I would like to identify two different symbols (with different colors) on the same line in a legend. Below, I tried doing this with Proxy Artists, but the result is that they get stacked on top of each other in the legend. I want them next to each other or one above the other-- so they are both visible.
from pylab import *
import matplotlib.lines as mlines
#define two colors, one for 'r' data, one for 'a' data
rcolor=[69./255 , 115./255, 50.8/255 ]
acolor=[202./255, 115./255, 50.8/255 ]
#Plot theory:
ax2.plot(rho, g_r, '-',color=rcolor,lw=2)
ax2.plot(rho, g_a, '-',color=acolor,lw=2)
#Plot experiment:
ax2.scatter(X_r, Y_r,s=200, marker='s', facecolors='none', edgecolors=rcolor);
ax2.scatter(X_a, Y_a,s=200, marker='^', facecolors='none', edgecolors=acolor);
#Create Proxy Artists for legend
expt_r = mlines.Line2D([], [], fillstyle='none', color=rcolor, marker='s', linestyle='', markersize=15)
expt_a = mlines.Line2D([], [], fillstyle='none', color=acolor, marker='^', linestyle='', markersize=15)
thry_r = mlines.Line2D([], [], fillstyle='none', color=rcolor, marker='', markersize=15)
thry_a = mlines.Line2D([], [], fillstyle='none', color=acolor, marker='', markersize=15)
#Add legend
ax2.legend(((expt_r,expt_a),(thry_r,thry_a)), ('Experiment','Theory'))
I think my problem is almost exactly like this one: (Matplotlib, legend with multiple different markers with one label), but it seems like the problem is unsolved since the answer there just plots one patch on top of the other, which is exactly what happens for me too. I feel like maybe I need to make a composite patch somehow, but I had trouble finding how to do this. Thanks!
Also, I haven't found how to make the legend symbols look the same (line thickness, size) as the scatter symbols. Thanks again.
The problem of overlapping patches (aka artists) lies in how you have defined the handles and labels when creating the legend. To quote the matplotlib legend guide:
the default handler_map has a special tuple handler (legend_handler.HandlerTuple) which simply plots the handles on top of one another for each item in the given tuple
Let's first examine the structure of the legend you have given as an example. Any iterable object can be used for handles and labels, so I choose to store them as lists, in line with some examples given in the legend guide and to make the code clearer:
ax2.legend([(expt_r, expt_a), (thry_r, thry_a)], ['Experiment', 'Theory'])
handles_list = [(expt_r, expt_a), (thry_r, thry_a)]
handles1 = (expt_r, expt_a) # tuple of 2 handles (aka legend keys) representing the markers
handles2 = (thry_r, thry_a) # same, these represent the lines
labels_list = ['Experiment', 'Theory']
label1 = 'Experiment'
label2 = 'Theory'
Whatever the number of handles contained in handles1 or in handles2, they will all be drawn on top of one another by the corresponding label1 and label2, seeing as they are contained in a single tuple. To solve this issue and have the keys/symbols drawn separately, you must take them out of the tuples like this:
handles_list = [expt_r, expt_a, thry_r, thry_a]
But now you face the issue that only the expt_r, expt_a handles will be drawn because the labels list contains only two labels. Yet the goal here is to avoid needlessly repeating these labels. Here is an example of how to solve this issue. It is built on the code sample you have provided and makes use of the legend parameters:
import numpy as np # v 1.19.2
import matplotlib.pyplot as plt # v 3.3.2
# Set data parameters
rng = np.random.default_rng(seed=1)
data_points = 10
error_scale = 0.2
# Create variables
rho = np.arange(0, data_points)
g_r = rho**2
g_r_error = rng.uniform(-error_scale, error_scale, size=g_r.size)*g_r
g_a = rho**2 + 50
g_a_error = rng.uniform(-error_scale, error_scale, size=g_a.size)*g_a
X_r = rho
Y_r = g_r + g_r_error
X_a = rho
Y_a = g_a + g_a_error
# Define two colors, one for 'r' data, one for 'a' data
rcolor = [69./255 , 115./255, 50.8/255 ]
acolor = [202./255, 115./255, 50.8/255 ]
# Create figure with single axes
fig, ax = plt.subplots(figsize=(9,5))
# Plot theory: notice the comma after each variable to unpack the list
# containing one Line2D object returned by the plotting function
# (because it is possible to plot several lines in one function call)
thry_r, = ax.plot(rho, g_r, '-', color=rcolor, lw=2)
thry_a, = ax.plot(rho, g_a, '-', color=acolor, lw=2)
# Plot experiment: no need for a comma as the PathCollection object
# returned by the plotting function is not contained in a list
expt_r = ax.scatter(X_r, Y_r, s=100, marker='s', facecolors='none', edgecolors=rcolor)
expt_a = ax.scatter(X_a, Y_a, s=100, marker='^', facecolors='none', edgecolors=acolor)
