Python 实用宝典

Question 1

I need to add two subplots to a figure. One subplot needs to be about three times as wide as the second (same height). I accomplished this using GridSpec and the colspan argument but I would like to do this using figure so I can save to PDF. I can adjust the first figure using the figsize argument in the constructor, but how do I change the size of the second plot?

Question 2

Another way is to use the subplots function and pass the width ratio with gridspec_kw:

import numpy as np
import matplotlib.pyplot as plt 

# generate some data
x = np.arange(0, 10, 0.2)
y = np.sin(x)

# plot it
f, (a0, a1) = plt.subplots(1, 2, gridspec_kw={'width_ratios': [3, 1]})
a0.plot(x, y)
a1.plot(y, x)

f.tight_layout()
f.savefig('grid_figure.pdf')

Question 3

You can use gridspec and figure:

import numpy as np
import matplotlib.pyplot as plt 
from matplotlib import gridspec

# generate some data
x = np.arange(0, 10, 0.2)
y = np.sin(x)

# plot it
fig = plt.figure(figsize=(8, 6)) 
gs = gridspec.GridSpec(1, 2, width_ratios=[3, 1]) 
ax0 = plt.subplot(gs[0])
ax0.plot(x, y)
ax1 = plt.subplot(gs[1])
ax1.plot(y, x)

plt.tight_layout()
plt.savefig('grid_figure.pdf')

resulting plot

Question 4

Probably the simplest way is using subplot2grid, described in Customizing Location of Subplot Using GridSpec.

ax = plt.subplot2grid((2, 2), (0, 0))

is equal to

import matplotlib.gridspec as gridspec
gs = gridspec.GridSpec(2, 2)
ax = plt.subplot(gs[0, 0])

so bmu’s example becomes:

import numpy as np
import matplotlib.pyplot as plt

# generate some data
x = np.arange(0, 10, 0.2)
y = np.sin(x)

# plot it
fig = plt.figure(figsize=(8, 6))
ax0 = plt.subplot2grid((1, 3), (0, 0), colspan=2)
ax0.plot(x, y)
ax1 = plt.subplot2grid((1, 3), (0, 2))
ax1.plot(y, x)

plt.tight_layout()
plt.savefig('grid_figure.pdf')

Question 5

I used pyplot‘s axes object to manually adjust the sizes without using GridSpec:

import matplotlib.pyplot as plt
import numpy as np
x = np.arange(0, 10, 0.2)
y = np.sin(x)

# definitions for the axes
left, width = 0.07, 0.65
bottom, height = 0.1, .8
bottom_h = left_h = left+width+0.02

rect_cones = [left, bottom, width, height]
rect_box = [left_h, bottom, 0.17, height]

fig = plt.figure()

cones = plt.axes(rect_cones)
box = plt.axes(rect_box)

cones.plot(x, y)

box.plot(y, x)

plt.show()

Question 6

I’ve spent entirely too long researching how to get two subplots to share the same y-axis with a single colorbar shared between the two in Matplotlib.

What was happening was that when I called the colorbar() function in either subplot1 or subplot2, it would autoscale the plot such that the colorbar plus the plot would fit inside the ‘subplot’ bounding box, causing the two side-by-side plots to be two very different sizes.

To get around this, I tried to create a third subplot which I then hacked to render no plot with just a colorbar present. The only problem is, now the heights and widths of the two plots are uneven, and I can’t figure out how to make it look okay.

Here is my code:

from __future__ import division
import matplotlib.pyplot as plt
import numpy as np
from matplotlib import patches
from matplotlib.ticker import NullFormatter

# SIS Functions
TE = 1 # Einstein radius
g1 = lambda x,y: (TE/2) * (y**2-x**2)/((x**2+y**2)**(3/2)) 
g2 = lambda x,y: -1*TE*x*y / ((x**2+y**2)**(3/2))
kappa = lambda x,y: TE / (2*np.sqrt(x**2+y**2))

coords = np.linspace(-2,2,400)
X,Y = np.meshgrid(coords,coords)
g1out = g1(X,Y)
g2out = g2(X,Y)
kappaout = kappa(X,Y)
for i in range(len(coords)):
    for j in range(len(coords)):
        if np.sqrt(coords[i]**2+coords[j]**2) <= TE:
            g1out[i][j]=0
            g2out[i][j]=0

fig = plt.figure()
fig.subplots_adjust(wspace=0,hspace=0)

# subplot number 1
ax1 = fig.add_subplot(1,2,1,aspect='equal',xlim=[-2,2],ylim=[-2,2])
plt.title(r"$\gamma_{1}$",fontsize="18")
plt.xlabel(r"x ($\theta_{E}$)",fontsize="15")
plt.ylabel(r"y ($\theta_{E}$)",rotation='horizontal',fontsize="15")
plt.xticks([-2.0,-1.5,-1.0,-0.5,0,0.5,1.0,1.5])
plt.xticks([-2.0,-1.5,-1.0,-0.5,0,0.5,1.0,1.5])
plt.imshow(g1out,extent=(-2,2,-2,2))
plt.axhline(y=0,linewidth=2,color='k',linestyle="--")
plt.axvline(x=0,linewidth=2,color='k',linestyle="--")
e1 = patches.Ellipse((0,0),2,2,color='white')
ax1.add_patch(e1)

