I'm trying to have one large plot, and then use arrows to connect outliers to smaller plots. I used gridspec to create my plots, and on their own they show up perfectly.
However, when I try to use ConnectionPatch to add an artist for the arrows, it changes the size of all the subplots.
Here is my code setting up the figure and subplots using gridspec:
fig2 = plt.figure(constrained_layout=True)
gs = fig2.add_gridspec(ncols=3, nrows=9, figure=fig2)
f2_ax1 = fig2.add_subplot(gs[0:-4, :-1])
f2_ax2 = fig2.add_subplot(gs[2:-4, 2])
f2_ax3 = fig2.add_subplot(gs[5:-1, 0])
f2_ax4 = fig2.add_subplot(gs[5:-1, 1])
f2_ax5 = fig2.add_subplot(gs[5:-1,2])
This creates the plots I want (see picture):
Here is how I'm adding arrows to the plots, using a for loop to iterate through each subplot:
coordsA = "data" #The arrows are plotting using the same coordinates as the data points
coordsB = "data"
i = 0
for sub_plot in sub_plot_list: #Adding the arrows to the subplots
xy1 = (x_oulier_list[i], y_outlier_list[i]) #Fetches coordinates for beginning of arrow
xy2 = arrow_end_list[i] #Fetches coordinates for the pointy end of arrow
con = ConnectionPatch(xyA=(x_outlier_list[i],y_outlier_list[i]), xyB=arrow_end_list[i],
coordsA=coordsA, coordsB=coordsB,
axesA=f2_ax1, axesB=sub_plot,
arrowstyle="->", shrinkB=5)
f2_ax1.add_artist(con) #Adds arrow to plot
i += 1
Both versions use fig2.tight_layout().
Here is the plot after adding arrows:
Any idea how I can fix it so that the dimensions of the plot don't change when I add arrows?
Looks like the problem is with constrained_layout, which does not seem to handle ConnectionPatch very well (constrained_layout is still experimental, maybe it is worth raising an issue on matplotlib's github in case someone can figure the problem out).
An easy fix is to request that the ConnectionPatch objects are not taken into account when calculating the layout using:
(...)
con = ConnectionPatch(xyA=(x_outlier_list[i],y_outlier_list[i]), xyB=arrow_end_list[i],
coordsA=coordsA, coordsB=coordsB,
axesA=f2_ax1, axesB=sub_plot,
arrowstyle="->", shrinkB=5)
con.set_in_layout(False)
(...)
Related
I have two imshow() problems that I suspect are closely related.
First, I can't figure out how to use set_data() to update an image I've created with imshow().
Second, I can't figure out why the colorbar I add to the imshow() plot after I'm done updating the plot doesn't match the colorbar I add to an imshow() plot of the same data that I create from scratch after I'm done taking data. The colorbar of the second plot appears to be correct.
Background.
I'm collecting measurement data in two nested loops, with each loop controlling one of the measurement conditions. I'm using pyplot.imshow() to plot my results, and I'm updating the imshow() plot every time I take data in the inner loop.
What I have works in terms of updating the imshow() plot but it seems to be getting increasingly slower as I add more loop iterations, so it's not scaling well. (The program I've included in with this post creates a plot that is eight rows high and six columns wide. A "real" plot might be 10x or 20x this size, in both dimensions.)
I think what I want to do is use the image's set_data() method but I can't figure out how. What I've tried either throws an error or doesn't appear to have any effect on the imshow() plot.
Once I'm done with the "take the data" loops, I add a colorbar to the imshow() plot I've been updating. However, the colorbar scale is obviously bogus.
In contrast, if I take create an entirely new imshow() plot, using the data I took in the loops, and then add a colorbar, the colorbar appears to be correct.
My hunch is the problem is with the vmin and vmax values associated with the imshow() plot I'm updating in the loops but I can't figure out how to fix it.
I've already looked at several related StackOverflow posts. For example:
update a figure made with imshow(), contour() and quiver()
Update matplotlib image in a function
How to update matplotlib's imshow() window interactively?
These have helped, in that they've pointed me to set_data() and given me solutions to some other
problems I had, but I still have the two problems I mentioned at the start.
Here's a simplified version of my code. Note that there are repeated zero values on the X and Y axes. This is on purpose.
