be equal to the length of the 3rd to last
dimension of ``signal``.
:param signal_name: Label used in the legend
:param xlabel: Label for X axis
:param ylabel: Label for Y axis
:param zlabel: Label for Z axis
:param title: Graph title
"""
if self.__selector_is_connected:
self._selector.selectionChanged.disconnect(self._updateStack)
self.__selector_is_connected = False
self.__signal = signal
self.__signal_name = signal_name or ""
self.__x_axis = x_axis
self.__x_axis_name = xlabel
self.__y_axis = y_axis
self.__y_axis_name = ylabel
self.__z_axis = z_axis
self.__z_axis_name = zlabel
self._selector.setData(signal)
self._selector.setAxisNames(["Y", "X", "Z"])
self._stack_view.setGraphTitle(title or "")
# by default, the z axis is the image position (dimension not plotted)
self._stack_view.getPlotWidget().getXAxis().setLabel(self.__x_axis_name
or "X")
self._stack_view.getPlotWidget().getYAxis().setLabel(self.__y_axis_name
or "Y")
self._updateStack()
ndims = len(signal.shape)
self._stack_view.setFirstStackDimension(ndims - 3)
# the legend label shows the selection slice producing the volume
# (only interesting for ndim > 3)
if ndims > 3:
self._selector.setVisible(True)
self._legend.setVisible(True)
self._hline.setVisible(True)
else:
self._selector.setVisible(False)
self._legend.setVisible(False)
self._hline.setVisible(False)
if not self.__selector_is_connected:
self._selector.selectionChanged.connect(self._updateStack)
self.__selector_is_connected = True
@staticmethod
def _get_origin_scale(axis):
"""Assuming axis is a regularly spaced 1D array,
return a tuple (origin, scale) where:
- origin = axis[0]
- scale = (axis[n-1] - axis[0]) / (n -1)
:param axis: 1D numpy array
:return: Tuple (axis[0], (axis[-1] - axis[0]) / (len(axis) - 1))
"""
return axis[0], (axis[-1] - axis[0]) / (len(axis) - 1)
def _updateStack(self):
"""Update displayed stack according to the current axes selector
data."""
stk = self._selector.selectedData()
x_axis = self.__x_axis
y_axis = self.__y_axis
z_axis = self.__z_axis
calibrations = []
for axis in [z_axis, y_axis, x_axis]:
if axis is None:
calibrations.append(NoCalibration())
elif len(axis) == 2:
calibrations.append(
LinearCalibration(y_intercept=axis[0], slope=axis[1]))
else:
calibrations.append(ArrayCalibration(axis))
legend = self.__signal_name + "["
for sl in self._selector.selection():
if sl == slice(None):
legend += ":, "
else:
legend += str(sl) + ", "
legend = legend[:-2] + "]"
self._legend.setText("Displayed data: " + legend)
self._stack_view.setStack(stk, calibrations=calibrations)
self._stack_view.setLabels(labels=[
self.__z_axis_name, self.__y_axis_name, self.__x_axis_name
])
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self._stack_view.clear()
class ArrayImagePlot(qt.QWidget):
"""
Widget for plotting an image from a multi-dimensional signal array
and two 1D axes array.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last two dimensions must have the same length as
the axes arrays.
Sliders are provided to select indices on the first (n - 2) dimensions of
the signal array, and the plot is updated to show the image corresponding
to the selection.
If one or both of the axes does not have regularly spaced values, the
the image is plotted as a coloured scatter plot.
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayImagePlot, self).__init__(parent)
self.__signals = None
self.__signals_names = None
self.__x_axis = None
self.__x_axis_name = None
self.__y_axis = None
self.__y_axis_name = None
self._plot = Plot2D(self)
self._plot.setDefaultColormap(
Colormap(name="viridis",
vmin=None,
vmax=None,
normalization=Colormap.LINEAR))
self._plot.getIntensityHistogramAction().setVisible(True)
self._plot.setKeepDataAspectRatio(True)
# not closable
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self._selector.selectionChanged.connect(self._updateImage)
self._auxSigSlider = HorizontalSliderWithBrowser(parent=self)
self._auxSigSlider.setMinimum(0)
self._auxSigSlider.setValue(0)
self._auxSigSlider.valueChanged[int].connect(self._sliderIdxChanged)
self._auxSigSlider.setToolTip("Select auxiliary signals")
layout = qt.QVBoxLayout()
layout.addWidget(self._plot)
layout.addWidget(self._selector)
layout.addWidget(self._auxSigSlider)
self.setLayout(layout)
def _sliderIdxChanged(self, value):
self._updateImage()
def getPlot(self):
"""Returns the plot used for the display
:rtype: Plot2D
"""
return self._plot
def setImageData(self,
signals,
x_axis=None,
y_axis=None,
signals_names=None,
xlabel=None,
ylabel=None,
title=None,
isRgba=False,
xscale=None,
yscale=None):
"""
:param signals: list of n-D datasets, whose last 2 dimensions are used as the
image's values, or list of 3D datasets interpreted as RGBA image.
:param x_axis: 1-D dataset used as the image's x coordinates. If
provided, its lengths must be equal to the length of the last
dimension of ``signal``.
:param y_axis: 1-D dataset used as the image's y. If provided,
its lengths must be equal to the length of the 2nd to last
dimension of ``signal``.
:param signals_names: Names for each image, used as subtitle and legend.