# subplot number 2
ax2 = fig.add_subplot(1,2,2,sharey=ax1,xlim=[-2,2],ylim=[-2,2])
plt.title(r"$\gamma_{2}$",fontsize="18")
plt.xlabel(r"x ($\theta_{E}$)",fontsize="15")
ax2.yaxis.set_major_formatter( NullFormatter() )
plt.axhline(y=0,linewidth=2,color='k',linestyle="--")
plt.axvline(x=0,linewidth=2,color='k',linestyle="--")
plt.imshow(g2out,extent=(-2,2,-2,2))
e2 = patches.Ellipse((0,0),2,2,color='white')
ax2.add_patch(e2)

# subplot for colorbar
ax3 = fig.add_subplot(1,1,1)
ax3.axis('off')
cbar = plt.colorbar(ax=ax2)

plt.show()

Question 7

Just place the colorbar in its own axis and use subplots_adjust to make room for it.

As a quick example:

import numpy as np
import matplotlib.pyplot as plt

fig, axes = plt.subplots(nrows=2, ncols=2)
for ax in axes.flat:
    im = ax.imshow(np.random.random((10,10)), vmin=0, vmax=1)

fig.subplots_adjust(right=0.8)
cbar_ax = fig.add_axes([0.85, 0.15, 0.05, 0.7])
fig.colorbar(im, cax=cbar_ax)

plt.show()

enter image description here

Note that the color range will be set by the last image plotted (that gave rise to im) even if the range of values is set by vmin and vmax. If another plot has, for example, a higher max value, points with higher values than the max of im will show in uniform color.

Question 8

You can simplify Joe Kington’s code using the axparameter of figure.colorbar() with a list of axes. From the documentation:

ax

None | parent axes object(s) from which space for a new colorbar axes will be stolen. If a list of axes is given they will all be resized to make room for the colorbar axes.

import numpy as np
import matplotlib.pyplot as plt

fig, axes = plt.subplots(nrows=2, ncols=2)
for ax in axes.flat:
    im = ax.imshow(np.random.random((10,10)), vmin=0, vmax=1)

fig.colorbar(im, ax=axes.ravel().tolist())

plt.show()

Question 9

This solution does not require manual tweaking of axes locations or colorbar size, works with multi-row and single-row layouts, and works with tight_layout(). It is adapted from a gallery example, using ImageGrid from matplotlib’s AxesGrid Toolbox.

import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import ImageGrid

# Set up figure and image grid
fig = plt.figure(figsize=(9.75, 3))

grid = ImageGrid(fig, 111,          # as in plt.subplot(111)
                 nrows_ncols=(1,3),
                 axes_pad=0.15,
                 share_all=True,
                 cbar_location="right",
                 cbar_mode="single",
                 cbar_size="7%",
                 cbar_pad=0.15,
                 )

# Add data to image grid
for ax in grid:
    im = ax.imshow(np.random.random((10,10)), vmin=0, vmax=1)

# Colorbar
ax.cax.colorbar(im)
ax.cax.toggle_label(True)

#plt.tight_layout()    # Works, but may still require rect paramater to keep colorbar labels visible
plt.show()

Question 10

Using make_axes is even easier and gives a better result. It also provides possibilities to customise the positioning of the colorbar. Also note the option of subplots to share x and y axes.

import numpy as np
import matplotlib.pyplot as plt
import matplotlib as mpl

fig, axes = plt.subplots(nrows=2, ncols=2, sharex=True, sharey=True)
for ax in axes.flat:
    im = ax.imshow(np.random.random((10,10)), vmin=0, vmax=1)

cax,kw = mpl.colorbar.make_axes([ax for ax in axes.flat])
plt.colorbar(im, cax=cax, **kw)

plt.show()

Question 11

As a beginner who stumbled across this thread, I’d like to add a python-for-dummies adaptation of abevieiramota‘s very neat answer (because I’m at the level that I had to look up ‘ravel’ to work out what their code was doing):

import numpy as np
import matplotlib.pyplot as plt

fig, ((ax1,ax2,ax3),(ax4,ax5,ax6)) = plt.subplots(2,3)

axlist = [ax1,ax2,ax3,ax4,ax5,ax6]

first = ax1.imshow(np.random.random((10,10)), vmin=0, vmax=1)
third = ax3.imshow(np.random.random((12,12)), vmin=0, vmax=1)

fig.colorbar(first, ax=axlist)

plt.show()

Much less pythonic, much easier for noobs like me to see what’s actually happening here.

Question 12

As pointed out in other answers, the idea is usually to define an axes for the colorbar to reside in. There are various ways of doing so; one that hasn’t been mentionned yet would be to directly specify the colorbar axes at subplot creation with plt.subplots(). The advantage is that the axes position does not need to be manually set and in all cases with automatic aspect the colorbar will be exactly the same height as the subplots. Even in many cases where images are used the result will be satisfying as shown below.

