def _get_data_for_plot(axes, workspace, kwargs, with_dy=False, with_dx=False):
if isinstance(workspace, mantid.dataobjects.MDHistoWorkspace):
(normalization, kwargs) = get_normalization(workspace, **kwargs)
indices, kwargs = get_indices(workspace, **kwargs)
(x, y, dy) = get_md_data1d(workspace, normalization, indices)
dx = None
axis = None
else:
axis = MantidAxType(kwargs.pop("axis", MantidAxType.SPECTRUM))
normalize_by_bin_width, kwargs = get_normalize_by_bin_width(
workspace, axes, **kwargs)
workspace_index, distribution, kwargs = get_wksp_index_dist_and_label(
workspace, axis, **kwargs)
if axis == MantidAxType.BIN:
# Overwrite any user specified xlabel
axes.set_xlabel("Spectrum")
x, y, dy, dx = get_bins(workspace, workspace_index, with_dy)
elif axis == MantidAxType.SPECTRUM:
x, y, dy, dx = get_spectrum(workspace, workspace_index,
normalize_by_bin_width, with_dy,
with_dx)
else:
raise ValueError(
"Axis {} is not a valid axis number.".format(axis))
indices = None
return x, y, dy, dx, indices, axis, kwargs
def _plot_impl(axes, workspace, args, kwargs):
"""
Compute data and labels for plot. Used by workspace
replacement handlers to recompute data. See plot for
argument details
"""
if 'LogName' in kwargs:
(x, y, FullTime, LogName, units,
kwargs) = get_sample_log(workspace, **kwargs)
axes.set_ylabel('{0} ({1})'.format(LogName, units))
axes.set_xlabel('Time (s)')
if FullTime:
axes.xaxis_date()
axes.xaxis.set_major_formatter(
mdates.DateFormatter('%H:%M:%S\n%b-%d'))
axes.set_xlabel('Time')
kwargs['drawstyle'] = 'steps-post'
else:
normalize_by_bin_width, kwargs = get_normalize_by_bin_width(
workspace, axes, **kwargs)
x, y, _, _, indices, axis, kwargs = _get_data_for_plot(
axes, workspace, kwargs)
if kwargs.pop('update_axes_labels', True):
_setLabels1D(axes,
workspace,
indices,
normalize_by_bin_width=normalize_by_bin_width,
axis=axis)
kwargs.pop('normalize_by_bin_width', None)
return x, y, args, kwargs
def imshow(axes, workspace, *args, **kwargs):
'''
Essentially the same as :meth:`matplotlib.axes.Axes.imshow`.
:param axes: :class:`matplotlib.axes.Axes` object that will do the plotting
:param workspace: :class:`mantid.api.MatrixWorkspace` or :class:`mantid.api.IMDHistoWorkspace`
to extract the data from
:param distribution: ``None`` (default) asks the workspace. ``False`` means
divide by bin width. ``True`` means do not divide by bin width.
Applies only when the the matrix workspace is a histogram.
:param normalization: ``None`` (default) ask the workspace. Applies to MDHisto workspaces. It can override
the value from displayNormalizationHisto. It checks only if
the normalization is mantid.api.MDNormalization.NumEventsNormalization
:param indices: Specify which slice of an MDHistoWorkspace to use when plotting. Needs to be a tuple
and will be interpreted as a list of indices. You need to use ``slice(None)`` to
select which dimensions to plot. *e.g.* to select the last two axes to plot from a
3D volume use ``indices=(5, slice(None), slice(None))`` where the 5 is the bin selected
for the first axis.
:param slicepoint: Specify which slice of an MDHistoWorkspace to use when plotting in the dimension units.
You need to use ``None`` to select which dimension to plot. *e.g.* to select the last
two axes to plot from a 3D volume use ``slicepoint=(1.0, None, None)`` where the 1.0 is
the value of the dimension selected for the first axis.