# Create custom legend: input handles and labels in desired order and
# set ncol=2 to line up the legend keys of the same type.
# Note that seeing as the labels are added here with explicitly defined
# handles, it is not necessary to define the labels in the plotting functions.
ax.legend(handles=[thry_r, expt_r, thry_a, expt_a],
labels=['', '', 'Theory','Experiment'],
loc='upper left', ncol=2, handlelength=3, edgecolor='black',
borderpad=0.7, handletextpad=1.5, columnspacing=0)
plt.show()
The problem is solved but the code can be simplified to automate the legend creation. It is possible to avoid storing the output of each plotting function as a new variable by making use of the get_legend_handles_labels function. Here is an example built on the same data. Note that a third type of plot (error band) is added to make the processing of handles and labels more clear:
# Define parameters used to process handles and labels
nb_plot_types = 3 # theory, experiment, error band
nb_experiments = 2 # r and a
# Create figure with single axes
fig, ax = plt.subplots(figsize=(9,5))
# Note that contrary to the previous example, here it is necessary to
# define a label in the plotting functions seeing as the returned
# handles/artists are this time not stored as variables. No labels means
# no handles in the handles list returned by the
# ax.get_legend_handles_labels() function.
# Plot theory
ax.plot(rho, g_r, '-', color=rcolor, lw=2, label='Theory')
ax.plot(rho, g_a, '-', color=acolor, lw=2, label='Theory')
# Plot experiment
ax.scatter(X_r, Y_r, s=100, marker='s', facecolors='none',
edgecolors=rcolor, label='Experiment')
ax.scatter(X_a, Y_a, s=100, marker='^', facecolors='none',
edgecolors=acolor, label='Experiment')
# Plot error band
g_r_lower = g_r - error_scale*g_r
g_r_upper = g_r + error_scale*g_r
ax.fill_between(X_r, g_r_lower, g_r_upper,
color=rcolor, alpha=0.2, label='Uncertainty')
g_a_lower = g_a - error_scale*g_a
g_a_upper = g_a + error_scale*g_a
ax.fill_between(X_a, g_a_lower, g_a_upper,
color=acolor, alpha=0.2, label='Uncertainty')
# Extract handles and labels and reorder/process them for the custom legend,
# based on the number of types of plots and the number of experiments.
# The handles list returned by ax.get_legend_handles_labels() appears to be
# always ordered the same way with lines placed first, followed by collection
# objects in alphabetical order, regardless of the order of the plotting
# functions calls. So if you want to display the legend keys in a different
# order (e.g. put lines on the bottom line) you will have to process the
# handles list in another way.
handles, labels = ax.get_legend_handles_labels()
handles_ordered_arr = np.array(handles).reshape(nb_plot_types, nb_experiments).T
handles_ordered = handles_ordered_arr.flatten()
# Note the use of asterisks to unpack the lists of strings
labels_trimmed = *nb_plot_types*[''], *labels[::nb_experiments]
# Create custom legend with the same format as in the previous example
ax.legend(handles_ordered, labels_trimmed,
loc='upper left', ncol=nb_experiments, handlelength=3, edgecolor='black',
borderpad=0.7, handletextpad=1.5, columnspacing=0)
plt.show()
Additional documentation: legend class, artist class
I do not answer your main question, sorry.
However, regarding your last point
how to make the legend symbols look the same (line thickness, size) as the scatter symbols
you can use the keyword markerscale of the legend command. So for equal size
ax2.legend( ...<handles_and_labels>... , markerscale=1)
or a change of legend.markerscale in the rcParams should do.
Related
I am plotting the same type of information, but for different countries, with multiple subplots with Matplotlib. That is, I have nine plots on a 3x3 grid, all with the same for lines (of course, different values per line).
However, I have not figured out how to put a single legend (since all nine subplots have the same lines) on the figure just once.
How do I do that?
There is also a nice function get_legend_handles_labels() you can call on the last axis (if you iterate over them) that would collect everything you need from label= arguments:
handles, labels = ax.get_legend_handles_labels()
fig.legend(handles, labels, loc='upper center')
figlegend may be what you're looking for: matplotlib.pyplot.figlegend
An example is at Figure legend demo.
Another example:
plt.figlegend(lines, labels, loc = 'lower center', ncol=5, labelspacing=0.)
Or:
fig.legend(lines, labels, loc = (0.5, 0), ncol=5)
TL;DR
lines_labels = [ax.get_legend_handles_labels() for ax in fig.axes]
lines, labels = [sum(lol, []) for lol in zip(*lines_labels)]
fig.legend(lines, labels)
I have noticed that none of the other answers displays an image with a single legend referencing many curves in different subplots, so I have to show you one... to make you curious...
Now, if I've teased you enough, here it is the code
from numpy import linspace
import matplotlib.pyplot as plt
# each Axes has a brand new prop_cycle, so to have differently
# colored curves in different Axes, we need our own prop_cycle
# Note: we CALL the axes.prop_cycle to get an itertoools.cycle
color_cycle = plt.rcParams['axes.prop_cycle']()
# I need some curves to plot
x = linspace(0, 1, 51)
functs = [x*(1-x), x**2*(1-x),
0.25-x*(1-x), 0.25-x**2*(1-x)]
labels = ['$x-x²$', '$x²-x³$',
'$\\frac{1}{4} - (x-x²)$', '$\\frac{1}{4} - (x²-x³)$']
# the plot,
fig, (a1,a2) = plt.subplots(2)
for ax, f, l, cc in zip((a1,a1,a2,a2), functs, labels, color_cycle):
ax.plot(x, f, label=l, **cc)
ax.set_aspect(2) # superfluos, but nice
# So far, nothing special except the managed prop_cycle. Now the trick:
lines_labels = [ax.get_legend_handles_labels() for ax in fig.axes]
lines, labels = [sum(lol, []) for lol in zip(*lines_labels)]
# Finally, the legend (that maybe you'll customize differently)
fig.legend(lines, labels, loc='upper center', ncol=4)
plt.show()
If you want to stick with the official Matplotlib API, this is
perfect, otherwise see note no.1 below (there is a private
method...)