When using plt.subplots(), the use of gridspec_kw argument allows to make the colorbar axes much smaller than the other axes.

fig, (ax, ax2, cax) = plt.subplots(ncols=3,figsize=(5.5,3), 
                  gridspec_kw={"width_ratios":[1,1, 0.05]})

Example:

import matplotlib.pyplot as plt
import numpy as np; np.random.seed(1)

fig, (ax, ax2, cax) = plt.subplots(ncols=3,figsize=(5.5,3), 
                  gridspec_kw={"width_ratios":[1,1, 0.05]})
fig.subplots_adjust(wspace=0.3)
im  = ax.imshow(np.random.rand(11,8), vmin=0, vmax=1)
im2 = ax2.imshow(np.random.rand(11,8), vmin=0, vmax=1)
ax.set_ylabel("y label")

fig.colorbar(im, cax=cax)

plt.show()

This works well, if the plots’ aspect is autoscaled or the images are shrunk due to their aspect in the width direction (as in the above). If, however, the images are wider then high, the result would look as follows, which might be undesired.

A solution to fix the colorbar height to the subplot height would be to use mpl_toolkits.axes_grid1.inset_locator.InsetPosition to set the colorbar axes relative to the image subplot axes.

import matplotlib.pyplot as plt
import numpy as np; np.random.seed(1)
from mpl_toolkits.axes_grid1.inset_locator import InsetPosition

fig, (ax, ax2, cax) = plt.subplots(ncols=3,figsize=(7,3), 
                  gridspec_kw={"width_ratios":[1,1, 0.05]})
fig.subplots_adjust(wspace=0.3)
im  = ax.imshow(np.random.rand(11,16), vmin=0, vmax=1)
im2 = ax2.imshow(np.random.rand(11,16), vmin=0, vmax=1)
ax.set_ylabel("y label")

ip = InsetPosition(ax2, [1.05,0,0.05,1]) 
cax.set_axes_locator(ip)

fig.colorbar(im, cax=cax, ax=[ax,ax2])

plt.show()

Question 13

The solution of using a list of axes by abevieiramota works very well until you use only one row of images, as pointed out in the comments. Using a reasonable aspect ratio for figsize helps, but is still far from perfect. For example:

import numpy as np
import matplotlib.pyplot as plt

fig, axes = plt.subplots(nrows=1, ncols=3, figsize=(9.75, 3))
for ax in axes.flat:
    im = ax.imshow(np.random.random((10,10)), vmin=0, vmax=1)

fig.colorbar(im, ax=axes.ravel().tolist())

plt.show()

The colorbar function provides the shrink parameter which is a scaling factor for the size of the colorbar axes. It does require some manual trial and error. For example:

fig.colorbar(im, ax=axes.ravel().tolist(), shrink=0.75)

Question 14

To add to @abevieiramota’s excellent answer, you can get the euqivalent of tight_layout with constrained_layout. You will still get large horizontal gaps if you use imshow instead of pcolormesh because of the 1:1 aspect ratio imposed by imshow.

import numpy as np
import matplotlib.pyplot as plt

fig, axes = plt.subplots(nrows=2, ncols=2, constrained_layout=True)
for ax in axes.flat:
    im = ax.pcolormesh(np.random.random((10,10)), vmin=0, vmax=1)

fig.colorbar(im, ax=axes.flat)
plt.show()

Question 15

I noticed that almost every solution posted involved ax.imshow(im, ...) and did not normalize the colors displayed to the colorbar for the multiple subfigures. The im mappable is taken from the last instance, but what if the values of the multiple im-s are different? (I’m assuming these mappables are treated in the same way that the contour-sets and surface-sets are treated.) I have an example using a 3d surface plot below that creates two colorbars for a 2×2 subplot (one colorbar per one row). Although the question asks explicitly for a different arrangement, I think the example helps clarify some things. I haven’t found a way to do this using plt.subplots(...) yet because of the 3D axes unfortunately.

If only I could position the colorbars in a better way… (There is probably a much better way to do this, but at least it should be not too difficult to follow.)

import matplotlib
from matplotlib import cm
import matplotlib.pyplot as plt
import numpy as np
from mpl_toolkits.mplot3d import Axes3D

cmap = 'plasma'
ncontours = 5

def get_data(row, col):
    """ get X, Y, Z, and plot number of subplot
        Z > 0 for top row, Z < 0 for bottom row """
    if row == 0:
        x = np.linspace(1, 10, 10, dtype=int)
        X, Y = np.meshgrid(x, x)
        Z = np.sqrt(X**2 + Y**2)
        if col == 0:
            pnum = 1
        else:
            pnum = 2
    elif row == 1:
        x = np.linspace(1, 10, 10, dtype=int)
        X, Y = np.meshgrid(x, x)
        Z = -np.sqrt(X**2 + Y**2)
        if col == 0:
            pnum = 3
        else:
            pnum = 4
    print("\nPNUM: {}, Zmin = {}, Zmax = {}\n".format(pnum, np.min(Z), np.max(Z)))
    return X, Y, Z, pnum

fig = plt.figure()
nrows, ncols = 2, 2
zz = []
axes = []
for row in range(nrows):
    for col in range(ncols):
        X, Y, Z, pnum = get_data(row, col)
        ax = fig.add_subplot(nrows, ncols, pnum, projection='3d')
        ax.set_title('row = {}, col = {}'.format(row, col))
        fhandle = ax.plot_surface(X, Y, Z, cmap=cmap)
        zz.append(Z)
        axes.append(ax)

