from collections import OrderedDict
from glue.core.hub import HubListener
import numpy as np
from glue.core import Subset
from echo import delay_callback
from glue.viewers.matplotlib.state import (MatplotlibDataViewerState,
MatplotlibLayerState,
DeferredDrawCallbackProperty as DDCProperty,
DeferredDrawSelectionCallbackProperty as DDSCProperty)
from glue.core.data_combo_helper import ManualDataComboHelper, ComponentIDComboHelper
from glue.utils import defer_draw, avoid_circular
from glue.core.data import BaseData
from glue.core.link_manager import is_convertible_to_single_pixel_cid
from glue.core.exceptions import IncompatibleDataException
from glue.core.message import SubsetUpdateMessage
from glue.core.units import find_unit_choices, UnitConverter
__all__ = ['ProfileViewerState', 'ProfileLayerState']
FUNCTIONS = OrderedDict([('maximum', 'Maximum'),
('minimum', 'Minimum'),
('mean', 'Mean'),
('median', 'Median'),
('sum', 'Sum')])
[docs]class ProfileViewerState(MatplotlibDataViewerState):
"""
A state class that includes all the attributes for a Profile viewer.
"""
x_att_pixel = DDCProperty(docstring='The component ID giving the pixel component '
'shown on the x axis')
x_att = DDSCProperty(docstring='The component ID giving the pixel or world component '
'shown on the x axis')
x_display_unit = DDSCProperty(docstring='The units to use to display the x-axis.')
y_display_unit = DDSCProperty(docstring='The units to use to display the y-axis')
reference_data = DDSCProperty(docstring='The dataset that is used to define the '
'available pixel/world components, and '
'which defines the coordinate frame in '
'which the images are shown')
function = DDSCProperty(docstring='The function to use for collapsing data')
normalize = DDCProperty(False, docstring='Whether to normalize all profiles '
'to the [0:1] range')
# TODO: add function to use
def __init__(self, **kwargs):
super(ProfileViewerState, self).__init__()
self.ref_data_helper = ManualDataComboHelper(self, 'reference_data')
self.add_callback('layers', self._layers_changed)
self.add_callback('reference_data', self._reference_data_changed, echo_old=True)
self.add_callback('x_att', self._update_att)
self.add_callback('x_display_unit', self._convert_units_x_limits, echo_old=True)
self.add_callback('y_display_unit', self._convert_units_y_limits, echo_old=True)
self.add_callback('normalize', self._reset_y_limits)
self.add_callback('function', self._reset_y_limits)
self.x_att_helper = ComponentIDComboHelper(self, 'x_att',
numeric=False, datetime=False, categorical=False,
pixel_coord=True)
ProfileViewerState.function.set_choices(self, list(FUNCTIONS))
ProfileViewerState.function.set_display_func(self, FUNCTIONS.get)
def format_unit(unit):
if unit is None:
return 'Native units'
else:
return unit
ProfileViewerState.x_display_unit.set_display_func(self, format_unit)
ProfileViewerState.y_display_unit.set_display_func(self, format_unit)
self.update_from_dict(kwargs)
def _convert_units_x_limits(self, old_unit, new_unit):
if (
getattr(self, '_previous_x_att', None) is self.x_att and
old_unit != new_unit and
self.reference_data is not None
):
limits = np.array([self.x_min, self.x_max])
converter = UnitConverter()
limits_native = converter.to_native(self.reference_data,
self.x_att, limits,
old_unit)
limits_new = converter.to_unit(self.reference_data,
self.x_att, limits_native,
new_unit)
with delay_callback(self, 'x_min', 'x_max'):
self.x_min, self.x_max = sorted(limits_new)