:param xlabel: Label for X axis
:param ylabel: Label for Y axis
:param title: Graph title
:param isRgba: True if data is a 3D RGBA image
:param str xscale: Scale of X axis in (None, 'linear', 'log')
:param str yscale: Scale of Y axis in (None, 'linear', 'log')
"""
self._selector.selectionChanged.disconnect(self._updateImage)
self._auxSigSlider.valueChanged.disconnect(self._sliderIdxChanged)
self.__signals = signals
self.__signals_names = signals_names
self.__x_axis = x_axis
self.__x_axis_name = xlabel
self.__y_axis = y_axis
self.__y_axis_name = ylabel
self.__title = title
self._selector.clear()
if not isRgba:
self._selector.setAxisNames(["Y", "X"])
img_ndim = 2
else:
self._selector.setAxisNames(["Y", "X", "RGB(A) channel"])
img_ndim = 3
self._selector.setData(signals[0])
if len(signals[0].shape) <= img_ndim:
self._selector.hide()
else:
self._selector.show()
self._auxSigSlider.setMaximum(len(signals) - 1)
if len(signals) > 1:
self._auxSigSlider.show()
else:
self._auxSigSlider.hide()
self._auxSigSlider.setValue(0)
self._axis_scales = xscale, yscale
self._updateImage()
self._plot.resetZoom()
self._selector.selectionChanged.connect(self._updateImage)
self._auxSigSlider.valueChanged.connect(self._sliderIdxChanged)
def _updateImage(self):
selection = self._selector.selection()
auxSigIdx = self._auxSigSlider.value()
legend = self.__signals_names[auxSigIdx]
images = [img[selection] for img in self.__signals]
image = images[auxSigIdx]
x_axis = self.__x_axis
y_axis = self.__y_axis
if x_axis is None and y_axis is None:
xcalib = NoCalibration()
ycalib = NoCalibration()
else:
if x_axis is None:
# no calibration
x_axis = numpy.arange(image.shape[1])
elif numpy.isscalar(x_axis) or len(x_axis) == 1:
# constant axis
x_axis = x_axis * numpy.ones((image.shape[1], ))
elif len(x_axis) == 2:
# linear calibration
x_axis = x_axis[0] * numpy.arange(image.shape[1]) + x_axis[1]
if y_axis is None:
y_axis = numpy.arange(image.shape[0])
elif numpy.isscalar(y_axis) or len(y_axis) == 1:
y_axis = y_axis * numpy.ones((image.shape[0], ))
elif len(y_axis) == 2:
y_axis = y_axis[0] * numpy.arange(image.shape[0]) + y_axis[1]
xcalib = ArrayCalibration(x_axis)
ycalib = ArrayCalibration(y_axis)
self._plot.remove(kind=(
"scatter",
"image",
))
if xcalib.is_affine() and ycalib.is_affine():
# regular image
xorigin, xscale = xcalib(0), xcalib.get_slope()
yorigin, yscale = ycalib(0), ycalib.get_slope()
origin = (xorigin, yorigin)
scale = (xscale, yscale)
self._plot.getXAxis().setScale('linear')
self._plot.getYAxis().setScale('linear')
self._plot.addImage(image,
legend=legend,
origin=origin,
scale=scale,
replace=True,
resetzoom=False)
else:
xaxisscale, yaxisscale = self._axis_scales
if xaxisscale is not None:
self._plot.getXAxis().setScale('log' if xaxisscale ==
'log' else 'linear')
if yaxisscale is not None:
self._plot.getYAxis().setScale('log' if yaxisscale ==
'log' else 'linear')
scatterx, scattery = numpy.meshgrid(x_axis, y_axis)
# fixme: i don't think this can handle "irregular" RGBA images
self._plot.addScatter(numpy.ravel(scatterx),
numpy.ravel(scattery),
numpy.ravel(image),
legend=legend)
if self.__title:
title = self.__title
if len(self.__signals_names) > 1:
# Append dataset name only when there is many datasets
title += '\n' + self.__signals_names[auxSigIdx]
else:
title = self.__signals_names[auxSigIdx]
self._plot.setGraphTitle(title)
self._plot.getXAxis().setLabel(self.__x_axis_name)
self._plot.getYAxis().setLabel(self.__y_axis_name)
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self._plot.clear()
class ArrayCurvePlot(qt.QWidget):
"""
Widget for plotting a curve from a multi-dimensional signal array
and a 1D axis array.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last dimension must have the same length as
the axis array.
The widget provides sliders to select indices on the first (n - 1)
dimensions of the signal array, and buttons to add/replace selected
curves to the plot.
This widget also handles simple 2D or 3D scatter plots (third dimension
displayed as colour of points).
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayCurvePlot, self).__init__(parent)
self.__signals = None
self.__signals_names = None
self.__signal_errors = None
self.__axis = None
self.__axis_name = None
self.__x_axis_errors = None
self.__values = None
self._plot = Plot1D(self)
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self.__selector_is_connected = False
self._plot.sigActiveCurveChanged.connect(
self._setYLabelFromActiveLegend)
layout = qt.QVBoxLayout()
layout.setContentsMargins(0, 0, 0, 0)
layout.addWidget(self._plot)
layout.addWidget(self._selector)
self.setLayout(layout)
def getPlot(self):
"""Returns the plot used for the display
:rtype: Plot1D
"""
return self._plot
def setCurvesData(self,
ys,
x=None,
yerror=None,
xerror=None,
ylabels=None,
xlabel=None,
title=None,
xscale=None,
yscale=None):
"""
:param List[ndarray] ys: List of arrays to be represented by the y (vertical) axis.
It can be multiple n-D array whose last dimension must
have the same length as x (and values must be None)
:param ndarray x: 1-D dataset used as the curve's x values. If provided,
its lengths must be equal to the length of the last dimension of
``y`` (and equal to the length of ``value``, for a scatter plot).
:param ndarray yerror: Single array of errors for y (same shape), or None.
There can only be one array, and it applies to the first/main y
(no y errors for auxiliary_signals curves).
:param ndarray xerror: 1-D dataset of errors for x, or None
:param str ylabels: Labels for each curve's Y axis
:param str xlabel: Label for X axis
:param str title: Graph title
:param str xscale: Scale of X axis in (None, 'linear', 'log')
:param str yscale: Scale of Y axis in (None, 'linear', 'log')
"""
self.__signals = ys
self.__signals_names = ylabels or (["Y"] * len(ys))
self.__signal_errors = yerror
self.__axis = x
self.__axis_name = xlabel
self.__x_axis_errors = xerror
if self.__selector_is_connected:
self._selector.selectionChanged.disconnect(self._updateCurve)
self.__selector_is_connected = False
self._selector.setData(ys[0])
self._selector.setAxisNames(["Y"])
if len(ys[0].shape) < 2:
self._selector.hide()
else:
self._selector.show()
self._plot.setGraphTitle(title or "")
if xscale is not None:
self._plot.getXAxis().setScale('log' if xscale ==
'log' else 'linear')
if yscale is not None:
self._plot.getYAxis().setScale('log' if yscale ==
'log' else 'linear')
self._updateCurve()
if not self.__selector_is_connected:
self._selector.selectionChanged.connect(self._updateCurve)
self.__selector_is_connected = True
def _updateCurve(self):
selection = self._selector.selection()
ys = [sig[selection] for sig in self.__signals]
y0 = ys[0]
len_y = len(y0)
x = self.__axis
if x is None:
x = numpy.arange(len_y)
elif numpy.isscalar(x) or len(x) == 1:
# constant axis
x = x * numpy.ones_like(y0)
elif len(x) == 2 and len_y != 2:
# linear calibration a + b * x
x = x[0] + x[1] * numpy.arange(len_y)
self._plot.remove(kind=("curve", ))
for i in range(len(self.__signals)):
legend = self.__signals_names[i]
# errors only supported for primary signal in NXdata
y_errors = None
if i == 0 and self.__signal_errors is not None:
y_errors = self.__signal_errors[self._selector.selection()]
self._plot.addCurve(x,
ys[i],
legend=legend,
xerror=self.__x_axis_errors,
yerror=y_errors)
if i == 0:
self._plot.setActiveCurve(legend)
self._plot.resetZoom()
self._plot.getXAxis().setLabel(self.__axis_name)
self._plot.getYAxis().setLabel(self.__signals_names[0])
def _setYLabelFromActiveLegend(self, previous_legend, new_legend):
for ylabel in self.__signals_names:
if new_legend is not None and new_legend == ylabel:
self._plot.getYAxis().setLabel(ylabel)
break
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self._plot.clear()
class ArrayCurvePlot(qt.QWidget):
"""
Widget for plotting a curve from a multi-dimensional signal array
and a 1D axis array.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last dimension must have the same length as
the axis array.
The widget provides sliders to select indices on the first (n - 1)
dimensions of the signal array, and buttons to add/replace selected
curves to the plot.
This widget also handles simple 2D or 3D scatter plots (third dimension
displayed as colour of points).
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayCurvePlot, self).__init__(parent)
self.__signals = None
self.__signals_names = None
self.__signal_errors = None
self.__axis = None
self.__axis_name = None
self.__x_axis_errors = None
self.__values = None
self._plot = Plot1D(self)
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self.__selector_is_connected = False
self._plot.sigActiveCurveChanged.connect(self._setYLabelFromActiveLegend)
layout = qt.QVBoxLayout()
layout.setContentsMargins(0, 0, 0, 0)
layout.addWidget(self._plot)
layout.addWidget(self._selector)
self.setLayout(layout)
def getPlot(self):
"""Returns the plot used for the display
:rtype: Plot1D
"""
return self._plot
def setCurvesData(self, ys, x=None,
yerror=None, xerror=None,
ylabels=None, xlabel=None, title=None):
"""
:param List[ndarray] ys: List of arrays to be represented by the y (vertical) axis.