## get full range of Z data as flat list for top and bottom rows
zz_top = zz[0].reshape(-1).tolist() + zz[1].reshape(-1).tolist()
zz_btm = zz[2].reshape(-1).tolist() + zz[3].reshape(-1).tolist()
## get top and bottom axes
ax_top = [axes[0], axes[1]]
ax_btm = [axes[2], axes[3]]
## normalize colors to minimum and maximum values of dataset
norm_top = matplotlib.colors.Normalize(vmin=min(zz_top), vmax=max(zz_top))
norm_btm = matplotlib.colors.Normalize(vmin=min(zz_btm), vmax=max(zz_btm))
cmap = cm.get_cmap(cmap, ncontours) # number of colors on colorbar
mtop = cm.ScalarMappable(cmap=cmap, norm=norm_top)
mbtm = cm.ScalarMappable(cmap=cmap, norm=norm_btm)
for m in (mtop, mbtm):
    m.set_array([])

# ## create cax to draw colorbar in
# cax_top = fig.add_axes([0.9, 0.55, 0.05, 0.4])
# cax_btm = fig.add_axes([0.9, 0.05, 0.05, 0.4])
cbar_top = fig.colorbar(mtop, ax=ax_top, orientation='vertical', shrink=0.75, pad=0.2) #, cax=cax_top)
cbar_top.set_ticks(np.linspace(min(zz_top), max(zz_top), ncontours))
cbar_btm = fig.colorbar(mbtm, ax=ax_btm, orientation='vertical', shrink=0.75, pad=0.2) #, cax=cax_btm)
cbar_btm.set_ticks(np.linspace(min(zz_btm), max(zz_btm), ncontours))

plt.show()
plt.close(fig)
## orientation of colorbar = 'horizontal' if done by column

Question 16

This topic is well covered but I still would like to propose another approach in a slightly different philosophy.

It is a bit more complex to set-up but it allow (in my opinion) a bit more flexibility. For example, one can play with the respective ratios of each subplots / colorbar:

import matplotlib.pyplot as plt
import numpy as np
from matplotlib.gridspec import GridSpec

# Define number of rows and columns you want in your figure
nrow = 2
ncol = 3

# Make a new figure
fig = plt.figure(constrained_layout=True)

# Design your figure properties
widths = [3,4,5,1]
gs = GridSpec(nrow, ncol + 1, figure=fig, width_ratios=widths)

# Fill your figure with desired plots
axes = []
for i in range(nrow):
    for j in range(ncol):
        axes.append(fig.add_subplot(gs[i, j]))
        im = axes[-1].pcolormesh(np.random.random((10,10)))

# Shared colorbar    
axes.append(fig.add_subplot(gs[:, ncol]))
fig.colorbar(im, cax=axes[-1])

plt.show()

Question 17

I created a histogram plot using data from a file and no problem. Now I wanted to superpose data from another file in the same histogram, so I do something like this

n,bins,patchs = ax.hist(mydata1,100)
n,bins,patchs = ax.hist(mydata2,100)

but the problem is that for each interval, only the bar with the highest value appears, and the other is hidden. I wonder how could I plot both histograms at the same time with different colors.

Question 18

Here you have a working example:

import random
import numpy
from matplotlib import pyplot

x = [random.gauss(3,1) for _ in range(400)]
y = [random.gauss(4,2) for _ in range(400)]

bins = numpy.linspace(-10, 10, 100)

pyplot.hist(x, bins, alpha=0.5, label='x')
pyplot.hist(y, bins, alpha=0.5, label='y')
pyplot.legend(loc='upper right')
pyplot.show()

enter image description here

Question 19

The accepted answers gives the code for a histogram with overlapping bars, but in case you want each bar to be side-by-side (as I did), try the variation below:

import numpy as np
import matplotlib.pyplot as plt
plt.style.use('seaborn-deep')

x = np.random.normal(1, 2, 5000)
y = np.random.normal(-1, 3, 2000)
bins = np.linspace(-10, 10, 30)

plt.hist([x, y], bins, label=['x', 'y'])
plt.legend(loc='upper right')
plt.show()

Reference: http://matplotlib.org/examples/statistics/histogram_demo_multihist.html

EDIT [2018/03/16]: Updated to allow plotting of arrays of different sizes, as suggested by @stochastic_zeitgeist

Question 20

In the case you have different sample sizes, it may be difficult to compare the distributions with a single y-axis. For example:

import numpy as np
import matplotlib.pyplot as plt

#makes the data
y1 = np.random.normal(-2, 2, 1000)
y2 = np.random.normal(2, 2, 5000)
colors = ['b','g']

#plots the histogram
fig, ax1 = plt.subplots()
ax1.hist([y1,y2],color=colors)
ax1.set_xlim(-10,10)
ax1.set_ylabel("Count")
plt.tight_layout()
plt.show()

In this case, you can plot your two data sets on different axes. To do so, you can get your histogram data using matplotlib, clear the axis, and then re-plot it on two separate axes (shifting the bin edges so that they don’t overlap):

#sets up the axis and gets histogram data
fig, ax1 = plt.subplots()
ax2 = ax1.twinx()
ax1.hist([y1, y2], color=colors)
n, bins, patches = ax1.hist([y1,y2])
ax1.cla() #clear the axis