The two lines
lines_labels = [ax.get_legend_handles_labels() for ax in fig.axes]
lines, labels = [sum(lol, []) for lol in zip(*lines_labels)]
deserve an explanation, see note 2 below.
I tried the method proposed by the most up-voted and accepted answer,
# fig.legend(lines, labels, loc='upper center', ncol=4)
fig.legend(*a2.get_legend_handles_labels(),
loc='upper center', ncol=4)
and this is what I've got
Note 1
If you don't mind using a private method of the matplotlib.legend module ... it's really much much much easier
from matplotlib.legend import _get_legend_handles_labels
...
fig.legend(*_get_legend_handles_and_labels(fig.axes), ...)
Note 2
I have encapsulated the two tricky lines in a function, just four lines of code, but heavily commented
def fig_legend(fig, **kwdargs):
# Generate a sequence of tuples, each contains
# - a list of handles (lohand) and
# - a list of labels (lolbl)
tuples_lohand_lolbl = (ax.get_legend_handles_labels() for ax in fig.axes)
# E.g., a figure with two axes, ax0 with two curves, ax1 with one curve
# yields: ([ax0h0, ax0h1], [ax0l0, ax0l1]) and ([ax1h0], [ax1l0])
# The legend needs a list of handles and a list of labels,
# so our first step is to transpose our data,
# generating two tuples of lists of homogeneous stuff(tolohs), i.e.,
# we yield ([ax0h0, ax0h1], [ax1h0]) and ([ax0l0, ax0l1], [ax1l0])
tolohs = zip(*tuples_lohand_lolbl)
# Finally, we need to concatenate the individual lists in the two
# lists of lists: [ax0h0, ax0h1, ax1h0] and [ax0l0, ax0l1, ax1l0]
# a possible solution is to sum the sublists - we use unpacking
handles, labels = (sum(list_of_lists, []) for list_of_lists in tolohs)
# Call fig.legend with the keyword arguments, return the legend object
return fig.legend(handles, labels, **kwdargs)
I recognize that sum(list_of_lists, []) is a really inefficient method to flatten a list of lists, but ① I love its compactness, ② usually is a few curves in a few subplots and ③ Matplotlib and efficiency? ;-)
For the automatic positioning of a single legend in a figure with many axes, like those obtained with subplots(), the following solution works really well:
plt.legend(lines, labels, loc = 'lower center', bbox_to_anchor = (0, -0.1, 1, 1),
bbox_transform = plt.gcf().transFigure)
With bbox_to_anchor and bbox_transform=plt.gcf().transFigure, you are defining a new bounding box of the size of your figureto be a reference for loc. Using (0, -0.1, 1, 1) moves this bounding box slightly downwards to prevent the legend to be placed over other artists.
OBS: Use this solution after you use fig.set_size_inches() and before you use fig.tight_layout()
You just have to ask for the legend once, outside of your loop.
For example, in this case I have 4 subplots, with the same lines, and a single legend.
from matplotlib.pyplot import *
ficheiros = ['120318.nc', '120319.nc', '120320.nc', '120321.nc']
fig = figure()
fig.suptitle('concentration profile analysis')
for a in range(len(ficheiros)):
# dados is here defined
level = dados.variables['level'][:]
ax = fig.add_subplot(2,2,a+1)
xticks(range(8), ['0h','3h','6h','9h','12h','15h','18h','21h'])
ax.set_xlabel('time (hours)')
ax.set_ylabel('CONC ($\mu g. m^{-3}$)')
for index in range(len(level)):
conc = dados.variables['CONC'][4:12,index] * 1e9
ax.plot(conc,label=str(level[index])+'m')
dados.close()
ax.legend(bbox_to_anchor=(1.05, 0), loc='lower left', borderaxespad=0.)
# it will place the legend on the outer right-hand side of the last axes
show()
If you are using subplots with bar charts, with a different colour for each bar, it may be faster to create the artefacts yourself using mpatches.
Say you have four bars with different colours as r, m, c, and k, you can set the legend as follows:
import matplotlib.patches as mpatches
import matplotlib.pyplot as plt
labels = ['Red Bar', 'Magenta Bar', 'Cyan Bar', 'Black Bar']
#####################################
# Insert code for the subplots here #
#####################################
# Now, create an artist for each color
red_patch = mpatches.Patch(facecolor='r', edgecolor='#000000') # This will create a red bar with black borders, you can leave out edgecolor if you do not want the borders
black_patch = mpatches.Patch(facecolor='k', edgecolor='#000000')
magenta_patch = mpatches.Patch(facecolor='m', edgecolor='#000000')
cyan_patch = mpatches.Patch(facecolor='c', edgecolor='#000000')
fig.legend(handles = [red_patch, magenta_patch, cyan_patch, black_patch], labels=labels,
loc="center right",
borderaxespad=0.1)
plt.subplots_adjust(right=0.85) # Adjust the subplot to the right for the legend
To build on top of gboffi's and Ben Usman's answer:
In a situation where one has different lines in different subplots with the same color and label, one can do something along the lines of:
labels_handles = {
label: handle for ax in fig.axes for handle, label in zip(*ax.get_legend_handles_labels())
}
fig.legend(
labels_handles.values(),
labels_handles.keys(),
loc = "upper center",
bbox_to_anchor = (0.5, 0),
bbox_transform = plt.gcf().transFigure,
)
Using Matplotlib 2.2.2, this can be achieved using the gridspec feature.