It can be multiple n-D array whose last dimension must
have the same length as x (and values must be None)
:param ndarray x: 1-D dataset used as the curve's x values. If provided,
its lengths must be equal to the length of the last dimension of
``y`` (and equal to the length of ``value``, for a scatter plot).
:param ndarray yerror: Single array of errors for y (same shape), or None.
There can only be one array, and it applies to the first/main y
(no y errors for auxiliary_signals curves).
:param ndarray xerror: 1-D dataset of errors for x, or None
:param str ylabels: Labels for each curve's Y axis
:param str xlabel: Label for X axis
:param str title: Graph title
"""
self.__signals = ys
self.__signals_names = ylabels or (["Y"] * len(ys))
self.__signal_errors = yerror
self.__axis = x
self.__axis_name = xlabel
self.__x_axis_errors = xerror
if self.__selector_is_connected:
self._selector.selectionChanged.disconnect(self._updateCurve)
self.__selector_is_connected = False
self._selector.setData(ys[0])
self._selector.setAxisNames(["Y"])
if len(ys[0].shape) < 2:
self._selector.hide()
else:
self._selector.show()
self._plot.setGraphTitle(title or "")
self._updateCurve()
if not self.__selector_is_connected:
self._selector.selectionChanged.connect(self._updateCurve)
self.__selector_is_connected = True
def _updateCurve(self):
selection = self._selector.selection()
ys = [sig[selection] for sig in self.__signals]
y0 = ys[0]
len_y = len(y0)
x = self.__axis
if x is None:
x = numpy.arange(len_y)
elif numpy.isscalar(x) or len(x) == 1:
# constant axis
x = x * numpy.ones_like(y0)
elif len(x) == 2 and len_y != 2:
# linear calibration a + b * x
x = x[0] + x[1] * numpy.arange(len_y)
self._plot.remove(kind=("curve",))
for i in range(len(self.__signals)):
legend = self.__signals_names[i]
# errors only supported for primary signal in NXdata
y_errors = None
if i == 0 and self.__signal_errors is not None:
y_errors = self.__signal_errors[self._selector.selection()]
self._plot.addCurve(x, ys[i], legend=legend,
xerror=self.__x_axis_errors,
yerror=y_errors)
if i == 0:
self._plot.setActiveCurve(legend)
self._plot.resetZoom()
self._plot.getXAxis().setLabel(self.__axis_name)
self._plot.getYAxis().setLabel(self.__signals_names[0])
def _setYLabelFromActiveLegend(self, previous_legend, new_legend):
for ylabel in self.__signals_names:
if new_legend is not None and new_legend == ylabel:
self._plot.getYAxis().setLabel(ylabel)
break
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self._plot.clear()
class ArrayStackPlot(qt.QWidget):
"""
Widget for plotting a n-D array (n >= 3) as a stack of images.
Three axis arrays can be provided to calibrate the axes.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last 3 dimensions must have the same length as
the axes arrays.
Sliders are provided to select indices on the first (n - 3) dimensions of
the signal array, and the plot is updated to load the stack corresponding
to the selection.
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayStackPlot, self).__init__(parent)
self.__signal = None
self.__signal_name = None
# the Z, Y, X axes apply to the last three dimensions of the signal
# (in that order)
self.__z_axis = None
self.__z_axis_name = None
self.__y_axis = None
self.__y_axis_name = None
self.__x_axis = None
self.__x_axis_name = None
self._stack_view = StackView(self)
self._hline = qt.QFrame(self)
self._hline.setFrameStyle(qt.QFrame.HLine)
self._hline.setFrameShadow(qt.QFrame.Sunken)
self._legend = qt.QLabel(self)
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self.__selector_is_connected = False
layout = qt.QVBoxLayout()
layout.addWidget(self._stack_view)
layout.addWidget(self._hline)
layout.addWidget(self._legend)
layout.addWidget(self._selector)
self.setLayout(layout)
def getStackView(self):
"""Returns the plot used for the display
:rtype: StackView
"""
return self._stack_view
def setStackData(self, signal,
x_axis=None, y_axis=None, z_axis=None,
signal_name=None,
xlabel=None, ylabel=None, zlabel=None,
title=None):
"""
:param signal: n-D dataset, whose last 3 dimensions are used as the
3D stack values.
:param x_axis: 1-D dataset used as the image's x coordinates. If
provided, its lengths must be equal to the length of the last
dimension of ``signal``.
:param y_axis: 1-D dataset used as the image's y. If provided,
its lengths must be equal to the length of the 2nd to last
dimension of ``signal``.
:param z_axis: 1-D dataset used as the image's z. If provided,
its lengths must be equal to the length of the 3rd to last
dimension of ``signal``.
:param signal_name: Label used in the legend
:param xlabel: Label for X axis
:param ylabel: Label for Y axis
:param zlabel: Label for Z axis
:param title: Graph title
"""
if self.__selector_is_connected:
self._selector.selectionChanged.disconnect(self._updateStack)
self.__selector_is_connected = False
self.__signal = signal
self.__signal_name = signal_name or ""
self.__x_axis = x_axis
self.__x_axis_name = xlabel
self.__y_axis = y_axis
self.__y_axis_name = ylabel
self.__z_axis = z_axis
self.__z_axis_name = zlabel
self._selector.setData(signal)
self._selector.setAxisNames(["Y", "X", "Z"])
self._stack_view.setGraphTitle(title or "")
# by default, the z axis is the image position (dimension not plotted)
self._stack_view.getPlot().getXAxis().setLabel(self.__x_axis_name or "X")
self._stack_view.getPlot().getYAxis().setLabel(self.__y_axis_name or "Y")
self._updateStack()
ndims = len(signal.shape)
self._stack_view.setFirstStackDimension(ndims - 3)
# the legend label shows the selection slice producing the volume
# (only interesting for ndim > 3)
if ndims > 3:
self._selector.setVisible(True)
self._legend.setVisible(True)
self._hline.setVisible(True)
else:
self._selector.setVisible(False)
self._legend.setVisible(False)
self._hline.setVisible(False)
if not self.__selector_is_connected:
self._selector.selectionChanged.connect(self._updateStack)
self.__selector_is_connected = True
@staticmethod
def _get_origin_scale(axis):
"""Assuming axis is a regularly spaced 1D array,
return a tuple (origin, scale) where:
- origin = axis[0]
- scale = (axis[n-1] - axis[0]) / (n -1)
:param axis: 1D numpy array
:return: Tuple (axis[0], (axis[-1] - axis[0]) / (len(axis) - 1))
"""
return axis[0], (axis[-1] - axis[0]) / (len(axis) - 1)
def _updateStack(self):
"""Update displayed stack according to the current axes selector
data."""
stk = self._selector.selectedData()
x_axis = self.__x_axis
y_axis = self.__y_axis
z_axis = self.__z_axis
calibrations = []
for axis in [z_axis, y_axis, x_axis]:
if axis is None:
calibrations.append(NoCalibration())
elif len(axis) == 2:
calibrations.append(
LinearCalibration(y_intercept=axis[0],
slope=axis[1]))
else:
calibrations.append(ArrayCalibration(axis))
legend = self.__signal_name + "["
for sl in self._selector.selection():
if sl == slice(None):
legend += ":, "
else:
legend += str(sl) + ", "
legend = legend[:-2] + "]"
self._legend.setText("Displayed data: " + legend)
self._stack_view.setStack(stk, calibrations=calibrations)
self._stack_view.setLabels(
labels=[self.__z_axis_name,
self.__y_axis_name,
self.__x_axis_name])
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self._stack_view.clear()
class ArrayVolumePlot(qt.QWidget):
"""
Widget for plotting a n-D array (n >= 3) as a 3D scalar field.