#plots the histogram data
width = (bins[1] - bins[0]) * 0.4
bins_shifted = bins + width
ax1.bar(bins[:-1], n[0], width, align='edge', color=colors[0])
ax2.bar(bins_shifted[:-1], n[1], width, align='edge', color=colors[1])

#finishes the plot
ax1.set_ylabel("Count", color=colors[0])
ax2.set_ylabel("Count", color=colors[1])
ax1.tick_params('y', colors=colors[0])
ax2.tick_params('y', colors=colors[1])
plt.tight_layout()
plt.show()

Question 21

As a completion to Gustavo Bezerra’s answer:

If you want each histogram to be normalized (normed for mpl<=2.1 and density for mpl>=3.1) you cannot just use normed/density=True, you need to set the weights for each value instead:

import numpy as np
import matplotlib.pyplot as plt

x = np.random.normal(1, 2, 5000)
y = np.random.normal(-1, 3, 2000)
x_w = np.empty(x.shape)
x_w.fill(1/x.shape[0])
y_w = np.empty(y.shape)
y_w.fill(1/y.shape[0])
bins = np.linspace(-10, 10, 30)

plt.hist([x, y], bins, weights=[x_w, y_w], label=['x', 'y'])
plt.legend(loc='upper right')
plt.show()

As a comparison, the exact same x and y vectors with default weights and density=True:

Question 22

You should use bins from the values returned by hist:

import numpy as np
import matplotlib.pyplot as plt

foo = np.random.normal(loc=1, size=100) # a normal distribution
bar = np.random.normal(loc=-1, size=10000) # a normal distribution

_, bins, _ = plt.hist(foo, bins=50, range=[-6, 6], normed=True)
_ = plt.hist(bar, bins=bins, alpha=0.5, normed=True)

Question 23

Here is a simple method to plot two histograms, with their bars side-by-side, on the same plot when the data has different sizes:

def plotHistogram(p, o):
    """
    p and o are iterables with the values you want to 
    plot the histogram of
    """
    plt.hist([p, o], color=['g','r'], alpha=0.8, bins=50)
    plt.show()

Question 24

It sounds like you might want just a bar graph:

Alternatively, you can use subplots.

Question 25

Just in case you have pandas (import pandas as pd) or are ok with using it:

test = pd.DataFrame([[random.gauss(3,1) for _ in range(400)], 
                     [random.gauss(4,2) for _ in range(400)]])
plt.hist(test.values.T)
plt.show()

Question 26

There is one caveat when you want to plot the histogram from a 2-d numpy array. You need to swap the 2 axes.

import numpy as np
import matplotlib.pyplot as plt

data = np.random.normal(size=(2, 300))
# swapped_data.shape == (300, 2)
swapped_data = np.swapaxes(x, axis1=0, axis2=1)
plt.hist(swapped_data, bins=30, label=['x', 'y'])
plt.legend()
plt.show()

Question 27

This question has been answered before, but wanted to add another quick/easy workaround that might help other visitors to this question.

import seasborn as sns 
sns.kdeplot(mydata1)
sns.kdeplot(mydata2)

Some helpful examples are here for kde vs histogram comparison.

Question 28

Inspired by Solomon’s answer, but to stick with the question, which is related to histogram, a clean solution is:

sns.distplot(bar)
sns.distplot(foo)
plt.show()

Make sure to plot the taller one first, otherwise you would need to set plt.ylim(0,0.45) so that the taller histogram is not chopped off.

Question 29

Also an option which is quite similar to joaquin answer:

import random
from matplotlib import pyplot

#random data
x = [random.gauss(3,1) for _ in range(400)]
y = [random.gauss(4,2) for _ in range(400)]

#plot both histograms(range from -10 to 10), bins set to 100
pyplot.hist([x,y], bins= 100, range=[-10,10], alpha=0.5, label=['x', 'y'])
#plot legend
pyplot.legend(loc='upper right')
#show it
pyplot.show()

Gives the following output:

Question 30

I am trying to plot some data from a camera in real time using OpenCV. However, the real-time plotting (using matplotlib) doesn’t seem to be working.

I’ve isolated the problem into this simple example:

fig = plt.figure()
plt.axis([0, 1000, 0, 1])

i = 0
x = list()
y = list()

while i < 1000:
    temp_y = np.random.random()
    x.append(i)
    y.append(temp_y)
    plt.scatter(i, temp_y)
    i += 1
    plt.show()

I would expect this example to plot 1000 points individually. What actually happens is that the window pops up with the first point showing (ok with that), then waits for the loop to finish before it populates the rest of the graph.

Any thoughts why I am not seeing points populated one at a time?

Question 31

Here’s the working version of the code in question (requires at least version Matplotlib 1.1.0 from 2011-11-14):

import numpy as np
import matplotlib.pyplot as plt

plt.axis([0, 10, 0, 1])

for i in range(10):
    y = np.random.random()
    plt.scatter(i, y)
    plt.pause(0.05)

plt.show()

Note some of the changes:

Call plt.pause(0.05) to both draw the new data and it runs the GUI’s event loop (allowing for mouse interaction).