I'm running Python 3.5.1, matplotlib 1.5.1, and numpy 1.10.4. (Yes, some of these are quite old. Corporate IT reasons.)
import numpy as np
import matplotlib.pyplot as plt
import random
import time
import warnings
warnings.filterwarnings("ignore", ".*GUI is implemented.*") # Filter out bogus matplotlib warning.
# Create the simulated data for plotting
v_max = 120
v_step_size = 40
h_max = 50
h_step_size = 25
scale = 8
v_points = np.arange(-1*abs(v_max), 0, abs(v_step_size))
v_points = np.append(v_points, [-0.0])
reversed_v_points = -1 * v_points[::-1] # Not just reverse order, but reversed sign
v_points = np.append(v_points, reversed_v_points)
h_points = np.arange(-1*abs(h_max), 0, abs(h_step_size))
h_points = np.append(h_points, [-0.0])
reversed_h_points = -1 * h_points[::-1] # Not just reverse order, but reversed sign
h_points = np.append(h_points, reversed_h_points)
h = 0 # Initialize
v = 0 # Initialize
plt.ion() # Turn on interactive mode.
fig, ax = plt.subplots() # So I have access to the figure and the axes of the plot.
# Initialize the data_points
data_points = np.zeros((v_points.size, h_points.size))
im = ax.imshow(data_points, cmap='hot', interpolation='nearest') # Specify the color map and interpolation
ax.set_title('Dummy title for initial plot')
# Set up the X-axis ticks and label them
ax.set_xticks(np.arange(len(h_points)))
ax.set_xticklabels(h_points)
ax.set_xlabel('Horizontal axis measurement values')
# Set up the Y-axis ticks and label them
ax.set_yticks(np.arange(len(v_points)))
ax.set_yticklabels(v_points)
ax.set_ylabel('Vertical axis measurement values')
plt.pause(0.0001) # In interactive mode, need a small delay to get the plot to appear
plt.show()
for v, v_value in enumerate(v_points):
for h, h_value in enumerate(h_points):
# Measurement goes here.
time.sleep(0.1) # Simulate the measurement delay.
measured_value = scale * random.uniform(0.0, 1.0) # Create simulated data
data_points[v][h] = measured_value # Update data_points with the simulated data
# Update the heat map with the latest point.
# - I *think* I want to use im.set_data() here, not ax.imshow(), but how?
ax.imshow(data_points, cmap='hot', interpolation='nearest') # Specify the color map and interpolation
plt.pause(0.0001) # In interactive mode, need a small delay to get the plot to appear
plt.draw()
# Create a colorbar
# - Except the colorbar here is wrong. It goes from -0.10 to +0.10 instead
# of matching the colorbar in the second imshow() plot, which goes from
# 0.0 to "scale". Why?
cbar = ax.figure.colorbar(im, ax=ax)
cbar.ax.set_ylabel('Default heatmap colorbar label')
plt.pause(0.0001) # In interactive mode, need a small delay to get the colorbar to appear
plt.show()
fig2, ax2 = plt.subplots() # So I have access to the figure and the axes of the plot.
im = ax2.imshow(data_points, cmap='hot', interpolation='nearest') # Specify the color map and interpolation
ax2.set_title('Dummy title for plot with pseudo-data')
# Set up the X-axis ticks and label them
ax2.set_xticks(np.arange(len(h_points)))
ax2.set_xticklabels(h_points)
ax2.set_xlabel('Horizontal axis measurement values')
# Set up the Y-axis ticks and label them
ax2.set_yticks(np.arange(len(v_points)))
ax2.set_yticklabels(v_points)
ax2.set_ylabel('Vertical axis measurement values')
# Create a colorbar
cbar = ax2.figure.colorbar(im, ax=ax2)
cbar.ax.set_ylabel('Default heatmap colorbar label')
plt.pause(0.0001) # In interactive mode, need a small delay to get the plot to appear
plt.show()
dummy = input("In interactive mode, press the Enter key when you're done with the plots.")
OK, I Googled some more. More importantly, I Googled smarter and figured out my own answer.