In the example below, the aim is to have four subplots arranged in a 2x2 fashion with the legend shown at the bottom. A 'faux' axis is created at the bottom to place the legend in a fixed spot. The 'faux' axis is then turned off so only the legend shows. Result:
import matplotlib.pyplot as plt
import matplotlib.gridspec as gridspec
# Gridspec demo
fig = plt.figure()
fig.set_size_inches(8, 9)
fig.set_dpi(100)
rows = 17 # The larger the number here, the smaller the spacing around the legend
start1 = 0
end1 = int((rows-1)/2)
start2 = end1
end2 = int(rows-1)
gspec = gridspec.GridSpec(ncols=4, nrows=rows)
axes = []
axes.append(fig.add_subplot(gspec[start1:end1, 0:2]))
axes.append(fig.add_subplot(gspec[start2:end2, 0:2]))
axes.append(fig.add_subplot(gspec[start1:end1, 2:4]))
axes.append(fig.add_subplot(gspec[start2:end2, 2:4]))
axes.append(fig.add_subplot(gspec[end2, 0:4]))
line, = axes[0].plot([0, 1], [0, 1], 'b') # Add some data
axes[-1].legend((line,), ('Test',), loc='center') # Create legend on bottommost axis
axes[-1].set_axis_off() # Don't show the bottom-most axis
fig.tight_layout()
plt.show()
This answer is a complement to user707650's answer on the legend position.
My first try on user707650's solution failed due to overlaps of the legend and the subplot's title.
In fact, the overlaps are caused by fig.tight_layout(), which changes the subplots' layout without considering the figure legend. However, fig.tight_layout() is necessary.
In order to avoid the overlaps, we can tell fig.tight_layout() to leave spaces for the figure's legend by fig.tight_layout(rect=(0,0,1,0.9)).
Description of tight_layout() parameters.
All of the previous answers are way over my head, at this state of my coding journey, so I just added another Matplotlib aspect called patches:
import matplotlib.patches as mpatches
first_leg = mpatches.Patch(color='red', label='1st plot')
second_leg = mpatches.Patch(color='blue', label='2nd plot')
thrid_leg = mpatches.Patch(color='green', label='3rd plot')
plt.legend(handles=[first_leg ,second_leg ,thrid_leg ])
The patches aspect put all the data i needed on my final plot (it was a line plot that combined three different line plots all in the same cell in Jupyter Notebook).
Result
(I changed the names form what I named my own legend.)
I would like to plot two dfs with two different colors. For each df, I would need to add two markers. Here is what I have tried:
for stats_file in stats_files:
data = Graph(stats_file)
Graph.compute(data)
data.servers_df.plot(x="time", y="percentage", linewidth=1, kind='line')
plt.plot(data.first_measurement['time'], data.first_measurement['percentage'], 'o-', color='orange')
plt.plot(data.second_measurement['time'], data.second_measurement['percentage'], 'o-', color='green')
plt.show()
Using this piece of code, I get the servers_df plotted with markers, but on separate graphs.
How I can have both graphs in a single one to compare them better?
Thanks.
TL;DR
Your call to data.servers_df.plot() always creates a new plot, and plt.plot() plots on the latest plot that was created. The solution is to create dedicated axis for everything to plot onto.
Preface
I assumed your variables are the following
data.servers_df: Dataframe with two float columns "time" and "percentage"
data.first_measurements: A dictionary with keys "time" and `"percentage", which each are a list of floats
data.second_measurements: A dictionary with keys "time" and "percentage", which each are a list of floats
I skipped generating stat_files as you did not show what Graph() does, but just created a list of dummy data.
If data.first_measurements and data.second_measurements are also dataframes, let me know and there is an even nicer solution.
Theory - Behind the curtains
Each matplotlib plot (line, bar, etc.) lives on a matplotlib.axes.Axes element. These are like regular axes of a coordinate system. Now two things happen here:
When you use plt.plot(), there are no axes specified and thus, matplotlib looks up the current axes element (in the background), and if there is none, it will create an empty one and use it, and set is as default. The second call to plt.plot() then finds these axes and uses them.
DataFrame.plot() on the other hand, always creates a new axes element if none is given to it (possible through the ax argument)
So in your code, data.servers_df.plot() first creates an axes element behind the curtains (which is then the default), and the two following plt.plot() calls get the default axes and plot onto it - which is why you get two plots instead of one.
Solution
The following solution first creates a dedicated matplotlib.axes.Axes using plt.subplots(). This axis element is then used to draw all lines onto. Note especially the ax=ax in data.server_df.plot(). Note that I changed the display of your markers from o- to o (as we don't want to display a line (-) but only markers (o)).