Question 32

If you’re interested in realtime plotting, I’d recommend looking into matplotlib’s animation API. In particular, using blit to avoid redrawing the background on every frame can give you substantial speed gains (~10x):

#!/usr/bin/env python

import numpy as np
import time
import matplotlib
matplotlib.use('GTKAgg')
from matplotlib import pyplot as plt


def randomwalk(dims=(256, 256), n=20, sigma=5, alpha=0.95, seed=1):
    """ A simple random walk with memory """

    r, c = dims
    gen = np.random.RandomState(seed)
    pos = gen.rand(2, n) * ((r,), (c,))
    old_delta = gen.randn(2, n) * sigma

    while True:
        delta = (1. - alpha) * gen.randn(2, n) * sigma + alpha * old_delta
        pos += delta
        for ii in xrange(n):
            if not (0. <= pos[0, ii] < r):
                pos[0, ii] = abs(pos[0, ii] % r)
            if not (0. <= pos[1, ii] < c):
                pos[1, ii] = abs(pos[1, ii] % c)
        old_delta = delta
        yield pos


def run(niter=1000, doblit=True):
    """
    Display the simulation using matplotlib, optionally using blit for speed
    """

    fig, ax = plt.subplots(1, 1)
    ax.set_aspect('equal')
    ax.set_xlim(0, 255)
    ax.set_ylim(0, 255)
    ax.hold(True)
    rw = randomwalk()
    x, y = rw.next()

    plt.show(False)
    plt.draw()

    if doblit:
        # cache the background
        background = fig.canvas.copy_from_bbox(ax.bbox)

    points = ax.plot(x, y, 'o')[0]
    tic = time.time()

    for ii in xrange(niter):

        # update the xy data
        x, y = rw.next()
        points.set_data(x, y)

        if doblit:
            # restore background
            fig.canvas.restore_region(background)

            # redraw just the points
            ax.draw_artist(points)

            # fill in the axes rectangle
            fig.canvas.blit(ax.bbox)

        else:
            # redraw everything
            fig.canvas.draw()

    plt.close(fig)
    print "Blit = %s, average FPS: %.2f" % (
        str(doblit), niter / (time.time() - tic))

if __name__ == '__main__':
    run(doblit=False)
    run(doblit=True)

Output:

Blit = False, average FPS: 54.37
Blit = True, average FPS: 438.27

Question 33

I know I’m a bit late to answer this question. Nevertheless, I’ve made some code a while ago to plot live graphs, that I would like to share:

Code for PyQt4:

###################################################################
#                                                                 #
#                    PLOT A LIVE GRAPH (PyQt4)                    #
#                  -----------------------------                  #
#            EMBED A MATPLOTLIB ANIMATION INSIDE YOUR             #
#            OWN GUI!                                             #
#                                                                 #
###################################################################


import sys
import os
from PyQt4 import QtGui
from PyQt4 import QtCore
import functools
import numpy as np
import random as rd
import matplotlib
matplotlib.use("Qt4Agg")
from matplotlib.figure import Figure
from matplotlib.animation import TimedAnimation
from matplotlib.lines import Line2D
from matplotlib.backends.backend_qt4agg import FigureCanvasQTAgg as FigureCanvas
import time
import threading


def setCustomSize(x, width, height):
    sizePolicy = QtGui.QSizePolicy(QtGui.QSizePolicy.Fixed, QtGui.QSizePolicy.Fixed)
    sizePolicy.setHorizontalStretch(0)
    sizePolicy.setVerticalStretch(0)
    sizePolicy.setHeightForWidth(x.sizePolicy().hasHeightForWidth())
    x.setSizePolicy(sizePolicy)
    x.setMinimumSize(QtCore.QSize(width, height))
    x.setMaximumSize(QtCore.QSize(width, height))

''''''

class CustomMainWindow(QtGui.QMainWindow):

    def __init__(self):

        super(CustomMainWindow, self).__init__()

        # Define the geometry of the main window
        self.setGeometry(300, 300, 800, 400)
        self.setWindowTitle("my first window")

        # Create FRAME_A
        self.FRAME_A = QtGui.QFrame(self)
        self.FRAME_A.setStyleSheet("QWidget { background-color: %s }" % QtGui.QColor(210,210,235,255).name())
        self.LAYOUT_A = QtGui.QGridLayout()
        self.FRAME_A.setLayout(self.LAYOUT_A)
        self.setCentralWidget(self.FRAME_A)

        # Place the zoom button
        self.zoomBtn = QtGui.QPushButton(text = 'zoom')
        setCustomSize(self.zoomBtn, 100, 50)
        self.zoomBtn.clicked.connect(self.zoomBtnAction)
        self.LAYOUT_A.addWidget(self.zoomBtn, *(0,0))

        # Place the matplotlib figure
        self.myFig = CustomFigCanvas()
        self.LAYOUT_A.addWidget(self.myFig, *(0,1))

        # Add the callbackfunc to ..
        myDataLoop = threading.Thread(name = 'myDataLoop', target = dataSendLoop, daemon = True, args = (self.addData_callbackFunc,))
        myDataLoop.start()

        self.show()

    ''''''


    def zoomBtnAction(self):
        print("zoom in")
        self.myFig.zoomIn(5)

