# (C) British Crown Copyright 2017, Met Office
#
# This code is free software: you can redistribute it and/or modify it under
# the terms of the GNU Lesser General Public License as published by the
# Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This code is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Lesser General Public License for more details.
#
import numpy
import iris
import matplotlib
import matplotlib.colors
import scipy
import Meteorographica.utils as utils
# Plot a single field as a standard contour plot
[docs]def plot_contour(ax,pe,**kwargs):
"""Plots a variable as a contour plot.
This is the same as :meth:`matplotlib.axes.Axes.contour`, except that it takes an :class:`iris.cube.Cube` instead of an array of values, and its defaults are chosen for plots of mean-sea-level pressure.
Args:
ax (:obj:`cartopy.mpl.geoaxes.GeoAxes`): Axes on which to draw.
pe (:obj:`iris.cube.Cube`): Variable to plot - must have dimensions 'latitude' and 'longitude'.
Keyword Args:
label (:obj:`bool`): Label contour lines? Defaults to False. If it's 'video' use stablised label locations.
resolution (:obj:`float`): What lat:lon resolution (in degrees) to interpolate pe.data to before plotting. Defaults to None - use original resolution.
scale (:obj:`float`): This function is tuned for data in hPa. For data in Pa, set this to 0.01. Defaults to 1.
colors (see :mod:`matplotlib.colors`) contour line colour. Defaults to 'black'.
linewidths (:obj:`float`): Line width for contour lines. Defaults to 0.5.
alpha (:obj:`float`): Colour alpha blend. Defaults to 1 (opaque).
fontsize (:obj:`int`): Font size for contour labels. Defaults to 12.
zorder (:obj:`float`): Standard matplotlib parameter determining which things are plotted on top (high zorder), and which underneath (low zorder), Defaults to 30.
Returns:
See :meth:`matplotlib.axes.Axes.contour` - also adds the lines to the plot.
|
"""
kwargs.setdefault('raw' ,False)
kwargs.setdefault('label' ,True)
kwargs.setdefault('resolution' ,None)
kwargs.setdefault('scale' ,1.0)
kwargs.setdefault('colors' ,'black')
kwargs.setdefault('alpha' ,1.0)
kwargs.setdefault('linewidths' ,0.5)
kwargs.setdefault('fontsize' ,12)
kwargs.setdefault('levels' ,numpy.arange(870,1050,10))
kwargs.setdefault('zorder' ,30)
if kwargs.get('resolution') is None:
contour_p=pe
else:
plot_cube=utils.dummy_cube(ax,kwargs.get('resolution'))
contour_p = pe.regrid(plot_cube,iris.analysis.Linear())
contour_p.data=contour_p.data*kwargs.get('scale')
lats = contour_p.coord('latitude').points
lons = contour_p.coord('longitude').points
lons,lats = numpy.meshgrid(lons,lats)
CS=ax.contour(lons, lats, contour_p.data,
colors=kwargs.get('colors'),
linewidths=kwargs.get('linewidths'),
alpha=kwargs.get('alpha'),
levels=kwargs.get('levels'),
zorder=kwargs.get('zorder'))
# Label the contours
if kwargs.get('label')=='video':
cl=ax.clabel(CS, inline=1, fontsize=kwargs.get('fontsize'),
manual=make_label_hints(ax,CS),
fmt='%d',zorder=kwargs.get('zorder'))
elif kwargs.get('label'):
cl=ax.clabel(CS, inline=1, fontsize=kwargs.get('fontsize'),
fmt='%d',zorder=kwargs.get('zorder'))
return CS
# Plot a set of fields as a spaghetti plot
[docs]def plot_spaghetti_contour(ax,pe,**kwargs):
"""Plots a multi-contour (spaghetti) plot.
Calls :meth:`plot_pressure_contour` multiple times with sensible defaults colurs and styles
Args:
ax (:obj:`cartopy.mpl.geoaxes.GeoAxes`): Axes on which to draw.
pe (:obj:`iris.cube.Cube`): Variable to plot.- must have dimensions <ensemble_dimension>, 'latitude' and 'longitude'.
Keyword Args:
ensemble_dimension (:obj:`float`): name of the ensemble dimension. Defaults to 'member'.
colors (see :mod:`matplotlib.colors`) contour line colour. Defaults to 'blue'.
linewidths (:obj:`float`): Line width for contour lines. Defaults to 0.2.
label (:obj:`bool`): Label contour lines? Defaults to False.