Three axis arrays can be provided to calibrate the axes.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last 3 dimensions must have the same length as
the axes arrays.
Sliders are provided to select indices on the first (n - 3) dimensions of
the signal array, and the plot is updated to load the stack corresponding
to the selection.
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayVolumePlot, self).__init__(parent)
self.__signal = None
self.__signal_name = None
# the Z, Y, X axes apply to the last three dimensions of the signal
# (in that order)
self.__z_axis = None
self.__z_axis_name = None
self.__y_axis = None
self.__y_axis_name = None
self.__x_axis = None
self.__x_axis_name = None
from silx.gui.plot3d.ScalarFieldView import ScalarFieldView
from silx.gui.plot3d import SFViewParamTree
self._view = ScalarFieldView(self)
def computeIsolevel(data):
data = data[numpy.isfinite(data)]
if len(data) == 0:
return 0
else:
return numpy.mean(data) + numpy.std(data)
self._view.addIsosurface(computeIsolevel, '#FF0000FF')
# Create a parameter tree for the scalar field view
options = SFViewParamTree.TreeView(self._view)
options.setSfView(self._view)
# Add the parameter tree to the main window in a dock widget
dock = qt.QDockWidget()
dock.setWidget(options)
self._view.addDockWidget(qt.Qt.RightDockWidgetArea, dock)
self._hline = qt.QFrame(self)
self._hline.setFrameStyle(qt.QFrame.HLine)
self._hline.setFrameShadow(qt.QFrame.Sunken)
self._legend = qt.QLabel(self)
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self.__selector_is_connected = False
layout = qt.QVBoxLayout()
layout.addWidget(self._view)
layout.addWidget(self._hline)
layout.addWidget(self._legend)
layout.addWidget(self._selector)
self.setLayout(layout)
def getVolumeView(self):
"""Returns the plot used for the display
:rtype: ScalarFieldView
"""
return self._view
def normalizeComplexData(self, data):
"""
Converts a complex data array to its amplitude, if necessary.
:param data: the data to normalize
:return:
"""
if hasattr(data, "dtype"):
isComplex = numpy.issubdtype(data.dtype, numpy.complexfloating)
else:
isComplex = isinstance(data, numbers.Complex)
if isComplex:
data = numpy.absolute(data)
return data
def setData(self,
signal,
x_axis=None,
y_axis=None,
z_axis=None,
signal_name=None,
xlabel=None,
ylabel=None,
zlabel=None,
title=None):
"""
:param signal: n-D dataset, whose last 3 dimensions are used as the
3D stack values.
:param x_axis: 1-D dataset used as the image's x coordinates. If
provided, its lengths must be equal to the length of the last
dimension of ``signal``.
:param y_axis: 1-D dataset used as the image's y. If provided,
its lengths must be equal to the length of the 2nd to last
dimension of ``signal``.
:param z_axis: 1-D dataset used as the image's z. If provided,
its lengths must be equal to the length of the 3rd to last
dimension of ``signal``.
:param signal_name: Label used in the legend
:param xlabel: Label for X axis
:param ylabel: Label for Y axis
:param zlabel: Label for Z axis
:param title: Graph title
"""
signal = self.normalizeComplexData(signal)
if self.__selector_is_connected:
self._selector.selectionChanged.disconnect(self._updateVolume)
self.__selector_is_connected = False
self.__signal = signal
self.__signal_name = signal_name or ""
self.__x_axis = x_axis
self.__x_axis_name = xlabel
self.__y_axis = y_axis
self.__y_axis_name = ylabel
self.__z_axis = z_axis
self.__z_axis_name = zlabel
self._selector.setData(signal)
self._selector.setAxisNames(["Y", "X", "Z"])
self._view.setAxesLabels(self.__x_axis_name or 'X', self.__y_axis_name
or 'Y', self.__z_axis_name or 'Z')
self._updateVolume()
# the legend label shows the selection slice producing the volume
# (only interesting for ndim > 3)
if signal.ndim > 3:
self._selector.setVisible(True)
self._legend.setVisible(True)
self._hline.setVisible(True)
else:
self._selector.setVisible(False)
self._legend.setVisible(False)
self._hline.setVisible(False)
if not self.__selector_is_connected:
self._selector.selectionChanged.connect(self._updateVolume)
self.__selector_is_connected = True
def _updateVolume(self):
"""Update displayed stack according to the current axes selector
data."""
data = self._selector.selectedData()
x_axis = self.__x_axis
y_axis = self.__y_axis
z_axis = self.__z_axis
offset = []
scale = []
for axis in [x_axis, y_axis, z_axis]:
if axis is None:
calibration = NoCalibration()
elif len(axis) == 2:
calibration = LinearCalibration(y_intercept=axis[0],
slope=axis[1])
else:
calibration = ArrayCalibration(axis)
if not calibration.is_affine():
_logger.warning("Axis has not linear values, ignored")
offset.append(0.)
scale.append(1.)
else:
offset.append(calibration(0))
scale.append(calibration.get_slope())
legend = self.__signal_name + "["
for sl in self._selector.selection():
if sl == slice(None):
legend += ":, "
else:
legend += str(sl) + ", "
legend = legend[:-2] + "]"
self._legend.setText("Displayed data: " + legend)
self._view.setData(data, copy=False)
self._view.setScale(*scale)
self._view.setTranslation(*offset)
self._view.setAxesLabels(self.__x_axis_name, self.__y_axis_name,
self.__z_axis_name)
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self._view.setData(None)
class ArrayImagePlot(qt.QWidget):
"""
Widget for plotting an image from a multi-dimensional signal array
and two 1D axes array.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last two dimensions must have the same length as
the axes arrays.
Sliders are provided to select indices on the first (n - 2) dimensions of
the signal array, and the plot is updated to show the image corresponding
to the selection.
If one or both of the axes does not have regularly spaced values, the
the image is plotted as a coloured scatter plot.
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayImagePlot, self).__init__(parent)
self.__signals = None
self.__signals_names = None
self.__x_axis = None
self.__x_axis_name = None
self.__y_axis = None
self.__y_axis_name = None
self._plot = Plot2D(self)
self._plot.setDefaultColormap(Colormap(name="viridis",
vmin=None, vmax=None,
normalization=Colormap.LINEAR))
self._plot.getIntensityHistogramAction().setVisible(True)
# not closable
self._selector = NumpyAxesSelector(self)
self._selector.setNamedAxesSelectorVisibility(False)
self._selector.selectionChanged.connect(self._updateImage)
self._auxSigSlider = HorizontalSliderWithBrowser(parent=self)
self._auxSigSlider.setMinimum(0)
self._auxSigSlider.setValue(0)
self._auxSigSlider.valueChanged[int].connect(self._sliderIdxChanged)
self._auxSigSlider.setToolTip("Select auxiliary signals")
layout = qt.QVBoxLayout()
layout.addWidget(self._plot)
layout.addWidget(self._selector)
layout.addWidget(self._auxSigSlider)
self.setLayout(layout)
def _sliderIdxChanged(self, value):
self._updateImage()
def getPlot(self):
"""Returns the plot used for the display
:rtype: Plot2D
"""
return self._plot
def setImageData(self, signals,
x_axis=None, y_axis=None,
signals_names=None,
xlabel=None, ylabel=None,
title=None, isRgba=False):
"""
:param signals: list of n-D datasets, whose last 2 dimensions are used as the
image's values, or list of 3D datasets interpreted as RGBA image.