    ''''''

    def addData_callbackFunc(self, value):
        # print("Add data: " + str(value))
        self.myFig.addData(value)



''' End Class '''


class CustomFigCanvas(FigureCanvas, TimedAnimation):

    def __init__(self):

        self.addedData = []
        print(matplotlib.__version__)

        # The data
        self.xlim = 200
        self.n = np.linspace(0, self.xlim - 1, self.xlim)
        a = []
        b = []
        a.append(2.0)
        a.append(4.0)
        a.append(2.0)
        b.append(4.0)
        b.append(3.0)
        b.append(4.0)
        self.y = (self.n * 0.0) + 50

        # The window
        self.fig = Figure(figsize=(5,5), dpi=100)
        self.ax1 = self.fig.add_subplot(111)


        # self.ax1 settings
        self.ax1.set_xlabel('time')
        self.ax1.set_ylabel('raw data')
        self.line1 = Line2D([], [], color='blue')
        self.line1_tail = Line2D([], [], color='red', linewidth=2)
        self.line1_head = Line2D([], [], color='red', marker='o', markeredgecolor='r')
        self.ax1.add_line(self.line1)
        self.ax1.add_line(self.line1_tail)
        self.ax1.add_line(self.line1_head)
        self.ax1.set_xlim(0, self.xlim - 1)
        self.ax1.set_ylim(0, 100)


        FigureCanvas.__init__(self, self.fig)
        TimedAnimation.__init__(self, self.fig, interval = 50, blit = True)

    def new_frame_seq(self):
        return iter(range(self.n.size))

    def _init_draw(self):
        lines = [self.line1, self.line1_tail, self.line1_head]
        for l in lines:
            l.set_data([], [])

    def addData(self, value):
        self.addedData.append(value)

    def zoomIn(self, value):
        bottom = self.ax1.get_ylim()[0]
        top = self.ax1.get_ylim()[1]
        bottom += value
        top -= value
        self.ax1.set_ylim(bottom,top)
        self.draw()


    def _step(self, *args):
        # Extends the _step() method for the TimedAnimation class.
        try:
            TimedAnimation._step(self, *args)
        except Exception as e:
            self.abc += 1
            print(str(self.abc))
            TimedAnimation._stop(self)
            pass

    def _draw_frame(self, framedata):
        margin = 2
        while(len(self.addedData) > 0):
            self.y = np.roll(self.y, -1)
            self.y[-1] = self.addedData[0]
            del(self.addedData[0])


        self.line1.set_data(self.n[ 0 : self.n.size - margin ], self.y[ 0 : self.n.size - margin ])
        self.line1_tail.set_data(np.append(self.n[-10:-1 - margin], self.n[-1 - margin]), np.append(self.y[-10:-1 - margin], self.y[-1 - margin]))
        self.line1_head.set_data(self.n[-1 - margin], self.y[-1 - margin])
        self._drawn_artists = [self.line1, self.line1_tail, self.line1_head]

''' End Class '''

# You need to setup a signal slot mechanism, to 
# send data to your GUI in a thread-safe way.
# Believe me, if you don't do this right, things
# go very very wrong..
class Communicate(QtCore.QObject):
    data_signal = QtCore.pyqtSignal(float)

''' End Class '''


def dataSendLoop(addData_callbackFunc):
    # Setup the signal-slot mechanism.
    mySrc = Communicate()
    mySrc.data_signal.connect(addData_callbackFunc)

    # Simulate some data
    n = np.linspace(0, 499, 500)
    y = 50 + 25*(np.sin(n / 8.3)) + 10*(np.sin(n / 7.5)) - 5*(np.sin(n / 1.5))
    i = 0

    while(True):
        if(i > 499):
            i = 0
        time.sleep(0.1)
        mySrc.data_signal.emit(y[i]) # <- Here you emit a signal!
        i += 1
    ###
###


if __name__== '__main__':
    app = QtGui.QApplication(sys.argv)
    QtGui.QApplication.setStyle(QtGui.QStyleFactory.create('Plastique'))
    myGUI = CustomMainWindow()
    sys.exit(app.exec_())

''''''

I recently rewrote the code for PyQt5.
Code for PyQt5:

###################################################################
#                                                                 #
#                    PLOT A LIVE GRAPH (PyQt5)                    #
#                  -----------------------------                  #
#            EMBED A MATPLOTLIB ANIMATION INSIDE YOUR             #
#            OWN GUI!                                             #
#                                                                 #
###################################################################

import sys
import os
from PyQt5.QtWidgets import *
from PyQt5.QtCore import *
from PyQt5.QtGui import *
import functools
import numpy as np
import random as rd
import matplotlib
matplotlib.use("Qt5Agg")
from matplotlib.figure import Figure
from matplotlib.animation import TimedAnimation
from matplotlib.lines import Line2D
from matplotlib.backends.backend_qt5agg import FigureCanvasQTAgg as FigureCanvas
import time
import threading