Other keyword arguments are passed to :meth:`plot_pressure_contour`
Returns:
See :meth:`matplotlib.axes.Axes.contour` - except it's an array, one for each member. Also adds the lines to the plot.
|
"""
kwargs.setdefault('ensemble_dimension','member')
kwargs.setdefault('colors' ,'blue')
kwargs.setdefault('linewidths' ,0.1)
kwargs.setdefault('label' ,False)
CS=[]
for m in pe.coord(kwargs.get('ensemble_dimension')).points:
pe_e=pe.extract(iris.Constraint(member=m))
CS.append(plot_contour(ax,pe_e,**kwargs))
return CS
mean_contour_cmap= matplotlib.colors.LinearSegmentedColormap('mc_cmap',
{'red' : ((0.0, 0.0, 0.0),
(1.0, 0.0, 0.0)),
'green' : ((0.0, 0.3, 0.3),
(1.0, 0.3, 0.3)),
'blue' : ((0.0, 0.0, 0.0),
(1.0, 0.0, 0.0)),
'alpha' : ((0.0, 0.0, 0.0),
(1.0, 0.75, 0.75))})
# Plot ensemble mean contours, using transparency as an uncertainty indicator
[docs]def plot_mean_spread(ax,pe,**kwargs):
"""Plots a variable as a contour plot.
Plots contours of the mean of an ensemble, mark uncertainty by fading out the contours where the ensemble spread is large.
Args:
ax (:obj:`cartopy.mpl.geoaxes.GeoAxes`): Axes on which to draw.
pe (:obj:`iris.cube.Cube`): Variable to plot.- must have dimensions <ensemble_dimension>, 'latitude' and 'longitude'.
Keyword Args:
ensemble_dimension (:obj:`float`): name of the ensemble dimension. Defaults to 'member'.
resolution (:obj:`float`): What lat:lon resolution (in degrees) to interpolate pe.data to before plotting. Defaults to None - use original resolution.
colors (see :mod:`matplotlib.colors`) contour line colour. Defaults to 'black'.
linewidths (:obj:`float`): Line width for contour lines. Defaults to 0.2.
label (:obj:`bool`): Label contour lines? Defaults to False.
cmap (:obj:`matplotlib.colors.LinearSegmentedColormap`): Mapping of pe.data to plot colour. Defaults to blackblack semi-transparent.
threshold (:obj:`float`): Ranges shown are regions where the probability that a contour line passes through the region is greater than this. Defaults to 0.05 (5%).
vmax (:obj:`float`): Show as 'most likely', regions where the prob of a contour is greater than this Defaults to 0.4 (40%).
line_threshold (:obj:`float`): Only draw contours where the local standard deviation is less than this. Defaults to None - draw contours everywhere.
alpha (:obj:`float`): Colour alpha blend. Defaults to 1 (opaque).
fontsize (:obj:`int`): Font size for contour labels. Defaults to 12.
zorder (:obj:`float`): Standard matplotlib parameter determining which things are plotted on top (high zorder), and which underneath (low zorder), Defaults to 40.
label (:obj:`bool`): Label contour lines? Defaults to True.
Returns:
See :meth:`matplotlib.axes.Axes.contour` - also adds the lines to the plot.
|
"""
kwargs.setdefault('label' ,True)
kwargs.setdefault('ensemble_dimension','member')
kwargs.setdefault('resolution' ,None)
kwargs.setdefault('scale' ,1.0)
kwargs.setdefault('cmap' ,mean_contour_cmap)
kwargs.setdefault('colors' ,'black')
kwargs.setdefault('alpha' ,1.0)
kwargs.setdefault('linewidths' ,0.5)
kwargs.setdefault('fontsize' ,12)
kwargs.setdefault('levels' ,numpy.arange(870,1050,10))
kwargs.setdefault('threshold' ,0.05)
kwargs.setdefault('vmax' ,0.4)
kwargs.setdefault('line_threshold' ,None)
kwargs.setdefault('zorder' ,40)
pe.data=pe.data*kwargs.get('scale')
pe_m=pe.collapsed(kwargs.get('ensemble_dimension'), iris.analysis.MEAN)
pe_s=pe.collapsed(kwargs.get('ensemble_dimension'), iris.analysis.STD_DEV)
if kwargs.get('resolution') is not None:
plot_cube=utils.dummy_cube(ax,kwargs.get('resolution'))
pe_m=pe_m.regrid(plot_cube,iris.analysis.Linear())
pe_s=pe_s.regrid(plot_cube,iris.analysis.Linear())