To update my plot inside my nested loops, I was using the command:
ax.imshow(data_points, cmap='hot', interpolation='nearest') # Specify the color map and interpolation
What I tried using to update my plot more efficiently was:
im.set_data(data_points)
What I should have used was:
im.set_data(data_points)
im.autoscale()
This updates the pixel scaling, which fixed both my "plot doesn't update" problem and my "colorbar has the wrong scale" problem.
I've only been using Python for about a month now, so I'm sorry if there's some simple solution to this that I overlooked.
Basically I have a figure with 4 subplots, the 2 on the left show longitudinal plots and the ones on the right show scatter plots at certain points of the longitudinal plots. You can click through the scatter plots at different points of the longitudinal plot with buttons, and the tick label of the longitudinal plot you're currently at will be highlighted in blue.
Coloring a certain tick label already works with this:
xlabels = []
labelcolors = []
for i, item in enumerate(mr.segmentlst):
if re.search('SFX|MC|MQ|MS|MKC', item):
xlabels.append(mr.segmentlst[i])
else:
xlabels.append('')
for i, item in enumerate(mr.segmentlst):
if re.search('SFX', item):
labelcolors.append('black')
else:
labelcolors.append('gray')
labelcolors[self.ind]='blue'
[t.set_color(i) for (i,t) in zip(labelcolors, ax1.xaxis.get_ticklabels())]
[t.set_color(i) for (i,t) in zip(labelcolors, ax2.xaxis.get_ticklabels())]
It only shows certain tick labels and changes their colors accordingly (I don't know if there is another solution for this, it's the only one I could find). Don't mind the mr.segmentlist, I've currently hardcoded the plot to use an attribute from another method so I can easily keep testing it in Spyder.
I'd like to also change the grid line color of the currently highlighted tick label (only xgridlines are visible) in the longitudinal plots, is there some kind of similar way of doing this? I've searched the internet for a solution for about 2 hours now and didn't really find anything helpful.
I thought something like ax1.get_xgridlines() might be used, but I have no idea how I could transform it into a useful list.
Thanks,
Tamara
get_xgridlines() returns a list of Line2D objects, so if you can locate which line you want to modify, you can modify any of their properties
x = np.random.random_sample((10,))
fig = plt.figure()
ax = fig.add_subplot(111)
ax.scatter(x,x)
ax.grid()
a = ax.get_xgridlines()
b = a[2]
b.set_color('red')
b.set_linewidth(3)
since the above solution only works with major gridlines
(since get_gridlines() is currently hardcoded to use only the major ones),
here's how you can also access the minor gridlines by adapting
the get_gridlines() function (from here):
from matplotlib import cbook
def get_gridlines(ax, which):
'''
Parameters:
ax : ax.xaxis or ax.yaxis instance
which : 'major' or 'minor'
Returns:
The grid lines as a list of Line2D instance
'''
if which == 'major':
ticks = ax.get_major_ticks()
if which == 'minor':
ticks = ax.get_minor_ticks()
return cbook.silent_list('Line2D gridline',
[tick.gridline for tick in ticks])
Does someone perhaps know if it is possible to rotate a legend on a plot in matplotlib? I made a simple plot with the below code, and edited the graph in paint to show what I want.
plt.plot([4,5,6], label = 'test')
ax = plt.gca()
ax.legend()
plt.show()
I went to a similar problem and solved it by writing the function legendAsLatex that generates a latex code to be used as the label of the y-axis. The function gathers the color, the marker, the line style, and the label provided to the plot function. It requires enabling the latex and loading the required packages. Here is the code to generate your plot with extra curves that use both vertical axis.
from matplotlib import pyplot as plt
import matplotlib.colors as cor
plt.rc('text', usetex=True)
plt.rc('text.latex', preamble=r'\usepackage{amsmath} \usepackage{wasysym}'+
r'\usepackage[dvipsnames]{xcolor} \usepackage{MnSymbol} \usepackage{txfonts}')
def legendAsLatex(axes, rotation=90) :
'''Generate a latex code to be used instead of the legend.
Uses the label, color, marker and linestyle provided to the pyplot.plot.
The marker and the linestyle must be defined using the one or two character
abreviations shown in the help of pyplot.plot.
Rotation of the markers must be multiple of 90.