Mock data can be found below
fig, ax = plt.subplots() # Here we create the axes that all data will plot onto
for i, data in enumerate(stat_files):
y_column = f'percentage_{i}' # Make the columns identifiable
data.servers_df \
.rename(columns={'percentage': y_column}) \
.plot(x='time', y=y_column, linewidth=1, kind='line', ax=ax)
ax.plot(data.first_measurement['time'], data.first_measurement['percentage'], 'o', color='orange')
ax.plot(data.second_measurement['time'], data.second_measurement['percentage'], 'o', color='green')
plt.show()
Mock data
import random
import pandas as pd
import matplotlib.pyplot as plt
# Generation of dummy data
random.seed(1)
NUMBER_OF_DATA_FILES = 2
X_LENGTH = 10
class Data:
def __init__(self):
self.servers_df = pd.DataFrame(
{
'time': range(X_LENGTH),
'percentage': [random.randint(0, 10) for _ in range(X_LENGTH)]
}
)
self.first_measurement = {
'time': self.servers_df['time'].values[:X_LENGTH // 2],
'percentage': self.servers_df['percentage'].values[:X_LENGTH // 2]
}
self.second_measurement = {
'time': self.servers_df['time'].values[X_LENGTH // 2:],
'percentage': self.servers_df['percentage'].values[X_LENGTH // 2:]
}
stat_files = [Data() for _ in range(NUMBER_OF_DATA_FILES)]
DataFrame.plot() by default returns a matplotlib.axes.Axes object. You should then plot the other two plots on this object:
for stats_file in stats_files:
data = Graph(stats_file)
Graph.compute(data)
ax = data.servers_df.plot(x="time", y="percentage", linewidth=1, kind='line')
ax.plot(data.first_measurement['time'], data.first_measurement['percentage'], 'o-', color='orange')
ax.plot(data.second_measurement['time'], data.second_measurement['percentage'], 'o-', color='green')
plt.show()
If you want to plot them one on top of the others with different colors you can do something like this:
colors = ['C0', 'C1', 'C2'] # matplotlib default color palette
# assuming that len(stats_files) = 3
# if not you need to specify as many colors as necessary
ax = plt.subplot(111)
for stats_file, c in zip(stats_files, colors):
data = Graph(stats_file)
Graph.compute(data)
data.servers_df.plot(x="time", y="percentage", linewidth=1, kind='line', ax=ax)
ax.plot(data.first_measurement['time'], data.first_measurement['percentage'], 'o-', color=c)
ax.plot(data.second_measurement['time'], data.second_measurement['percentage'], 'o-', color='green')
plt.show()
This just changes the color of the servers_df.plot. If you want to change the color of the other two you can just to the same logic: create a list of colors that you want them to take at each iteration, iterate over that list and pass the color value to the color param at each iteration.
You can create an Axes object for plotting in the first place, for example
import pandas as pd
import numpy as np
from matplotlib import pyplot as plt
df_one = pd.DataFrame({'a':np.linspace(1,10,10),'b':np.linspace(1,10,10)})
df_two = pd.DataFrame({'a':np.random.randint(0,20,10),'b':np.random.randint(0,5,10)})
dfs = [df_one,df_two]
fig,ax = plt.subplots(figsize=(8,6))
colors = ['navy','darkviolet']
markers = ['x','o']
for ind,item in enumerate(dfs):
ax.plot(item['a'],item['b'],c=colors[ind],marker=markers[ind])
as you can see, in the same ax, the two dataframes are plotted with different colors and markers.
You need to create the plot before.
Afterwards, you can explicitly refer to this plot while plotting the graphs.
df.plot(..., ax=ax) or ax.plot(x, y)
import matplotlib.pyplot as plt
(fig, ax) = plt.subplots(figsize=(20,5))
for stats_file in stats_files:
data = Graph(stats_file)
Graph.compute(data)
data.servers_df.plot(x="time", y="percentage", linewidth=1, kind='line', ax=ax)
ax.plot(data.first_measurement['time'], data.first_measurement['percentage'], 'o-', color='orange')
ax.plot(data.second_measurement['time'], data.second_measurement['percentage'], 'o-', color='green')
plt.show()
I'm working on a plot with translucent 'x' markers (20% alpha). How do I make the marker appear at 100% opacity in the legend?
import matplotlib.pyplot as plt
plt.plot_date( x = xaxis, y = yaxis, marker = 'x', color=[1, 0, 0, .2], label='Data Series' )
plt.legend(loc=3, mode="expand", numpoints=1, scatterpoints=1 )
UPDATED:
There is an easier way! First, assign your legend to a variable when you create it:
leg = plt.legend()
Then:
for lh in leg.legendHandles:
lh.set_alpha(1)
OR if the above doesn't work (you may be using an older version of matplotlib):
for lh in leg.legendHandles:
lh._legmarker.set_alpha(1)
to make your markers opaque for a plt.plot or a plt.scatter, respectively.
Note that using simply lh.set_alpha(1) on a plt.plot will make the lines in your legend opaque rather than the markers. You should be able to adapt these two possibilities for the other plot types.
Sources:
Synthesized from some good advice by DrV about marker sizes. Update was inspired by useful comment from Owen.
Following up on cosmosis's answer, to make the "fake" lines for the legend invisible on the plot, you can use NaNs, and they will still work for generating legend entries:
import numpy as np
import matplotlib.pyplot as plt
# Plot data with alpha=0.2
plt.plot((0,1), (0,1), marker = 'x', color=[1, 0, 0, .2])
# Plot non-displayed NaN line for legend, leave alpha at default of 1.0
legend_line_1 = plt.plot( np.NaN, np.NaN, marker = 'x', color=[1, 0, 0], label='Data Series' )
plt.legend()
Other answers here give good practical solutions by either changing the alpha value in the legend after creation, or changing the alpha of the line after legend creation.