:param axisaligned: ``False`` (default). If ``True``, or if the workspace has a variable
number of bins, the polygons will be aligned with the axes
:param transpose: ``bool`` to transpose the x and y axes of the plotted dimensions of an MDHistoWorkspace
'''
transpose = kwargs.pop('transpose', False)
if isinstance(workspace, mantid.dataobjects.MDHistoWorkspace):
(normalization, kwargs) = get_normalization(workspace, **kwargs)
indices, kwargs = get_indices(workspace, **kwargs)
x, y, z = get_md_data2d_bin_bounds(workspace, normalization, indices,
transpose)
_setLabels2D(axes, workspace, indices, transpose)
else:
(aligned, kwargs) = check_resample_to_regular_grid(workspace, **kwargs)
(normalize_by_bin_width,
kwargs) = get_normalize_by_bin_width(workspace, axes, **kwargs)
(distribution, kwargs) = get_distribution(workspace, **kwargs)
if aligned:
(x, y, z) = get_matrix_2d_ragged(workspace,
normalize_by_bin_width,
histogram2D=True,
transpose=transpose)
else:
(x, y, z) = get_matrix_2d_data(workspace,
distribution=distribution,
histogram2D=True,
transpose=transpose)
_setLabels2D(axes,
workspace,
transpose=transpose,
normalize_by_bin_width=normalize_by_bin_width)
if 'extent' not in kwargs:
if x.ndim == 2 and y.ndim == 2:
kwargs['extent'] = [x[0, 0], x[0, -1], y[0, 0], y[-1, 0]]
else:
kwargs['extent'] = [x[0], x[-1], y[0], y[-1]]
return mantid.plots.modest_image.imshow(axes, z, *args, **kwargs)
def set_normalization(self, ws, **kwargs):
normalize_by_bin_width, _ = get_normalize_by_bin_width(
ws, self.ax, **kwargs)
is_normalized = normalize_by_bin_width or ws.isDistribution()
self.presenter.normalization = is_normalized
if is_normalized:
self.norm_opts.setCurrentIndex(1)
else:
self.norm_opts.setCurrentIndex(0)
def set_normalization(self, ws, **kwargs):
normalize_by_bin_width, _ = get_normalize_by_bin_width(ws, self.ax, **kwargs)
is_normalized = normalize_by_bin_width or ws.isDistribution()
if is_normalized:
self.presenter.normalization = mantid.api.MDNormalization.VolumeNormalization
self.norm_opts.setCurrentIndex(1)
else:
self.presenter.normalization = mantid.api.MDNormalization.NoNormalization
self.norm_opts.setCurrentIndex(0)
def errorbar(axes, workspace, *args, **kwargs):
"""
Unpack mantid workspace and render it with matplotlib. ``args`` and
``kwargs`` are passed to :py:meth:`matplotlib.axes.Axes.errorbar` after special
keyword arguments are removed. This will automatically label the
line according to the spectrum number unless specified otherwise.
:param axes: :class:`matplotlib.axes.Axes` object that will do the plotting
:param workspace: :class:`mantid.api.MatrixWorkspace` or :class:`mantid.api.IMDHistoWorkspace`
to extract the data from
:param specNum: spectrum number to plot if MatrixWorkspace
:param wkspIndex: workspace index to plot if MatrixWorkspace
:param distribution: ``None`` (default) asks the global setting. ``False`` means
divide by bin width. ``True`` means do not divide by bin width.
Applies only when the the workspace is a MatrixWorkspace histogram.
:param normalize_by_bin_width: Plot the workspace as a distribution. If None default to global
setting: config['graph1d.autodistribution']
:param normalization: ``None`` (default) ask the workspace. Applies to MDHisto workspaces. It can override
the value from displayNormalizationHisto. It checks only if
the normalization is mantid.api.MDNormalization.NumEventsNormalization
:param axis: Specify which axis will be plotted. Use axis=MantidAxType.BIN to plot a bin,
and axis=MantidAxType.SPECTRUM to plot a spectrum.
The default value is axis=1, plotting spectra by default.
:param indices: Specify which slice of an MDHistoWorkspace to use when plotting. Needs to be a tuple
and will be interpreted as a list of indices. You need to use ``slice(None)`` to
select which dimension to plot. *e.g.* to select the second axis to plot from a
3D volume use ``indices=(5, slice(None), 10)`` where the 5/10 are the bins selected
for the other 2 axes.
:param slicepoint: Specify which slice of an MDHistoWorkspace to use when plotting in the dimension units.