'''
latexLine = {'-':'\\textbf{\Large ---}',
'-.':'\\textbf{\Large --\:\!$\\boldsymbol{\cdot}$\:\!--}',
'--':'\\textbf{\Large --\,--}',':':'\\textbf{\Large -\:\!-}'}
latexSymbol = {'o':'medbullet', 'd':'diamond', 's':'filledmedsquare',
'D':'Diamondblack', '*':'bigstar', '+':'boldsymbol{\plus}',
'x':'boldsymbol{\\times}', 'p':'pentagon', 'h':'hexagon',
',':'boldsymbol{\cdot}', '_':'boldsymbol{\minus}','<':'LHD',
'>':'RHD','v':'blacktriangledown', '^':'blacktriangle'}
rot90=['^','<','v','>']
di = [0,-1,2,1][rotation%360//90]
latexSymbol.update({rot90[i]:latexSymbol[rot90[(i+di)%4]] for i in range(4)})
return ', '.join(['\\textcolor[rgb]{'\
+ ','.join([str(x) for x in cor.to_rgb(handle.get_color())]) +'}{'
+ '$\\'+latexSymbol.get(handle.get_marker(),';')+'$'
+ latexLine.get(handle.get_linestyle(),'') + '} ' + label
for handle,label in zip(*axes.get_legend_handles_labels())])
ax = plt.axes()
ax.plot(range(0,10), 'b-', label = 'Blue line')
ax.plot(range(10,0,-1), 'sm', label = 'Magenta squares')
ax.set_ylabel(legendAsLatex(ax))
ax2 = plt.twinx()
ax2.plot([x**0.5 for x in range(0,10)], 'ro', label = 'Red circles')
ax2.plot([x**0.5 for x in range(10,0,-1)],'g--', label = 'Green dashed line')
ax2.set_ylabel(legendAsLatex(ax2))
plt.savefig('legend.eps')
plt.close()
Figure generated by the code:
I spent a few hours chipping away at this yesterday, and made a bit of progress so I'll share that below along with some suggestions moving forward.
First, it seems that we can certainly rotate and translate the bounding box (bbox) or frame around the legend. In the first example below you can see that a transform can be applied, albeit requiring some oddly large translation numbers after applying the 90 degree rotation. But, there are actually problems saving the translated legend frame to an image file so I had to take a screenshot from the IPython notebook. I've added some comments as well.
import matplotlib.pyplot as plt
%matplotlib inline
import numpy as np
import matplotlib.transforms
fig = plt.figure()
ax = fig.add_subplot('121') #make room for second subplot, where we are actually placing the legend
ax2 = fig.add_subplot('122') #blank subplot to make space for legend
ax2.axis('off')
ax.plot([4,5,6], label = 'test')
transform = matplotlib.transforms.Affine2D(matrix=np.eye(3)) #start with the identity transform, which does nothing
transform.rotate_deg(90) #add the desired 90 degree rotation
transform.translate(410,11) #for some reason we need to play with some pretty extreme translation values to position the rotated legend
legend = ax.legend(bbox_to_anchor=[1.5,1.0])
legend.set_title('test title')
legend.get_frame().set_transform(transform) #This actually works! But, only for the frame of the legend (see below)
frame = legend.get_frame()
fig.subplots_adjust(wspace = 0.4, right = 0.9)
fig.savefig('rotate_legend_1.png',bbox_extra_artists=(legend,frame),bbox_inches='tight', dpi = 300) #even with the extra bbox parameters the legend frame is still getting clipped
Next, I thought it would be smart to explore the get_methods() of other legend components. You can sort of dig through these things with dir(legend) and legend.__dict__ and so on. In particular, I noticed that you can do this: legend.get_title().set_transform(transform), which would seem to imply that we could translate the legend text (and not just the frame as above). Let's see what happens when I tried that:
fig2 = plt.figure()
ax = fig2.add_subplot('121')
ax2 = fig2.add_subplot('122')
ax2.axis('off')
ax.plot([4,5,6], label = 'test')
transform = matplotlib.transforms.Affine2D(matrix=np.eye(3))
transform.rotate_deg(90)
transform.translate(410,11)
legend = ax.legend(bbox_to_anchor=[1.5,1.0])
legend.set_title('test title')
legend.get_frame().set_transform(transform)
legend.get_title().set_transform(transform) #one would expect this to apply the same transformation to the title text in the legend, rotating it 90 degrees and translating it
frame = legend.get_frame()
fig2.subplots_adjust(wspace = 0.4, right = 0.9)
fig2.savefig('rotate_legend_1.png',bbox_extra_artists=(legend,frame),bbox_inches='tight', dpi = 300)
The legend title seems to have disappeared in the screenshot from the IPython notebook. But, if we look at the saved file the legend title is now in the bottom left corner and seems to have ignored the rotation component of the transformation (why?):
I had similar technical difficulties with this type of approach:
bbox = matplotlib.transforms.Bbox([[0.,1],[1,1]])
trans_bbox = matplotlib.transforms.TransformedBbox(bbox, transform)
legend.set_bbox_to_anchor(trans_bbox)
Other notes and suggestions:
It might be a sensible idea to dig into the differences in behaviour between the legend title and frame objects--why do they both accept transforms, but only the frame accepts a rotation? Perhaps it would be possible to subclass the legend object in the source code and make some adjustments.