# Estimate, at each point, the probability that a contour goes through it.
pe_u = pe_m.copy()
pe_u.data=pe_m.data*0.0
pe_t = pe_u.copy()
for level in kwargs.get('levels'):
pe_t.data=1-scipy.stats.norm.cdf(numpy.absolute(pe_m.data-level)/pe_s.data)
pe_u.data=numpy.maximum(pe_u.data,pe_t.data)
# Plot this probability as a colormap
lats = pe_u.coord('latitude').points
lons = pe_u.coord('longitude').points
u_img=ax.pcolorfast(lons, lats, pe_u.data,
cmap=kwargs.get('cmap'),
vmin=kwargs.get('threshold')/2.0-0.01,
vmax=kwargs.get('vmax'),
zorder=kwargs.get('zorder')-1)
# Generate the mean contour lines, but don't draw them (linewidth=0)
CS=ax.contour(lons, lats, pe_m.data,
colors=kwargs.get('colors'),
linewidths=0,
alpha=kwargs.get('alpha'),
levels=kwargs.get('levels'),
zorder=kwargs.get('zorder'))
# Label the mean contours - transparency dependent on spread
interpolator = iris.analysis.Linear().interpolator(pe_s,
['latitude', 'longitude'])
if kwargs.get('label'):
cl=ax.clabel(CS, inline=1,
fontsize=kwargs.get('fontsize'),
fmt='%d',
zorder=kwargs.get('zorder')+1)
if kwargs.get('line_threshold') is not None:
for label in cl:
pos=label.get_position()
local_spread=interpolator([pos[1],pos[0]]).data
alpha_s=numpy.sqrt(max(0.04,1-local_spread/
kwargs.get('line_threshold')))
label.set_alpha(kwargs.get('alpha')*alpha_s)
# Draw the mean contours, with transparency dependent on spread
base_col=matplotlib.colors.colorConverter.to_rgb(kwargs.get('colors'))
for collection in CS.collections:
segments=collection.get_segments()
for segment in segments:
for idx in range(segment.shape[0]-1):
alpha_s=1
if kwargs.get('line_threshold') is not None:
local_spread=interpolator(
[(segment[idx,1]+segment[idx+1,1])/2.0,
(segment[idx,0]+segment[idx+1,0])/2.0]).data
alpha_s=numpy.sqrt(max(0.04,1-local_spread/
kwargs.get('line_threshold')))
clr=(base_col[0],
base_col[1],
base_col[2],kwargs.get('alpha')*alpha_s)
ax.add_line(matplotlib.lines.Line2D(
xdata=segment[idx:(idx+2),0],
ydata=segment[idx:(idx+2),1],
linestyle='solid',
linewidth=kwargs.get('linewidths'),
color=clr,
zorder=kwargs.get('zorder')))
return CS
# Plot pressure
[docs]def plot(ax,pe,**kwargs):
"""Plot pressure.
Generic function for plotting pressure. Use the 'type' argument to choose the plot style.
Args:
ax (:obj:`cartopy.mpl.geoaxes.GeoAxes`): Axes on which to draw.
pe (:obj:`iris.cube.Cube`): Variable to plot - must have dimensions 'latitude' and 'longitude'.
Keyword Args:
type (:obj:`str`): Style to plot. Options are:'contour' (default), which delegates plotting to :meth:`plot_contour`, 'spaghetti', which delegates plotting to :meth:`plot_spaghetti_contour`, 'spread', which delegates plotting to :meth:`plot_mean_spread`.
Other keyword arguments are passed to the style-specific plotting function.
Returns:
See :meth:`matplotlib.axes.Axes.contour` - Also adds the lines to the plot.
|
"""
kwargs.setdefault('type','contour')
if kwargs.get('type')=='contour':
return plot_contour(ax,pe,**kwargs)
if kwargs.get('type')=='spaghetti':
return plot_spaghetti_contour(ax,pe,**kwargs)
if kwargs.get('type')=='spread':
return plot_mean_spread(ax,pe,**kwargs)
raise Exception('Unsupported pressure plot type %s' %
kwargs.get('type'))
[docs]def make_label_hints(ax,CS):
"""Make hints for the contour label placement algorithm - these stabilise the positions of the contour labels between frames in videos. They don't eliminate the problem of jittery and flickering contour labels, but they do help.
Args:
ax (:obj:`cartopy.mpl.geoaxes.GeoAxes`): Axes on which to draw.
CS (:obj:`matplotlib.contour.ContourSet`): Contours to be labeled.