You need to use ``None`` to select which dimension to plot. *e.g.* to select the second
axis to plot from a 3D volume use ``slicepoint=(1.0, None, 2.0)`` where the 1.0/2.0 are
the dimension selected for the other 2 axes.
For matrix workspaces with more than one spectra, either ``specNum`` or ``wkspIndex``
needs to be specified. Giving both will generate a :class:`RuntimeError`. There is no similar
keyword for MDHistoWorkspaces. These type of workspaces have to have exactly one non integrated
dimension
"""
normalize_by_bin_width, kwargs = get_normalize_by_bin_width(
workspace, axes, **kwargs)
x, y, dy, dx, indices, axis, kwargs = _get_data_for_plot(axes,
workspace,
kwargs,
with_dy=True,
with_dx=False)
if kwargs.pop('update_axes_labels', True):
_setLabels1D(axes,
workspace,
indices,
normalize_by_bin_width=normalize_by_bin_width,
axis=axis)
kwargs.pop('normalize_by_bin_width', None)
return axes.errorbar(x, y, dy, dx, *args, **kwargs)
def imshow_sampling(axes,
workspace,
cmap=None,
alpha=None,
vmin=None,
vmax=None,
shape=None,
filternorm=1,
filterrad=4.0,
imlim=None,
url=None,
**kwargs):
"""Copy of imshow but replaced AxesImage with SamplingImage and added
callbacks and Mantid Workspace stuff.
See :meth:`matplotlib.axes.Axes.imshow`
To test:
from mantidqt.widgets.sliceviewer.samplingimage import imshow_sampling
fig, ax = plt.subplots()
im = imshow_sampling(ax, workspace, aspect='auto', origin='lower')
fig.show()
"""
normalize_by_bin_width, kwargs = get_normalize_by_bin_width(workspace, axes, **kwargs)
transpose = kwargs.pop('transpose', False)
extent = kwargs.pop('extent', None)
interpolation = kwargs.pop('interpolation', None)
origin = kwargs.pop('origin', None)
norm = kwargs.pop('norm', None)
resample = kwargs.pop('resample', False)
kwargs.pop('distribution', None)
if not extent:
x0, x1, y0, y1 = (workspace.getDimension(0).getMinimum(),
workspace.getDimension(0).getMaximum(),
workspace.getDimension(1).getMinimum(),
workspace.getDimension(1).getMaximum())
if isinstance(workspace, MatrixWorkspace) and not workspace.isCommonBins():
# for MatrixWorkspace the x extent obtained from dimension 0 corresponds to the first spectrum
# this is not correct in case of ragged workspaces, where we need to obtain the global xmin and xmax
# moreover the axis might be in ascending or descending order, so x[0] is not necessarily the minimum
xmax, xmin = None, None # don't initialise with values from first spectrum as could be a monitor
si = workspace.spectrumInfo()
for i in range(workspace.getNumberHistograms()):
if si.hasDetectors(i) and not si.isMonitor(i):
x_axis = workspace.readX(i)
x_i_first = x_axis[0]
x_i_last = x_axis[-1]
x_i_min = min([x_i_first, x_i_last])
x_i_max = max([x_i_first, x_i_last])
# effectively ignore spectra with nan or inf values
if np.isfinite(x_i_min):
xmin = min([x_i_min, xmin]) if xmin else x_i_min
if np.isfinite(x_i_max):
xmax = max([x_i_max, xmax]) if xmax else x_i_max
x0 = xmin if xmin else x0
x1 = xmax if xmax else x1
if workspace.getDimension(1).getNBins() == workspace.getAxis(1).length():
width = workspace.getDimension(1).getBinWidth()
y0 -= width / 2
y1 += width / 2
if origin == "upper":
y0, y1 = y1, y0
extent = (x0, x1, y0, y1)
if transpose:
e1, e2, e3, e4 = extent
extent = e3, e4, e1, e2
# from matplotlib.axes.Axes.imshow
if norm is not None and not isinstance(norm, matplotlib.colors.Normalize):
raise ValueError("'norm' must be an instance of 'mcolors.Normalize'")
aspect = kwargs.pop('aspect', matplotlib.rcParams['image.aspect'])
axes.set_aspect(aspect)
im = SamplingImage(axes,
workspace,
transpose,
cmap,
norm,
interpolation,
origin,
extent,
filternorm=filternorm,
filterrad=filterrad,
resample=resample,
normalize_by_bin_width=normalize_by_bin_width,
**kwargs)
im._resample_image(100, 100)
im.set_alpha(alpha)
im.set_url(url)
if im.get_clip_path() is None:
# image does not already have clipping set, clip to axes patch
im.set_clip_path(axes.patch)
if vmin is not None or vmax is not None:
if norm is not None and isinstance(norm, matplotlib.colors.LogNorm):
if vmin <= 0:
vmin = 0.0001
if vmax <= 0:
vmax = 1
im.set_clim(vmin, vmax)
else:
im.autoscale_None()