A solution to achieve a different opacity in the legend without manipulating anything afterwards would be the following. It uses a handler_map and an updating function.
import matplotlib.pyplot as plt
import numpy as np; np.random.seed(43)
from matplotlib.collections import PathCollection
from matplotlib.legend_handler import HandlerPathCollection, HandlerLine2D
plt.plot(np.linspace(0,1,8), np.random.rand(8), marker="o", markersize=12, label="A line", alpha=0.2)
plt.scatter(np.random.rand(8),np.random.rand(8), s=144,
c="red", marker=r"$\clubsuit$", label="A scatter", alpha=0.2)
def update(handle, orig):
handle.update_from(orig)
handle.set_alpha(1)
plt.legend(handler_map={PathCollection : HandlerPathCollection(update_func= update),
plt.Line2D : HandlerLine2D(update_func = update)})
plt.show()
If you want to have something specific in your legend, it's easier to define objects that you place in the legend with appropriate text. For example:
import matplotlib.pyplot as plt
import pylab
plt.plot_date( x = xaxis, y = yaxis, marker = 'x', color=[1, 0, 0, .2], label='Data Series' )
line1 = pylab.Line2D(range(1),range(1),color='white',marker='x',markersize=10, markerfacecolor="red",alpha=1.0)
line2 = pylab.Line2D(range(10),range(10),marker="_",linewidth=3.0,color="dodgerblue",alpha=1.0)
plt.legend((line1,line2),('Text','Other Text'),numpoints=1,loc=1)
Here, line1 defines a short, white line (so essentially invisible) with the marker 'x' in red and full opacity. As an example, line2 gives you a longer blue line with no markers visible. By creating this "lines," you are able to more easily control their properties within the legend.
It looks like matplotlib draws the plot lines after it copies the alpha level to the legend. That means that you can create the plot lines with the alpha level that you want in the legend, create the legend to copy that alpha level, then change the alpha level on the plot lines.
Here's a complete example:
import matplotlib.pyplot as plt
x = (0, 1, 2)
y = (0, 2, 1)
line, = plt.plot(x, y, 'ro', label='label') # Default alpha is 1.0.
plt.legend() # Copy alpha to legend.
line.set_alpha(0.2) # Change alpha for data points.
plt.show()
That plot looks like this when I run it with matplotlib 2.2.3 on Python 2.7.15:
I've found that the .set_alpha() function works on many legend objects, but unfortunately, many legend objects have several pieces (such as the output of errorbar()) and the .set_alpha() call will only affect one of them.
One can use .get_legend_handles_labels() and then loop through parts of the handles and .set_alpha(), but unfortunately, copy.deepcopy() does not seem to work on the list of handles, so the plot itself will be affected. The best workaround I could find was to save the original alphas, .set_alpha() to what I wanted, create the legend, then reset the plot alphas back to their original values. It would be much cleaner if I could deepcopy handles (I wouldn't have to save alpha values or reset them), but I could not do this in python2.7 (maybe this depends on what objects are in the legend).
f,ax=plt.subplots(1)
ax.plot( ... )
def legend_alpha(ax,newalpha=1.0):
#sets alpha of legends to some value
#this would be easier if deepcopy worked on handles, but it doesn't
handles,labels=ax.get_legend_handles_labels()
alphass=[None]*len(handles) #make a list to hold lists of saved alpha values
for k,handle in enumerate(handles): #loop through the legend entries
alphas=[None]*len(handle) #make a list to hold the alphas of the pieces of this legend entry
for i,h in enumerate(handle): #loop through the pieces of this legend entry (there could be a line and a marker, for example)
try: #if handle was a simple list of parts, then this will work
alphas[i]=h.get_alpha()
h.set_alpha(newalpha)
except: #if handle was a list of parts which themselves were made up of smaller subcomponents, then we must go one level deeper still.
#this was needed for the output of errorbar() and may not be needed for simpler plot objects
alph=[None]*len(h)
for j,hh in enumerate(h):
alph[j]=hh.get_alpha() #read the alpha values of the sub-components of the piece of this legend entry
hh.set_alpha(newalpha)
alphas[i]=alph #save the list of alpha values for the subcomponents of this piece of this legend entry
alphass[k]=alphas #save the list of alpha values for the pieces of this legend entry
leg=ax.legend(handles,labels) #create the legend while handles has updated alpha values
for k,handle in enumerate(handles): #loop through legend items to restore origina alphas on the plot
for i,h in enumerate(handle): #loop through pieces of this legend item to restore alpha values on the plot
try:
h.set_alpha(alphass[k][i])
except:
for j,hh in enumerate(h): #loop through sub-components of this piece of this legend item to restore alpha values
hh.set_alpha(alphass[k][i][j])
return leg
leg=legend_alpha(ax)
leg.draggable()
In my case, set_alpha(1) also modified the edgecolors, which I didn't want: I had "invisible" edges, and setting alpha to opaque made them visible in the legend. The following snippet (OOP) changes the opacity of the face without changing the border color:
leg = ax.legend()
for lh in leg.legendHandles:
fc_arr = lh.get_fc().copy()
fc_arr[:, -1] = 1 # set opacity here
lh.set_fc(fc_arr)
Note the call to .copy(), if we don't do this it will modify the opacity for the whole plot. Calling copy means we are only modifying the facecolor inside the legend box.
Alternatively, you can add this function to your library:
def opaque_legend(ax):
"""
Calls legend, and sets all the legend colors opacity to 100%.
Returns the legend handle.
"""
leg = ax.legend()
for lh in leg.legendHandles:
fc_arr = lh.get_fc().copy()
fc_arr[:, -1] = 1
lh.set_fc(fc_arr)
return leg
And then simply replace leg = ax.legend() with leg = opaque_legend(ax). Hope this helps!