:param x_axis: 1-D dataset used as the image's x coordinates. If
provided, its lengths must be equal to the length of the last
dimension of ``signal``.
:param y_axis: 1-D dataset used as the image's y. If provided,
its lengths must be equal to the length of the 2nd to last
dimension of ``signal``.
:param signals_names: Names for each image, used as subtitle and legend.
:param xlabel: Label for X axis
:param ylabel: Label for Y axis
:param title: Graph title
:param isRgba: True if data is a 3D RGBA image
"""
self._selector.selectionChanged.disconnect(self._updateImage)
self._auxSigSlider.valueChanged.disconnect(self._sliderIdxChanged)
self.__signals = signals
self.__signals_names = signals_names
self.__x_axis = x_axis
self.__x_axis_name = xlabel
self.__y_axis = y_axis
self.__y_axis_name = ylabel
self.__title = title
self._selector.clear()
if not isRgba:
self._selector.setAxisNames(["Y", "X"])
img_ndim = 2
else:
self._selector.setAxisNames(["Y", "X", "RGB(A) channel"])
img_ndim = 3
self._selector.setData(signals[0])
if len(signals[0].shape) <= img_ndim:
self._selector.hide()
else:
self._selector.show()
self._auxSigSlider.setMaximum(len(signals) - 1)
if len(signals) > 1:
self._auxSigSlider.show()
else:
self._auxSigSlider.hide()
self._auxSigSlider.setValue(0)
self._updateImage()
self._selector.selectionChanged.connect(self._updateImage)
self._auxSigSlider.valueChanged.connect(self._sliderIdxChanged)
def _updateImage(self):
selection = self._selector.selection()
auxSigIdx = self._auxSigSlider.value()
legend = self.__signals_names[auxSigIdx]
images = [img[selection] for img in self.__signals]
image = images[auxSigIdx]
x_axis = self.__x_axis
y_axis = self.__y_axis
if x_axis is None and y_axis is None:
xcalib = NoCalibration()
ycalib = NoCalibration()
else:
if x_axis is None:
# no calibration
x_axis = numpy.arange(image.shape[1])
elif numpy.isscalar(x_axis) or len(x_axis) == 1:
# constant axis
x_axis = x_axis * numpy.ones((image.shape[1], ))
elif len(x_axis) == 2:
# linear calibration
x_axis = x_axis[0] * numpy.arange(image.shape[1]) + x_axis[1]
if y_axis is None:
y_axis = numpy.arange(image.shape[0])
elif numpy.isscalar(y_axis) or len(y_axis) == 1:
y_axis = y_axis * numpy.ones((image.shape[0], ))
elif len(y_axis) == 2:
y_axis = y_axis[0] * numpy.arange(image.shape[0]) + y_axis[1]
xcalib = ArrayCalibration(x_axis)
ycalib = ArrayCalibration(y_axis)
self._plot.remove(kind=("scatter", "image",))
if xcalib.is_affine() and ycalib.is_affine():
# regular image
xorigin, xscale = xcalib(0), xcalib.get_slope()
yorigin, yscale = ycalib(0), ycalib.get_slope()
origin = (xorigin, yorigin)
scale = (xscale, yscale)
self._plot.addImage(image, legend=legend,
origin=origin, scale=scale,
replace=True)
else:
scatterx, scattery = numpy.meshgrid(x_axis, y_axis)
# fixme: i don't think this can handle "irregular" RGBA images
self._plot.addScatter(numpy.ravel(scatterx),
numpy.ravel(scattery),
numpy.ravel(image),
legend=legend)
title = ""
if self.__title:
title += self.__title
if not title.strip().endswith(self.__signals_names[auxSigIdx]):
title += "\n" + self.__signals_names[auxSigIdx]
self._plot.setGraphTitle(title)
self._plot.getXAxis().setLabel(self.__x_axis_name)
self._plot.getYAxis().setLabel(self.__y_axis_name)
self._plot.resetZoom()
def clear(self):
old = self._selector.blockSignals(True)
self._selector.clear()
self._selector.blockSignals(old)
self._plot.clear()
class XpadVisualisation(QWidget):
unfoldButtonClicked = pyqtSignal()
def __init__(self):
super(QWidget, self).__init__()
self.layout = QVBoxLayout(self)
self.raw_data = None
self.flatfield_image = None
self.path = None
self.diagram_data_array = []
self.angles = []
# Initialize tab screen
self.tabs = QTabWidget()
self.raw_data_tab = QWidget()
# Create an unfolding data tab
self.unfolded_data_tab = UnfoldingDataTab(self)
self.diagram_tab = QWidget()
self.fitting_data_tab = QWidget()
# Create raw data display tab
self.raw_data_tab.layout = QVBoxLayout(self.raw_data_tab)
self.raw_data_viewer = RawDataViewer(self.raw_data_tab)
# Create diagram plot data tab
self.diagram_tab.layout = QVBoxLayout(self.diagram_tab)
self.diagram_data_plot = Plot1D(self.diagram_tab)
# Create fitting curve tab
self.fitting_data_tab.layout = QVBoxLayout(self.fitting_data_tab)
self.fitting_data_selector = NumpyAxesSelector(self.fitting_data_tab)
self.fitting_data_plot = Plot1D(self.fitting_data_tab)
self.fitting_widget = self.fitting_data_plot.getFitAction()
self.fit_action = FitAction(plot=self.fitting_data_plot,
parent=self.fitting_data_plot)
self.toolbar = QToolBar("New")
# Create automatic fitting tab
self.automatic_fit_tab = FittingDataTab(self)
self.unfolded_data_tab.viewer.get_unfold_with_flatfield_action(
).unfoldWithFlatfieldClicked.connect(self.get_calibration)
self.unfolded_data_tab.viewer.get_unfold_action(
).unfoldClicked.connect(self.get_calibration)
self.unfolded_data_tab.unfoldingFinished.connect(
self.create_diagram_array)
self.init_UI()
def init_UI(self):
self.tabs.resize(400, 300)
# Add tabs
self.tabs.addTab(self.raw_data_tab, "Raw data")
self.tabs.addTab(self.unfolded_data_tab, "Unfolded data")
self.tabs.addTab(self.diagram_tab, "Diffraction diagram")
self.tabs.addTab(self.fitting_data_tab, "Fitted data")
self.tabs.addTab(self.automatic_fit_tab, "Automatic fit")
self.raw_data_tab.layout.addWidget(self.raw_data_viewer)
self.diagram_tab.layout.addWidget(self.diagram_data_plot)
self.diagram_data_plot.setGraphTitle(f"Diagram diffraction")
self.diagram_data_plot.setGraphXLabel("two-theta (°)")
self.diagram_data_plot.setGraphYLabel("intensity")
self.diagram_data_plot.setYAxisLogarithmic(True)
self.fitting_data_selector.setNamedAxesSelectorVisibility(False)
self.fitting_data_selector.setVisible(True)
self.fitting_data_selector.setAxisNames("12")
self.fitting_data_plot.setYAxisLogarithmic(True)
self.fitting_data_plot.setGraphXLabel("two-theta (°)")
self.fitting_data_plot.setGraphYLabel("intensity")
self.fitting_data_plot.getRoiAction().trigger()