# update ax.dataLim, and, if autoscaling, set viewLim
# to tightly fit the image, regardless of dataLim.
im.set_extent(im.get_extent())
axes.add_image(im)
if extent:
axes.set_xlim(extent[0], extent[1])
axes.set_ylim(extent[2], extent[3])
im.connect_events()
return im
def imshow_sampling(axes,
workspace,
cmap=None,
alpha=None,
vmin=None,
vmax=None,
shape=None,
filternorm=1,
filterrad=4.0,
imlim=None,
url=None,
**kwargs):
"""Copy of imshow but replaced AxesImage with SamplingImage and added
callbacks and Mantid Workspace stuff.
See :meth:`matplotlib.axes.Axes.imshow`
To test:
from mantidqt.widgets.sliceviewer.samplingimage import imshow_sampling
fig, ax = plt.subplots()
im = imshow_sampling(ax, workspace, aspect='auto', origin='lower')
fig.show()
"""
normalize_by_bin_width, kwargs = get_normalize_by_bin_width(
workspace, axes, **kwargs)
transpose = kwargs.pop('transpose', False)
extent = kwargs.pop('extent', None)
interpolation = kwargs.pop('interpolation', None)
origin = kwargs.pop('origin', None)
norm = kwargs.pop('norm', None)
resample = kwargs.pop('resample', False)
kwargs.pop('distribution', None)
if not extent:
x0, x1, y0, y1 = (workspace.getDimension(0).getMinimum(),
workspace.getDimension(0).getMaximum(),
workspace.getDimension(1).getMinimum(),
workspace.getDimension(1).getMaximum())
if workspace.getDimension(1).getNBins() == workspace.getAxis(
1).length():
width = workspace.getDimension(1).getBinWidth()
y0 -= width / 2
y1 += width / 2
if origin == "upper":
y0, y1 = y1, y0
extent = (x0, x1, y0, y1)
if transpose:
e1, e2, e3, e4 = extent
extent = e3, e4, e1, e2
# from matplotlib.axes.Axes.imshow
if norm is not None and not isinstance(norm, matplotlib.colors.Normalize):
raise ValueError("'norm' must be an instance of 'mcolors.Normalize'")
aspect = kwargs.pop('aspect', matplotlib.rcParams['image.aspect'])
axes.set_aspect(aspect)
im = SamplingImage(axes,
workspace,
transpose,
cmap,
norm,
interpolation,
origin,
extent,
filternorm=filternorm,
filterrad=filterrad,
resample=resample,
normalize_by_bin_width=normalize_by_bin_width,
**kwargs)
im._resample_image(100, 100)
im.set_alpha(alpha)
im.set_url(url)
if im.get_clip_path() is None:
# image does not already have clipping set, clip to axes patch
im.set_clip_path(axes.patch)
if vmin is not None or vmax is not None:
if norm is not None and isinstance(norm, matplotlib.colors.LogNorm):
if vmin <= 0:
vmin = 0.0001
if vmax <= 0:
vmax = 1
im.set_clim(vmin, vmax)
else:
im.autoscale_None()
# update ax.dataLim, and, if autoscaling, set viewLim
# to tightly fit the image, regardless of dataLim.
im.set_extent(im.get_extent())
axes.add_image(im)
if extent:
axes.set_xlim(extent[0], extent[1])
axes.set_ylim(extent[2], extent[3])
im.connect_events()
return im