Andres
Instead of messing up with the opacity of the legend, I found another way. Firstly, I create a plot line with the style I want the legend to be. Then I change the plot line style, and, miraculously, the legend style remains intact. MWE:
plt.plot(x, y, 'ro', label='label')
for lh in plt.gca().get_legend_handles_labels():
lh[0].set_alpha(new_alpha)
I'd like to explain, why it works, but I can't. Neither I'm sure that it works for all backends.
And yes, I know that the question is old. As it still appears in Google, I'll find it later and help my future self.
I would like to create a custom legend for multiple plots in matplotlib (python) in a pyqt GUI. (pyqt advises against using pyplot so the object-oriented method has to be used).
Multiple plots will be appear in a grid but the user can define how many plots to appear. I would like the legend to appear on the right hand side of all the plots therefore I cannot simply create the legend for last axes plotted. I would like the legend to be created for the entire figure not just the last axis (similarly to plt.figlegend in pyplot).
In examples I have seen elsewhere, this requires referencing the lines plotted. Again, I can't do this because the user has the possibility of choosing which lines to appear on the graphs, and I would rather the legend alway show all the possible lines whether they are currently displayed or not.
(Note the example code below uses pyplot but my final version cannot)
import matplotlib.pyplot as plt
import matplotlib.lines as mlines
import numpy as np
fig = plt.figure()
# Create plots in 2x2 grid
for plot in range(4):
# Create plots
x = np.arange(0, 10, 0.1)
y = np.random.randn(len(x))
y2 = np.random.randn(len(x))
ax = fig.add_subplot(2,2,plot+1)
plt.plot(x, y, label="y")
plt.plot(x, y2, label="y2")
# Create custom legend
blue_line = mlines.Line2D([], [], color='blue',markersize=15, label='Blue line')
green_line = mlines.Line2D([], [], color='green', markersize=15, label='Green line')
ax.legend(handles=[blue_line,green_line],bbox_to_anchor=(1.05, 0), loc='lower left', borderaxespad=0.)
If I change ax.legend to:
fig.legend(handles=[blue_line,green_line])
then python produces the error:
TypeError: legend() takes at least 3 arguments (2 given)
(I guess because the line points aren't referenced)
Thanks for any help offered - I've been looking at this for a week now!
the error you are getting is because Figure.legend requires you to pass it both the handles and the labels.
From the docs:
legend(handles, labels, *args, **kwargs)
Place a legend in the figure. labels are a sequence of strings, handles is a sequence of Line2D or Patch instances.
The following works:
# Create custom legend
blue_line = mlines.Line2D([], [], color='blue',markersize=15, label='Blue line')
green_line = mlines.Line2D([], [], color='green', markersize=15, label='Green line')
handles = [blue_line,green_line]
labels = [h.get_label() for h in handles]
fig.legend(handles=handles, labels=labels)
I'm working on a plot with translucent 'x' markers (20% alpha). How do I make the marker appear at 100% opacity in the legend?
import matplotlib.pyplot as plt
plt.plot_date( x = xaxis, y = yaxis, marker = 'x', color=[1, 0, 0, .2], label='Data Series' )
plt.legend(loc=3, mode="expand", numpoints=1, scatterpoints=1 )
UPDATED:
There is an easier way! First, assign your legend to a variable when you create it:
leg = plt.legend()
Then:
for lh in leg.legendHandles:
lh.set_alpha(1)
OR if the above doesn't work (you may be using an older version of matplotlib):
for lh in leg.legendHandles:
lh._legmarker.set_alpha(1)
to make your markers opaque for a plt.plot or a plt.scatter, respectively.
Note that using simply lh.set_alpha(1) on a plt.plot will make the lines in your legend opaque rather than the markers. You should be able to adapt these two possibilities for the other plot types.
Sources:
Synthesized from some good advice by DrV about marker sizes. Update was inspired by useful comment from Owen.
Following up on cosmosis's answer, to make the "fake" lines for the legend invisible on the plot, you can use NaNs, and they will still work for generating legend entries:
import numpy as np
import matplotlib.pyplot as plt
# Plot data with alpha=0.2
plt.plot((0,1), (0,1), marker = 'x', color=[1, 0, 0, .2])
# Plot non-displayed NaN line for legend, leave alpha at default of 1.0
legend_line_1 = plt.plot( np.NaN, np.NaN, marker = 'x', color=[1, 0, 0], label='Data Series' )
plt.legend()
Other answers here give good practical solutions by either changing the alpha value in the legend after creation, or changing the alpha of the line after legend creation.
A solution to achieve a different opacity in the legend without manipulating anything afterwards would be the following. It uses a handler_map and an updating function.