# Make sure that we keep track of what attribute the limits
# are for - if the attribute changes, we should not try and
# update the limits.
self._previous_x_att = self.x_att
def _convert_units_y_limits(self, old_unit, new_unit):
if old_unit != new_unit:
if old_unit is None or new_unit is None:
self._reset_y_limits()
return
limits = np.array([self.y_min, self.y_max])
converter = UnitConverter()
# We can use any layer, just find the first one that works and
# exit if none of them do.
for layer_state in self.layers:
try:
layer_state.profile
except Exception: # e.g. incompatible subset
continue
else:
break
else:
return
if isinstance(layer_state.layer, Subset):
data = layer_state.layer.data
else:
data = layer_state.layer
limits_native = converter.to_native(data,
layer_state.attribute, limits,
old_unit)
limits_new = converter.to_unit(data,
layer_state.attribute, limits_native,
new_unit)
with delay_callback(self, 'y_min', 'y_max'):
self.y_min, self.y_max = sorted(limits_new)
def _update_combo_ref_data(self):
self.ref_data_helper.set_multiple_data(self.layers_data)
[docs] def reset_limits(self):
with delay_callback(self, 'x_min', 'x_max', 'y_min', 'y_max'):
self._reset_x_limits()
self._reset_y_limits()
@property
def _display_world(self):
return getattr(self.reference_data, 'coords', None) is not None
@defer_draw
def _update_att(self, *args):
if self.x_att is not None:
if self._display_world:
if self.x_att in self.reference_data.pixel_component_ids:
self.x_att_pixel = self.x_att
else:
index = self.reference_data.world_component_ids.index(self.x_att)
self.x_att_pixel = self.reference_data.pixel_component_ids[index]
else:
self.x_att_pixel = self.x_att
self._update_x_display_unit_choices()
self._reset_x_limits()
def _reset_x_limits(self, *event):
# NOTE: we don't use AttributeLimitsHelper because we need to avoid
# trying to get the minimum of *all* the world coordinates in the
# dataset. Instead, we use the same approach as in the layer state below
# and in the case of world coordinates we use online the spine of the
# data.
if self.reference_data is None or self.x_att_pixel is None:
return
data = self.reference_data
if self.x_att in data.pixel_component_ids:
x_min, x_max = -0.5, data.shape[self.x_att.axis] - 0.5
else:
axis = data.world_component_ids.index(self.x_att)
axis_view = [0] * data.ndim
axis_view[axis] = slice(None)
axis_values = data[self.x_att, tuple(axis_view)]
x_min, x_max = np.nanmin(axis_values), np.nanmax(axis_values)
converter = UnitConverter()
x_min, x_max = converter.to_unit(self.reference_data,
self.x_att, np.array([x_min, x_max]),
self.x_display_unit)
with delay_callback(self, 'x_min', 'x_max'):
self.x_min = x_min
self.x_max = x_max
def _reset_y_limits(self, *event):
if self.normalize:
with delay_callback(self, 'y_min', 'y_max'):
self.y_min = -0.1
self.y_max = +1.1
else:
y_min, y_max = np.inf, -np.inf
for layer in self.layers:
try:
profile = layer.profile
except Exception: # e.g. incompatible subset
continue
if profile is not None:
x, y = profile
if len(y) > 0:
y_min = min(y_min, np.nanmin(y))
y_max = max(y_max, np.nanmax(y))
with delay_callback(self, 'y_min', 'y_max'):
if y_max > y_min:
self.y_min = y_min
self.y_max = y_max
else:
self.y_min = 0
self.y_max = 1
[docs] def flip_x(self):
"""
Flip the x_min/x_max limits.