Returns:
iterable of (x,y) tuples - each a position hint for a label.
|
"""
label_locations=[]
buffer=(ax.get_extent()[1]-ax.get_extent()[0])/20.0
for collection in CS.collections:
segments=collection.get_segments()
for segment in segments:
left_edge=None
right_edge=None
bottom_edge=None
top_edge=None
bottom_point=None
# Find a suitable spot to label the segment
for si in range(len(segment)):
if (left_edge is None) and (si>0):
if ((segment[si][0]>=(ax.get_extent()[0]+buffer)) and
(segment[si-1][0]<(ax.get_extent()[0]+buffer))):
weight=((segment[si][0]-(ax.get_extent()[0]+buffer))/
(segment[si][0]-segment[si-1][0]))
left_edge=(ax.get_extent()[0]+buffer,
segment[si][1]+
(segment[si][1]-segment[si-1][1])*weight)
if (left_edge is None) and (si<(len(segment)-1)):
if ((segment[si][0]>=(ax.get_extent()[0]+buffer)) and
(segment[si+1][0]<(ax.get_extent()[0]+buffer))):
weight=((segment[si][0]-(ax.get_extent()[0]+buffer))/
(segment[si][0]-segment[si+1][0]))
left_edge=(ax.get_extent()[0]+buffer,
segment[si][1]+
(segment[si][1]-segment[si+1][1])*weight)
if (right_edge is None) and (si>0):
if ((segment[si][0]>=(ax.get_extent()[1]-buffer)) and
(segment[si-1][0]<(ax.get_extent()[1]-buffer))):
weight=((segment[si][0]-(ax.get_extent()[1]-buffer))/
(segment[si][0]-segment[si-1][0]))
right_edge=(ax.get_extent()[1]-buffer,
segment[si][1]+
(segment[si][1]-segment[si-1][1])*weight)
if (right_edge is None) and (si<(len(segment)-1)):
if ((segment[si][0]>=(ax.get_extent()[1]-buffer)) and
(segment[si+1][0]<(ax.get_extent()[1]-buffer))):
weight=((segment[si][0]-(ax.get_extent()[1]-buffer))/
(segment[si][0]-segment[si+1][0]))
right_edge=(ax.get_extent()[1]-buffer,
segment[si][1]+
(segment[si][1]-segment[si+1][1])*weight)
if (bottom_edge is None) and (si>0):
if ((segment[si][1]>=(ax.get_extent()[2]+buffer)) and
(segment[si-1][1]<(ax.get_extent()[2]+buffer))):
weight=((segment[si][1]-(ax.get_extent()[2]+buffer))/
(segment[si][1]-segment[si-1][1]))
bottom_edge=(segment[si][0]+
(segment[si][0]-segment[si-1][0])*weight,
ax.get_extent()[2]+buffer)
if (bottom_edge is None) and (si<(len(segment)-1)):
if ((segment[si][1]>=(ax.get_extent()[2]+buffer)) and
(segment[si+1][1]<(ax.get_extent()[2]+buffer))):
weight=((segment[si][1]-(ax.get_extent()[2]+buffer))/
(segment[si][1]-segment[si+1][1]))
bottom_edge=(segment[si][0]+
(segment[si][0]-segment[si+1][0])*weight,
ax.get_extent()[2]+buffer)
if (top_edge is None) and (si>0):
if ((segment[si][1]>=(ax.get_extent()[3]-buffer)) and
(segment[si-1][1]<(ax.get_extent()[3]-buffer))):
weight=((segment[si][1]-(ax.get_extent()[3]-buffer))/
(segment[si][1]-segment[si-1][1]))
top_edge=(segment[si][0]+
(segment[si][0]-segment[si-1][0])*weight,
ax.get_extent()[3]-buffer)
if (top_edge is None) and (si<(len(segment)-1)):
if ((segment[si][1]>=(ax.get_extent()[3]-buffer)) and
(segment[si+1][1]<(ax.get_extent()[3]-buffer))):
weight=((segment[si][1]-(ax.get_extent()[3]-buffer))/
(segment[si][1]-segment[si+1][1]))
top_edge=(segment[si][0]+
(segment[si][0]-segment[si+1][0])*weight,
ax.get_extent()[3]-buffer)
if bottom_point is None or segment[si][1]<bottom_point[1]:
bottom_point=(segment[si][0],segment[si][1])
if left_edge is not None:
label_locations.append(left_edge)
if right_edge is not None:
label_locations.append(right_edge)
if left_edge is None and right_edge is None:
if bottom_edge is not None:
label_locations.append(bottom_edge)
if top_edge is not None:
label_locations.append(top_edge)
if bottom_edge is None and top_edge is None:
label_locations.append(bottom_point)
return label_locations