self.fitting_widget.setXRangeUpdatedOnZoom(False)
self.toolbar.addAction(self.fit_action)
self.fit_action.setVisible(True)
self.fitting_data_plot.addToolBar(self.toolbar)
self.fitting_data_tab.layout.addWidget(self.fitting_data_plot)
self.fitting_data_tab.layout.addWidget(self.fitting_data_selector)
# Add tabs to widget
self.layout.addWidget(self.tabs)
# self.unfold_timer.timeout.connect(self.unfold_data)
self.fitting_data_selector.selectionChanged.connect(self.fitting_curve)
self.fitting_data_plot.getCurvesRoiWidget().sigROIWidgetSignal.connect(
self.get_roi_list)
self.unfolded_data_tab.viewer.scatter_selector.selectionChanged.connect(
self.synchronize_visualisation)
@pyqtSlot()
def on_click(self):
print("\n")
for currentQTableWidgetItem in self.tableWidget.selectedItems():
print(currentQTableWidgetItem.row(),
currentQTableWidgetItem.column(),
currentQTableWidgetItem.text())
def set_data(self, path: str) -> None:
self.path = path
with File(os.path.join(path), mode='r') as h5file:
self.raw_data = get_dataset(h5file,
DataPath.IMAGE_INTERPRETATION.value)[:]
# We put the raw data in the dataviewer
self.raw_data_viewer.set_movie(self.raw_data, self.flatfield_image)
self.unfolded_data_tab.images = self.raw_data
self.unfolded_data_tab.path = self.path
# We allocate a number of view in the stack of unfolded data and fitting data
self.unfolded_data_tab.viewer.set_stack_slider(self.raw_data.shape[0])
self.fitting_data_selector.setData(
numpy.zeros((self.raw_data.shape[0], 1, 1)))
def set_calibration(self, calibration):
# Check if there is a empty list of coordinate in the direct beam calibration
if not [] in [value for value in calibration.values()]:
self.unfolded_data_tab.set_calibration(calibration)
if self.unfolded_data_tab.images is not None:
self.unfolded_data_tab.start_unfolding()
else:
print("Direct beam not calibrated yet.")
def get_calibration(self):
self.unfoldButtonClicked.emit()
def create_diagram_array(self):
self.diagram_data_array = []
numpy.seterr(divide='ignore', invalid='ignore')
for image in self.unfolded_data_tab.viewer.get_scatter_items():
self.diagram_data_array.append(
extract_diffraction_diagram(
image[0],
image[1],
image[2],
1.0 / self.unfolded_data_tab.geometry["calib"],
-100,
100,
patch_data_flag=True))
self.plot_diagram()
self.automatic_fit_tab.set_data_to_fit(self.diagram_data_array)
self.fitting_data_selector.selectionChanged.emit()
set_plot_limits(self.diagram_data_plot,
self.diagram_data_plot.getActiveCurve())
def plot_diagram(self, images_to_remove=[-1]):
self.diagram_data_plot.setGraphTitle(
f"Diagram diffraction of {self.path.split('/')[-1]}")
for index, curve in enumerate(self.diagram_data_array):
if index not in images_to_remove:
self.diagram_data_plot.addCurve(curve[0],
curve[1],
f'Data of image {index}',
color="#0000FF",
replace=False,
symbol='o')
def get_flatfield(self, flat_img: numpy.ndarray):
self.flatfield_image = flat_img
self.raw_data_viewer.get_action_flatfield().set_flatfield(
self.flatfield_image)
self.unfolded_data_tab.flatfield = flat_img
def synchronize_visualisation(self):
# When user change the unfolded view, it set the raw image to the same frame
self.raw_data_viewer.setFrameNumber(
self.unfolded_data_tab.viewer.scatter_selector.selection()[0])
def fitting_curve(self):
if len(self.diagram_data_array) > 0:
self.clear_plot_fitting_widget()
curve = self.diagram_data_array[
self.fitting_data_selector.selection()[0]]
self.fitting_data_plot.addCurve(curve[0], curve[1], symbol='o')
set_plot_limits(self.fitting_data_plot, curve)
def get_roi_list(self, events: dict):
self.rois_list = list(events["roilist"])
def clear_plot_fitting_widget(self):
self.fitting_data_plot.clear()
self.fitting_data_plot.clearMarkers()
class ArrayCurvePlot(qt.QWidget):
"""
Widget for plotting a curve from a multi-dimensional signal array
and a 1D axis array.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last dimension must have the same length as
the axis array.
The widget provides sliders to select indices on the first (n - 1)
dimensions of the signal array, and buttons to add/replace selected
curves to the plot.
This widget also handles simple 2D or 3D scatter plots (third dimension
displayed as colour of points).
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayCurvePlot, self).__init__(parent)
self.__signal = None
self.__signal_name = None
self.__signal_errors = None
self.__axis = None
self.__axis_name = None
self.__axis_errors = None
self.__values = None
self.__first_curve_added = False
self._plot = Plot1D(self)
self._plot.setDefaultColormap( # for scatters
{"name": "viridis",
"vmin": 0., "vmax": 1., # ignored (autoscale) but mandatory
"normalization": "linear",
"autoscale": True})
self.selectorDock = qt.QDockWidget("Data selector", self._plot)
# not closable
self.selectorDock.setFeatures(qt.QDockWidget.DockWidgetMovable
| qt.QDockWidget.DockWidgetFloatable)
self._selector = NumpyAxesSelector(self.selectorDock)
self._selector.setNamedAxesSelectorVisibility(False)
self.__selector_is_connected = False
self.selectorDock.setWidget(self._selector)
self._plot.addTabbedDockWidget(self.selectorDock)
layout = qt.QGridLayout()
layout.setContentsMargins(0, 0, 0, 0)
layout.addWidget(self._plot, 0, 0)
self.setLayout(layout)
def setCurveData(self,
y,
x=None,
values=None,
yerror=None,
xerror=None,
ylabel=None,
xlabel=None,
title=None):
"""
:param y: dataset to be represented by the y (vertical) axis.
For a scatter, this must be a 1D array and x and values must be
1-D arrays of the same size.
In other cases, it can be a n-D array whose last dimension must
have the same length as x (and values must be None)
:param x: 1-D dataset used as the curve's x values. If provided,
its lengths must be equal to the length of the last dimension of
``y`` (and equal to the length of ``value``, for a scatter plot).
:param values: Values, to be provided for a x-y-value scatter plot.
This will be used to compute the color map and assign colors
to the points.