"""
with delay_callback(self, 'x_min', 'x_max'):
self.x_min, self.x_max = self.x_max, self.x_min
@defer_draw
@avoid_circular
def _layers_changed(self, *args):
# By default if any of the state properties change, this triggers a
# callback on anything listening to changes on self.layers - but here
# we just want to know if any layers have been removed/added so we keep
# track of the UUIDs of the layers and check this before continuing.
current_layers = [layer_state.layer.uuid for layer_state in self.layers]
if not hasattr(self, '_last_layers') or self._last_layers != current_layers:
self._update_combo_ref_data()
self._update_y_display_unit_choices()
self._last_layers = current_layers
return
current_attributes = [id(layer_state.attribute) for layer_state in self.layers]
if not hasattr(self, '_last_attributes') or self._last_attributes != current_attributes:
self._update_y_display_unit_choices()
self._last_attributes = current_attributes
def _update_x_display_unit_choices(self):
if self.reference_data is None:
ProfileViewerState.x_display_unit.set_choices(self, [])
return
component = self.reference_data.get_component(self.x_att)
if component.units:
x_choices = find_unit_choices([(self.reference_data, self.x_att, component.units)])
else:
x_choices = ['']
ProfileViewerState.x_display_unit.set_choices(self, x_choices)
self.x_display_unit = component.units
def _update_y_display_unit_choices(self):
component_units = set()
for layer_state in self.layers:
if isinstance(layer_state.layer, BaseData):
component = layer_state.layer.get_component(layer_state.attribute)
if component.units:
component_units.add((layer_state.layer, layer_state.attribute, component.units))
y_choices = [None] + find_unit_choices(component_units)
ProfileViewerState.y_display_unit.set_choices(self, y_choices)
@defer_draw
def _reference_data_changed(self, before=None, after=None):
# A callback event for reference_data is triggered if the choices change
# but the actual selection doesn't - so we avoid resetting the WCS in
# this case.
if before is after:
return
for layer in self.layers:
layer.reset_cache()
# This signal can get emitted if just the choices but not the actual
# reference data change, so we check here that the reference data has
# actually changed
if self.reference_data is not getattr(self, '_last_reference_data', None):
self._last_reference_data = self.reference_data
with delay_callback(self, 'x_att'):
if self.reference_data is None:
self.x_att_helper.set_multiple_data([])
else:
self.x_att_helper.set_multiple_data([self.reference_data])
if self._display_world:
self.x_att_helper.world_coord = True
self.x_att = self.reference_data.world_component_ids[0]
else:
self.x_att_helper.world_coord = False
self.x_att = self.reference_data.pixel_component_ids[0]
self._update_att()
self.reset_limits()
def _update_priority(self, name):
if name == 'layers':
return 2
elif name == 'reference_data':
return 1.5
elif name.endswith(('_min', '_max')):
return 0
else:
return 1
[docs]class ProfileLayerState(MatplotlibLayerState, HubListener):
"""
A state class that includes all the attributes for layers in a Profile plot.
"""
linewidth = DDCProperty(1, docstring='The width of the line')
attribute = DDSCProperty(docstring='The attribute shown in the layer')
v_min = DDCProperty(docstring='The lower level shown')
v_max = DDCProperty(docstring='The upper level shown')
percentile = DDSCProperty(docstring='The percentile value used to '
'automatically calculate levels')
as_steps = DDCProperty(True, docstring='Whether to display the profile as steps')
_viewer_callbacks_set = False
_layer_subset_updates_subscribed = False
_profile_cache = None
def __init__(self, layer=None, viewer_state=None, **kwargs):
super(ProfileLayerState, self).__init__(layer=layer, viewer_state=viewer_state)
self.attribute_att_helper = ComponentIDComboHelper(self, 'attribute',
numeric=True, categorical=False)
percentile_display = {100: 'Min/Max',
99.5: '99.5%',
99: '99%',
95: '95%',
90: '90%',
'Custom': 'Custom'}
ProfileLayerState.percentile.set_choices(self, [100, 99.5, 99, 95, 90, 'Custom'])
ProfileLayerState.percentile.set_display_func(self, percentile_display.get)
self.add_callback('layer', self._on_layer_change, priority=1000)
self.add_callback('visible', self.reset_cache, priority=1000)
if layer is not None:
self._on_layer_change()
self.update_from_dict(kwargs)
def _on_layer_change(self, *args):
if self.layer is not None:
# Set the available attributes
self.attribute_att_helper.set_multiple_data([self.layer])