We also need to find a solution for the rotated / translated legend frame not being saved to output, even after following various related suggestion on SO (i.e., Matplotlib savefig with a legend outside the plot).
So I am back again with another silly question.
Consider this piece of code
x = linspace(-10,10,100);
[X,Y]=meshgrid(x,x)
g = np.exp(-(square(X)+square(Y))/2)
plt.imshow(g)
scat = plt.scatter(50,50,c='r',marker='+')
Is there a way to clear only the scatter point on the graph without clearing all the image?
In fact, I am writing a code where the appearance of the scatter point is bound with a Tkinter Checkbutton and I want it to appear/disappear when I click/unclick the button.
Thanks for your help!
The return handle of plt.scatter has several methods, including remove(). So all you need to do is call that. With your example:
x = np.linspace(-10,10,100);
[X,Y] = np.meshgrid(x,x)
g = np.exp(-(np.square(X) + np.square(Y))/2)
im_handle = plt.imshow(g)
scat = plt.scatter(50,50,c='r', marker='+')
# image, with scatter point overlayed
scat.remove()
plt.draw()
# underlying image, no more scatter point(s) now shown
# For completeness, can also remove the other way around:
plt.clf()
im_handle = plt.imshow(g)
scat = plt.scatter(50,50,c='r', marker='+')
# image with both components
im_handle.remove()
plt.draw()
# now just the scatter points remain.
(almost?) all matplotlib rendering functions return a handle, which have some method to remove the rendered item.
Note that you need the call to redraw to see the effects of remove() -- from the remove help (my emphasis):
Remove the artist from the figure if possible. The effect will not be
visible until the figure is redrawn, e.g., with
:meth:matplotlib.axes.Axes.draw_idle.
I am plotting a 2D view of a spacecraft orbit using matplotlib. On this orbit, I identify and mark certain events, and then list these events and the corresponding dates in a legend. Before saving the figure to a file, I autozoom on my orbit plot, which causes the legend to be printed directly on top of my plot. What I would like to do is, after autoscaling, somehow find out the width of my legend, and then expand my xaxis to "make room" for the legend on the right side of the plot. Conceptually, something like this;
# ... code that generates my plot up here, then:
ax.autoscale_view()
leg = ax.get_legend()
leg_width = # Somehow get the width of legend in units that I can use to modify my axes
xlims = ax.get_xlim()
ax.set_xlim( [xlims[0], xlims[1] + leg_width] )
fig.savefig('myplot.ps',format='ps')
The main problem I'm having is that ax.set_xlim() takes "data" specific values, whereas leg.get_window_extent reports in window pixels (I think), and even that only after the canvas has been drawn, so I'm not sure how I can get the legend "width" in a way that I can use similar to above.
You can save the figure once to get the real legend location, and then use transData.inverted() to transform screen coordinate to data coordinate.
import pylab as pl
ax = pl.subplot(111)
pl.plot(pl.randn(1000), pl.randn(1000), label="ok")
leg = pl.legend()
pl.savefig("test.png") # save once to get the legend location
x,y,w,h = leg.get_window_extent().bounds
# transform from screen coordinate to screen coordinate
tmp1, tmp2 = ax.transData.inverted().transform([0, w])
print abs(tmp1-tmp2) # this is the with of legend in data coordinate
pl.savefig("test.png")