Files
aesthetic-immich-starter/backend/.venv/Lib/site-packages/matplotlib/container.py
T

256 lines
7.9 KiB
Python

from matplotlib import cbook
from matplotlib.artist import Artist
class Container(tuple):
"""
Base class for containers.
Containers are classes that collect semantically related Artists such as
the bars of a bar plot.
"""
def __repr__(self):
return f"<{type(self).__name__} object of {len(self)} artists>"
def __new__(cls, *args, **kwargs):
return tuple.__new__(cls, args[0])
def __init__(self, kl, label=None):
self._callbacks = cbook.CallbackRegistry(signals=["pchanged"])
self._remove_method = None
self._label = str(label) if label is not None else None
def remove(self):
for c in cbook.flatten(
self, scalarp=lambda x: isinstance(x, Artist)):
if c is not None:
c.remove()
if self._remove_method:
self._remove_method(self)
def get_children(self):
return [child for child in cbook.flatten(self) if child is not None]
get_label = Artist.get_label
set_label = Artist.set_label
add_callback = Artist.add_callback
remove_callback = Artist.remove_callback
pchanged = Artist.pchanged
class BarContainer(Container):
"""
Container for the artists of bar plots (e.g. created by `.Axes.bar`).
The container can be treated as a tuple of the *patches* themselves.
Additionally, you can access these and further parameters by the
attributes.
Attributes
----------
patches : list of :class:`~matplotlib.patches.Rectangle`
The artists of the bars.
errorbar : None or :class:`~matplotlib.container.ErrorbarContainer`
A container for the error bar artists if error bars are present.
*None* otherwise.
datavalues : None or array-like
The underlying data values corresponding to the bars.
orientation : {'vertical', 'horizontal'}, default: None
If 'vertical', the bars are assumed to be vertical.
If 'horizontal', the bars are assumed to be horizontal.
"""
def __init__(self, patches, errorbar=None, *, datavalues=None,
orientation=None, **kwargs):
self.patches = patches
self.errorbar = errorbar
self.datavalues = datavalues
self.orientation = orientation
super().__init__(patches, **kwargs)
@property
def bottoms(self):
"""
Return the values at the lower end of the bars.
.. versionadded:: 3.11
"""
if self.orientation == 'vertical':
return [p.get_y() for p in self.patches]
elif self.orientation == 'horizontal':
return [p.get_x() for p in self.patches]
else:
raise ValueError("orientation must be 'vertical' or 'horizontal'.")
@property
def tops(self):
"""
Return the values at the upper end of the bars.
.. versionadded:: 3.11
"""
if self.orientation == 'vertical':
return [p.get_y() + p.get_height() for p in self.patches]
elif self.orientation == 'horizontal':
return [p.get_x() + p.get_width() for p in self.patches]
else:
raise ValueError("orientation must be 'vertical' or 'horizontal'.")
@property
def position_centers(self):
"""
Return the centers of bar positions.
.. versionadded:: 3.11
"""
if self.orientation == 'vertical':
return [p.get_x() + p.get_width() / 2 for p in self.patches]
elif self.orientation == 'horizontal':
return [p.get_y() + p.get_height() / 2 for p in self.patches]
else:
raise ValueError("orientation must be 'vertical' or 'horizontal'.")
class ErrorbarContainer(Container):
"""
Container for the artists of error bars (e.g. created by `.Axes.errorbar`).
The container can be treated as the *lines* tuple itself.
Additionally, you can access these and further parameters by the
attributes.
Attributes
----------
lines : tuple
Tuple of ``(data_line, caplines, barlinecols)``.
- data_line : A `~matplotlib.lines.Line2D` instance of x, y plot markers
and/or line.
- caplines : A tuple of `~matplotlib.lines.Line2D` instances of the error
bar caps.
- barlinecols : A tuple of `~matplotlib.collections.LineCollection` with the
horizontal and vertical error ranges.
has_xerr, has_yerr : bool
``True`` if the errorbar has x/y errors.
"""
def __init__(self, lines, has_xerr=False, has_yerr=False, **kwargs):
self.lines = lines
self.has_xerr = has_xerr
self.has_yerr = has_yerr
super().__init__(lines, **kwargs)
class PieContainer:
"""
Container for the artists of pie charts (e.g. created by `.Axes.pie`).
.. versionadded:: 3.11
.. warning::
The class name ``PieContainer`` name is provisional and may change in future
to reflect development of its functionality.
You can access the wedge patches and further parameters by the attributes.
Attributes
----------
wedges : list of `~matplotlib.patches.Wedge`
The artists of the pie wedges.
values : `numpy.ndarray`
The data that the pie is based on.
fracs : `numpy.ndarray`
The fraction of the pie that each wedge represents.
texts : list of list of `~matplotlib.text.Text`
The artists of any labels on the pie wedges. Each inner list has one
text label per wedge.
"""
def __init__(self, wedges, values, normalize):
self.wedges = wedges
self._texts = []
self._values = values
self._normalize = normalize
@property
def texts(self):
# Only return non-empty sublists. An empty sublist may have been added
# for backwards compatibility of the Axes.pie return value (see __getitem__).
return [t_list for t_list in self._texts if t_list]
@property
def values(self):
result = self._values.copy()
result.flags.writeable = False
return result
@property
def fracs(self):
if self._normalize:
result = self._values / self._values.sum()
else:
result = self._values
result.flags.writeable = False
return result
def add_texts(self, texts):
"""Add a list of `~matplotlib.text.Text` objects to the container."""
self._texts.append(texts)
def remove(self):
"""Remove all wedges and texts from the axes"""
for artist_list in self.wedges, self._texts:
for artist in cbook.flatten(artist_list):
artist.remove()
def __getitem__(self, key):
# needed to support unpacking into a tuple for backward compatibility of the
# Axes.pie return value
return (self.wedges, *self._texts)[key]
class StemContainer(Container):
"""
Container for the artists created in a :meth:`.Axes.stem` plot.
The container can be treated like a namedtuple ``(markerline, stemlines,
baseline)``.
Attributes
----------
markerline : `~matplotlib.lines.Line2D`
The artist of the markers at the stem heads.
stemlines : `~matplotlib.collections.LineCollection`
The artists of the vertical lines for all stems.
baseline : `~matplotlib.lines.Line2D`
The artist of the horizontal baseline.
"""
def __init__(self, markerline_stemlines_baseline, **kwargs):
"""
Parameters
----------
markerline_stemlines_baseline : tuple
Tuple of ``(markerline, stemlines, baseline)``.
``markerline`` contains the `.Line2D` of the markers,
``stemlines`` is a `.LineCollection` of the main lines,
``baseline`` is the `.Line2D` of the baseline.
"""
markerline, stemlines, baseline = markerline_stemlines_baseline
self.markerline = markerline
self.stemlines = stemlines
self.baseline = baseline
super().__init__(markerline_stemlines_baseline, **kwargs)