:param yerror: 1-D dataset of errors for y, or None
:param xerror: 1-D dataset of errors for x, or None
:param ylabel: Label for Y axis
:param xlabel: Label for X axis
:param title: Graph title
"""
self.__signal = y
self.__signal_name = ylabel
self.__signal_errors = yerror
self.__axis = x
self.__axis_name = xlabel
self.__axis_errors = xerror
self.__values = values
if self.__selector_is_connected:
self._selector.selectionChanged.disconnect(self._updateCurve)
self.__selector_is_connected = False
self._selector.setData(y)
self._selector.setAxisNames([ylabel or "Y"])
if len(y.shape) < 2:
self.selectorDock.hide()
else:
self.selectorDock.show()
self._plot.setGraphTitle(title or "")
self._plot.setGraphXLabel(self.__axis_name or "X")
self._plot.setGraphYLabel(self.__signal_name or "Y")
self._updateCurve()
if not self.__selector_is_connected:
self._selector.selectionChanged.connect(self._updateCurve)
self.__selector_is_connected = True
def _updateCurve(self):
y = self._selector.selectedData()
x = self.__axis
if x is None:
x = numpy.arange(len(y))
elif numpy.isscalar(x) or len(x) == 1:
# constant axis
x = x * numpy.ones_like(y)
elif len(x) == 2 and len(y) != 2:
# linear calibration a + b * x
x = x[0] + x[1] * numpy.arange(len(y))
legend = self.__signal_name + "["
for sl in self._selector.selection():
if sl == slice(None):
legend += ":, "
else:
legend += str(sl) + ", "
legend = legend[:-2] + "]"
if self.__signal_errors is not None:
y_errors = self.__signal_errors[self._selector.selection()]
else:
y_errors = None
self._plot.remove(kind=("curve", "scatter"))
# values: x-y-v scatter
if self.__values is not None:
self._plot.addScatter(x,
y,
self.__values,
legend=legend,
xerror=self.__axis_errors,
yerror=y_errors)
# x monotonically increasing: curve
elif numpy.all(numpy.diff(x) > 0):
self._plot.addCurve(x,
y,
legend=legend,
xerror=self.__axis_errors,
yerror=y_errors)
# scatter
else:
self._plot.addScatter(x,
y,
value=numpy.ones_like(y),
legend=legend,
xerror=self.__axis_errors,
yerror=y_errors)
self._plot.resetZoom()
self._plot.setGraphXLabel(self.__axis_name)
self._plot.setGraphYLabel(self.__signal_name)
def clear(self):
self._plot.clear()
class ArrayImagePlot(qt.QWidget):
"""
Widget for plotting an image from a multi-dimensional signal array
and two 1D axes array.
The signal array can have an arbitrary number of dimensions, the only
limitation being that the last two dimensions must have the same length as
the axes arrays.
Sliders are provided to select indices on the first (n - 2) dimensions of
the signal array, and the plot is updated to show the image corresponding
to the selection.
If one or both of the axes does not have regularly spaced values, the
the image is plotted as a coloured scatter plot.
"""
def __init__(self, parent=None):
"""
:param parent: Parent QWidget
"""
super(ArrayImagePlot, self).__init__(parent)
self.__signal = None
self.__signal_name = None
self.__x_axis = None
self.__x_axis_name = None
self.__y_axis = None
self.__y_axis_name = None
self._plot = Plot2D(self)
self._plot.setDefaultColormap({
"name": "viridis",
"vmin": 0.,
"vmax": 1., # ignored (autoscale) but mandatory
"normalization": "linear",
"autoscale": True
})
self.selectorDock = qt.QDockWidget("Data selector", self._plot)
# not closable
self.selectorDock.setFeatures(qt.QDockWidget.DockWidgetMovable
| qt.QDockWidget.DockWidgetFloatable)
self._legend = qt.QLabel(self)
self._selector = NumpyAxesSelector(self.selectorDock)
self._selector.setNamedAxesSelectorVisibility(False)
self.__selector_is_connected = False
layout = qt.QVBoxLayout()
layout.addWidget(self._plot)
layout.addWidget(self._legend)
self.selectorDock.setWidget(self._selector)
self._plot.addTabbedDockWidget(self.selectorDock)
self.setLayout(layout)
def setImageData(self,
signal,
x_axis=None,
y_axis=None,
signal_name=None,
xlabel=None,
ylabel=None,
title=None):
"""
:param signal: n-D dataset, whose last 2 dimensions are used as the
image's values.
:param x_axis: 1-D dataset used as the image's x coordinates. If
provided, its lengths must be equal to the length of the last
dimension of ``signal``.
:param y_axis: 1-D dataset used as the image's y. If provided,
its lengths must be equal to the length of the 2nd to last
dimension of ``signal``.
:param signal_name: Label used in the legend
:param xlabel: Label for X axis
:param ylabel: Label for Y axis
:param title: Graph title
"""
if self.__selector_is_connected:
self._selector.selectionChanged.disconnect(self._updateImage)
self.__selector_is_connected = False
self.__signal = signal
self.__signal_name = signal_name or ""
self.__x_axis = x_axis
self.__x_axis_name = xlabel
self.__y_axis = y_axis
self.__y_axis_name = ylabel
self._selector.setData(signal)
self._selector.setAxisNames([ylabel or "Y", xlabel or "X"])
if len(signal.shape) < 3:
self.selectorDock.hide()
else:
self.selectorDock.show()
self._plot.setGraphTitle(title or "")
self._plot.setGraphXLabel(self.__x_axis_name or "X")
self._plot.setGraphYLabel(self.__y_axis_name or "Y")
self._updateImage()
if not self.__selector_is_connected:
self._selector.selectionChanged.connect(self._updateImage)
self.__selector_is_connected = True
def _updateImage(self):
legend = self.__signal_name + "["
for sl in self._selector.selection():
if sl == slice(None):
legend += ":, "
else:
legend += str(sl) + ", "
legend = legend[:-2] + "]"
self._legend.setText("Displayed data: " + legend)
img = self._selector.selectedData()
x_axis = self.__x_axis
y_axis = self.__y_axis
if x_axis is None and y_axis is None:
xcalib = NoCalibration()
ycalib = NoCalibration()
else:
if x_axis is None:
# no calibration
x_axis = numpy.arange(img.shape[-1])
elif numpy.isscalar(x_axis) or len(x_axis) == 1:
# constant axis
x_axis = x_axis * numpy.ones((img.shape[-1], ))
elif len(x_axis) == 2:
# linear calibration
x_axis = x_axis[0] * numpy.arange(img.shape[-1]) + x_axis[1]
if y_axis is None:
y_axis = numpy.arange(img.shape[-2])
elif numpy.isscalar(y_axis) or len(y_axis) == 1:
y_axis = y_axis * numpy.ones((img.shape[-2], ))
elif len(y_axis) == 2:
y_axis = y_axis[0] * numpy.arange(img.shape[-2]) + y_axis[1]
xcalib = ArrayCalibration(x_axis)
ycalib = ArrayCalibration(y_axis)
self._plot.remove(kind=("scatter", "image"))
if xcalib.is_affine() and ycalib.is_affine():
# regular image
xorigin, xscale = xcalib(0), xcalib.get_slope()
yorigin, yscale = ycalib(0), ycalib.get_slope()
origin = (xorigin, yorigin)
scale = (xscale, yscale)
self._plot.addImage(img, legend=legend, origin=origin, scale=scale)
else:
scatterx, scattery = numpy.meshgrid(x_axis, y_axis)
self._plot.addScatter(numpy.ravel(scatterx),
numpy.ravel(scattery),
numpy.ravel(img),
legend=legend)
self._plot.setGraphXLabel(self.__x_axis_name)
self._plot.setGraphYLabel(self.__y_axis_name)
self._plot.resetZoom()
def clear(self):
self._plot.clear()
class FittingDataTab(QWidget):
def __init__(self, parent, data_to_fit=None):
super().__init__(parent)
self._data_to_fit = data_to_fit
self._fitted_data = []
self.layout = QVBoxLayout(self)
self.automatic_plot = Plot1D(self)
self.fitting_data_selector = NumpyAxesSelector(self)
self.fit = FitManager()
self.fit.addtheory("pearson7",
function=pearson7bg,
parameters=[
'backgr', 'slopeLin', 'amplitude', 'center',
'fwhm', 'exposant'
],
estimate=estimate_pearson7)
"""
self.fitting_widget = self.fitting_data_plot.getFitAction()
self.fit_action = FitAction(plot=self.fitting_data_plot, parent=self.fitting_data_plot)
self.toolbar = QToolBar("New")
"""
self.init_ui()
def init_ui(self):
self.setLayout(self.layout)
self.layout.addWidget(self.automatic_plot)
self.layout.addWidget(self.fitting_data_selector)
self.fitting_data_selector.setNamedAxesSelectorVisibility(False)
self.fitting_data_selector.setVisible(True)
self.fitting_data_selector.setAxisNames("12")
# self.fitting_data_selector.selectionChanged.connect(self.automatic_plot_fit)
def set_data_to_fit(self, data_to_fit):
self._data_to_fit = data_to_fit
self.fitting_data_selector.setData(
numpy.zeros((len(data_to_fit), 1, 1)))
self.start_automatic_fit()
def automatic_fit(self):
if self._data_to_fit is not None:
prominence = max(
self._data_to_fit[0][1][~numpy.isnan(self._data_to_fit[0][1])])
print(prominence)
peaks, _ = find_peaks(self._data_to_fit[0][1],
prominence=prominence / 3.0)
plt.plot(peaks, self._data_to_fit[0][1][peaks], "xr")
plt.plot(self._data_to_fit[0][1])
plt.legend(['Test detection with prominence'])
plt.show()
def plot_fit(self):
if len(self._fitted_data) > 0 and len(self._data_to_fit) > 0:
self.automatic_plot.addCurve(
self._data_to_fit[self.fitting_data_selector.selection()[0]]
[0], self._data_to_fit[self.fitting_data_selector.selection()
[0]][1], 'Data to fit')
self.automatic_plot.addCurve(
self._fitted_data[self.fitting_data_selector.selection()[0]]
[0], self._fitted_data[self.fitting_data_selector.selection()
[0]][1], 'Fitted data')
def start_automatic_fit(self):
self.fit.settheory("pearson7")
print("Start fitting...")