# We only subscribe to SubsetUpdateMessage the first time that 'layer'
# is not None, and then do any filtering in the callback function.
if not self._layer_subset_updates_subscribed and self.layer.hub is not None:
self.layer.hub.subscribe(self, SubsetUpdateMessage, handler=self._on_subset_update)
self._layer_subset_updates_subscribed = True
self.reset_cache()
def _on_subset_update(self, msg):
if msg.subset is self.layer:
self.reset_cache()
[docs] @property
def independent_x_att(self):
return is_convertible_to_single_pixel_cid(self.layer, self.viewer_state.x_att) is not None
[docs] def normalize_values(self, values):
return (np.asarray(values) - self.v_min) / (self.v_max - self.v_min)
[docs] def reset_cache(self, *args):
self._profile_cache = None
@property
def viewer_state(self):
return self._viewer_state
[docs] @viewer_state.setter
def viewer_state(self, viewer_state):
self._viewer_state = viewer_state
[docs] @property
def profile(self):
self.update_profile()
return self._profile_cache
[docs] def update_profile(self, update_limits=True):
if self._profile_cache is not None:
return self._profile_cache
if not self.visible:
return
if not self._viewer_callbacks_set:
self.viewer_state.add_callback('x_att', self.reset_cache, priority=100000)
self.viewer_state.add_callback('x_display_unit', self.reset_cache, priority=100000)
self.viewer_state.add_callback('y_display_unit', self.reset_cache, priority=100000)
self.viewer_state.add_callback('function', self.reset_cache, priority=100000)
if self.is_callback_property('attribute'):
self.add_callback('attribute', self.reset_cache, priority=100000)
self._viewer_callbacks_set = True
if self.viewer_state is None or self.viewer_state.x_att is None or self.attribute is None:
raise IncompatibleDataException()
# Check what pixel axis in the current dataset x_att corresponds to
pix_cid = is_convertible_to_single_pixel_cid(self.layer, self.viewer_state.x_att_pixel)
if pix_cid is None:
raise IncompatibleDataException()
# If we get here, then x_att does correspond to a single pixel axis in
# the cube, so we now prepare a list of axes to collapse over.
axes = tuple(i for i in range(self.layer.ndim) if i != pix_cid.axis)
# We now get the y values for the data
# TODO: in future we should optimize the case where the mask is much
# smaller than the data to just average the relevant 'spaxels' in the
# data rather than collapsing the whole cube.
if isinstance(self.layer, Subset):
data = self.layer.data
subset_state = self.layer.subset_state
else:
data = self.layer
subset_state = None
profile_values = data.compute_statistic(self.viewer_state.function, self.attribute, axis=axes, subset_state=subset_state)
if np.all(np.isnan(profile_values)):
self._profile_cache = [], []
else:
axis_view = [0] * data.ndim
axis_view[pix_cid.axis] = slice(None)
axis_values = data[self.viewer_state.x_att, tuple(axis_view)]
converter = UnitConverter()
axis_values = converter.to_unit(self.viewer_state.reference_data,
self.viewer_state.x_att, axis_values,
self.viewer_state.x_display_unit)
profile_values = converter.to_unit(data, self.attribute, profile_values,
self.viewer_state.y_display_unit)
self._profile_cache = axis_values, profile_values
if update_limits:
self.update_limits(update_profile=False)
[docs] def update_limits(self, update_profile=True):
with delay_callback(self, 'v_min', 'v_max'):
if update_profile:
self.update_profile(update_limits=False)
if self._profile_cache is not None and len(self._profile_cache[1]) > 0:
self.v_min = np.nanmin(self._profile_cache[1])
self.v_max = np.nanmax(self._profile_cache[1])