import matplotlib.pyplot as plt
import numpy as np; np.random.seed(43)
from matplotlib.collections import PathCollection
from matplotlib.legend_handler import HandlerPathCollection, HandlerLine2D
plt.plot(np.linspace(0,1,8), np.random.rand(8), marker="o", markersize=12, label="A line", alpha=0.2)
plt.scatter(np.random.rand(8),np.random.rand(8), s=144,
c="red", marker=r"$\clubsuit$", label="A scatter", alpha=0.2)
def update(handle, orig):
handle.update_from(orig)
handle.set_alpha(1)
plt.legend(handler_map={PathCollection : HandlerPathCollection(update_func= update),
plt.Line2D : HandlerLine2D(update_func = update)})
plt.show()
If you want to have something specific in your legend, it's easier to define objects that you place in the legend with appropriate text. For example:
import matplotlib.pyplot as plt
import pylab
plt.plot_date( x = xaxis, y = yaxis, marker = 'x', color=[1, 0, 0, .2], label='Data Series' )
line1 = pylab.Line2D(range(1),range(1),color='white',marker='x',markersize=10, markerfacecolor="red",alpha=1.0)
line2 = pylab.Line2D(range(10),range(10),marker="_",linewidth=3.0,color="dodgerblue",alpha=1.0)
plt.legend((line1,line2),('Text','Other Text'),numpoints=1,loc=1)
Here, line1 defines a short, white line (so essentially invisible) with the marker 'x' in red and full opacity. As an example, line2 gives you a longer blue line with no markers visible. By creating this "lines," you are able to more easily control their properties within the legend.
It looks like matplotlib draws the plot lines after it copies the alpha level to the legend. That means that you can create the plot lines with the alpha level that you want in the legend, create the legend to copy that alpha level, then change the alpha level on the plot lines.
Here's a complete example:
import matplotlib.pyplot as plt
x = (0, 1, 2)
y = (0, 2, 1)
line, = plt.plot(x, y, 'ro', label='label') # Default alpha is 1.0.
plt.legend() # Copy alpha to legend.
line.set_alpha(0.2) # Change alpha for data points.
plt.show()
That plot looks like this when I run it with matplotlib 2.2.3 on Python 2.7.15:
I've found that the .set_alpha() function works on many legend objects, but unfortunately, many legend objects have several pieces (such as the output of errorbar()) and the .set_alpha() call will only affect one of them.
One can use .get_legend_handles_labels() and then loop through parts of the handles and .set_alpha(), but unfortunately, copy.deepcopy() does not seem to work on the list of handles, so the plot itself will be affected. The best workaround I could find was to save the original alphas, .set_alpha() to what I wanted, create the legend, then reset the plot alphas back to their original values. It would be much cleaner if I could deepcopy handles (I wouldn't have to save alpha values or reset them), but I could not do this in python2.7 (maybe this depends on what objects are in the legend).
f,ax=plt.subplots(1)
ax.plot( ... )
def legend_alpha(ax,newalpha=1.0):
#sets alpha of legends to some value
#this would be easier if deepcopy worked on handles, but it doesn't
handles,labels=ax.get_legend_handles_labels()
alphass=[None]*len(handles) #make a list to hold lists of saved alpha values
for k,handle in enumerate(handles): #loop through the legend entries
alphas=[None]*len(handle) #make a list to hold the alphas of the pieces of this legend entry
for i,h in enumerate(handle): #loop through the pieces of this legend entry (there could be a line and a marker, for example)
try: #if handle was a simple list of parts, then this will work
alphas[i]=h.get_alpha()
h.set_alpha(newalpha)
except: #if handle was a list of parts which themselves were made up of smaller subcomponents, then we must go one level deeper still.
#this was needed for the output of errorbar() and may not be needed for simpler plot objects
alph=[None]*len(h)
for j,hh in enumerate(h):
alph[j]=hh.get_alpha() #read the alpha values of the sub-components of the piece of this legend entry
hh.set_alpha(newalpha)
alphas[i]=alph #save the list of alpha values for the subcomponents of this piece of this legend entry
alphass[k]=alphas #save the list of alpha values for the pieces of this legend entry
leg=ax.legend(handles,labels) #create the legend while handles has updated alpha values
for k,handle in enumerate(handles): #loop through legend items to restore origina alphas on the plot
for i,h in enumerate(handle): #loop through pieces of this legend item to restore alpha values on the plot
try:
h.set_alpha(alphass[k][i])
except:
for j,hh in enumerate(h): #loop through sub-components of this piece of this legend item to restore alpha values
hh.set_alpha(alphass[k][i][j])
return leg
leg=legend_alpha(ax)
leg.draggable()
In my case, set_alpha(1) also modified the edgecolors, which I didn't want: I had "invisible" edges, and setting alpha to opaque made them visible in the legend. The following snippet (OOP) changes the opacity of the face without changing the border color:
leg = ax.legend()
for lh in leg.legendHandles:
fc_arr = lh.get_fc().copy()
fc_arr[:, -1] = 1 # set opacity here
lh.set_fc(fc_arr)
Note the call to .copy(), if we don't do this it will modify the opacity for the whole plot. Calling copy means we are only modifying the facecolor inside the legend box.
Alternatively, you can add this function to your library:
def opaque_legend(ax):
"""
Calls legend, and sets all the legend colors opacity to 100%.
Returns the legend handle.
"""
leg = ax.legend()
for lh in leg.legendHandles:
fc_arr = lh.get_fc().copy()
fc_arr[:, -1] = 1
lh.set_fc(fc_arr)
return leg
And then simply replace leg = ax.legend() with leg = opaque_legend(ax). Hope this helps!
Andres
Instead of messing up with the opacity of the legend, I found another way. Firstly, I create a plot line with the style I want the legend to be. Then I change the plot line style, and, miraculously, the legend style remains intact. MWE:
plt.plot(x, y, 'ro', label='label')
for lh in plt.gca().get_legend_handles_labels():
lh[0].set_alpha(new_alpha)
I'd like to explain, why it works, but I can't. Neither I'm sure that it works for all backends.
And yes, I know that the question is old. As it still appears in Google, I'll find it later and help my future self.