for data in self._data_to_fit:
# copy the data arrays, with all the values even the nan ones to keep the size
# the copy is used to not erase data on arrays ploted in the last panel
x = data[0].copy()
y = data[1].copy()
# get the max of the y array without any nan value (it would be the maximum)
maximum = max(y[~numpy.isnan(y)])
# current maximum / peak we are looking for (for the first iteration it will be the max)
current_maximum = max(y[~numpy.isnan(y)])
try:
cpt_peak = 0
print("Searching peak and fitting it...")
# plot the original curve, were we are going to plot each fitted peak.
self.automatic_plot.addCurve(x, y, "Data to fit")
print("Max : ", maximum, " Current max : ", current_maximum)
# this threshold means that we only want peaks that are at least at 1/4 distance in y axis of the max.
while current_maximum > (maximum / 4.0):
peak = numpy.where(y == current_maximum)[0][0]
left = peak - 35 if peak - 35 > 0 else 0
right = peak + 35 if peak + 35 < len(x) else len(x) - 1
x_peak = x[left:right]
y_peak = y[left:right]
# set the data to fit, i.e only the peak without all the curve
self.fit.setdata(x=x_peak, y=y_peak)
# use the estimate function we made to make a first guess of the parameters of the function that will fit our peak.
self.fit.estimate()
# fit the function.
self.fit.runfit()
# draw the resulted function of the fit.
self.automatic_plot.addCurve(
x_peak,
pearson7bg(
x_peak,
*(param['fitresult']
for param in self.fit.fit_results)),
f"Peak number {cpt_peak}")
self._fitted_data.append(
self.automatic_plot.getActiveCurve())
backgr = self.fit.fit_results[0]['fitresult']
fwhm = self.fit.fit_results[4]['fitresult']
# erase the peak to make it easier to found other peaks.
y[left:right] = statistics.mean(y[~numpy.isnan(y)])
current_maximum = max(y[~numpy.isnan(y)])
cpt_peak += 1
print("new current_max : ", current_maximum)
except (numpy.linalg.LinAlgError, TypeError):
print(
"Singular matrix error: fit is impossible with the given parameters"
)
class UnfoldedDataViewer(QWidget):
def __init__(self, parent):
super().__init__(parent=parent)
self.scatter_view = ScatterView(self)
colormap = Colormap('viridis', normalization='log')
self.scatter_view.setGraphTitle("Stack of unfolded data")
self.scatter_view.setColormap(colormap)
self.plot = self.scatter_view.getPlotWidget()
self.plot.setGraphXLabel("two-theta (°)")
self.plot.setGraphYLabel("psi (°)")
self.plot.setKeepDataAspectRatio(False)
self.plot.setYAxisInverted(True)
self.scatter_selector = NumpyAxesSelector(self)
# Prevent user from changing dimensions for the plot
self.scatter_selector.setNamedAxesSelectorVisibility(False)
self.scatter_selector.setVisible(True)
self.scatter_selector.setAxisNames("12")
self.layout = QVBoxLayout(self)
self.layout.addWidget(self.scatter_view)
self.layout.addWidget(self.scatter_selector)
self.stack = []
self.initial_data_flag = True
self.toolbar = QToolBar("Custom toolbar 1")
self.scatter_view.addToolBar(self.toolbar)
self.action_unfold = Unfold(self.plot, parent=self)
self.action_unfold_with_flatfield = UnfoldWithFlatfield(self.plot,
parent=self)
self.action_save = SaveAction(self.plot, parent=self)
self.toolbar.addAction(self.action_unfold)
self.toolbar.addAction(self.action_unfold_with_flatfield)
self.toolbar.addAction(self.action_save)
self.scatter_selector.selectionChanged.connect(
self.change_displayed_data)
def add_scatter(self, scatter_image: tuple, scatter_factor: int):
# Add an image to the stack. If it is the first, emit the selectionChanged signal to plot the first image
self.stack.append((scatter_image[0][::scatter_factor],
scatter_image[1][::scatter_factor],
scatter_image[2][::scatter_factor]))
if self.initial_data_flag:
self.scatter_selector.selectionChanged.emit()
self.initial_data_flag = False
def set_stack_slider(self, nb_images: int):
# Set the size of the sliderbar that will let the user navigate the images
self.clear_scatter_view()
self.scatter_selector.setData(numpy.zeros((nb_images, 1, 1)))
def change_displayed_data(self):
# If there is at least one unfolded image, clear the view, unpack the data and plot a scatter view of the image
if len(self.stack) > 0:
self.clear_scatter_view()
tth_array, psi_array, intensity = self.stack[
self.scatter_selector.selection()[0]]
self.plot.setGraphXLimits(
min(tth_array) - 0.0,
max(tth_array) + 0.0)
self.plot.setGraphYLimits(
min(psi_array) - 5.0,
max(psi_array) + 5.0)
start = time.time()
self.scatter_view.setData(tth_array,
psi_array,
intensity,
copy=False)
end = time.time()
print("Setting the data took :", (end - start) * 1000.0, " ms")
def clear_scatter_view(self):
self.scatter_view.setData(None, None, None)
def reset_scatter_view(self):
self.clear_scatter_view()
self.stack = []
self.initial_data_flag = True
def get_scatter_item(self, index: int) -> tuple:
return self.stack[index]
def get_scatter_items(self) -> list:
return self.stack
def get_unfold_action(self):
return self.action_unfold
def get_unfold_with_flatfield_action(self):
return self.action_unfold_with_flatfield