class CustomMainWindow(QMainWindow):
    def __init__(self):
        super(CustomMainWindow, self).__init__()
        # Define the geometry of the main window
        self.setGeometry(300, 300, 800, 400)
        self.setWindowTitle("my first window")
        # Create FRAME_A
        self.FRAME_A = QFrame(self)
        self.FRAME_A.setStyleSheet("QWidget { background-color: %s }" % QColor(210,210,235,255).name())
        self.LAYOUT_A = QGridLayout()
        self.FRAME_A.setLayout(self.LAYOUT_A)
        self.setCentralWidget(self.FRAME_A)
        # Place the zoom button
        self.zoomBtn = QPushButton(text = 'zoom')
        self.zoomBtn.setFixedSize(100, 50)
        self.zoomBtn.clicked.connect(self.zoomBtnAction)
        self.LAYOUT_A.addWidget(self.zoomBtn, *(0,0))
        # Place the matplotlib figure
        self.myFig = CustomFigCanvas()
        self.LAYOUT_A.addWidget(self.myFig, *(0,1))
        # Add the callbackfunc to ..
        myDataLoop = threading.Thread(name = 'myDataLoop', target = dataSendLoop, daemon = True, args = (self.addData_callbackFunc,))
        myDataLoop.start()
        self.show()
        return

    def zoomBtnAction(self):
        print("zoom in")
        self.myFig.zoomIn(5)
        return

    def addData_callbackFunc(self, value):
        # print("Add data: " + str(value))
        self.myFig.addData(value)
        return

''' End Class '''


class CustomFigCanvas(FigureCanvas, TimedAnimation):
    def __init__(self):
        self.addedData = []
        print(matplotlib.__version__)
        # The data
        self.xlim = 200
        self.n = np.linspace(0, self.xlim - 1, self.xlim)
        a = []
        b = []
        a.append(2.0)
        a.append(4.0)
        a.append(2.0)
        b.append(4.0)
        b.append(3.0)
        b.append(4.0)
        self.y = (self.n * 0.0) + 50
        # The window
        self.fig = Figure(figsize=(5,5), dpi=100)
        self.ax1 = self.fig.add_subplot(111)
        # self.ax1 settings
        self.ax1.set_xlabel('time')
        self.ax1.set_ylabel('raw data')
        self.line1 = Line2D([], [], color='blue')
        self.line1_tail = Line2D([], [], color='red', linewidth=2)
        self.line1_head = Line2D([], [], color='red', marker='o', markeredgecolor='r')
        self.ax1.add_line(self.line1)
        self.ax1.add_line(self.line1_tail)
        self.ax1.add_line(self.line1_head)
        self.ax1.set_xlim(0, self.xlim - 1)
        self.ax1.set_ylim(0, 100)
        FigureCanvas.__init__(self, self.fig)
        TimedAnimation.__init__(self, self.fig, interval = 50, blit = True)
        return

    def new_frame_seq(self):
        return iter(range(self.n.size))

    def _init_draw(self):
        lines = [self.line1, self.line1_tail, self.line1_head]
        for l in lines:
            l.set_data([], [])
        return

    def addData(self, value):
        self.addedData.append(value)
        return

    def zoomIn(self, value):
        bottom = self.ax1.get_ylim()[0]
        top = self.ax1.get_ylim()[1]
        bottom += value
        top -= value
        self.ax1.set_ylim(bottom,top)
        self.draw()
        return

    def _step(self, *args):
        # Extends the _step() method for the TimedAnimation class.
        try:
            TimedAnimation._step(self, *args)
        except Exception as e:
            self.abc += 1
            print(str(self.abc))
            TimedAnimation._stop(self)
            pass
        return

    def _draw_frame(self, framedata):
        margin = 2
        while(len(self.addedData) > 0):
            self.y = np.roll(self.y, -1)
            self.y[-1] = self.addedData[0]
            del(self.addedData[0])

        self.line1.set_data(self.n[ 0 : self.n.size - margin ], self.y[ 0 : self.n.size - margin ])
        self.line1_tail.set_data(np.append(self.n[-10:-1 - margin], self.n[-1 - margin]), np.append(self.y[-10:-1 - margin], self.y[-1 - margin]))
        self.line1_head.set_data(self.n[-1 - margin], self.y[-1 - margin])
        self._drawn_artists = [self.line1, self.line1_tail, self.line1_head]
        return

''' End Class '''


# You need to setup a signal slot mechanism, to
# send data to your GUI in a thread-safe way.
# Believe me, if you don't do this right, things
# go very very wrong..
class Communicate(QObject):
    data_signal = pyqtSignal(float)

''' End Class '''



def dataSendLoop(addData_callbackFunc):
    # Setup the signal-slot mechanism.
    mySrc = Communicate()
    mySrc.data_signal.connect(addData_callbackFunc)

    # Simulate some data
    n = np.linspace(0, 499, 500)
    y = 50 + 25*(np.sin(n / 8.3)) + 10*(np.sin(n / 7.5)) - 5*(np.sin(n / 1.5))
    i = 0

    while(True):
        if(i > 499):
            i = 0
        time.sleep(0.1)
        mySrc.data_signal.emit(y[i]) # <- Here you emit a signal!
        i += 1
    ###
###

if __name__== '__main__':
    app = QApplication(sys.argv)
    QApplication.setStyle(QStyleFactory.create('Plastique'))
    myGUI = CustomMainWindow()
    sys.exit(app.exec_())

Just try it out. Copy-paste this code in a new python-file, and run it. You should get a beautiful, smoothly moving graph:

Question 34

show is probably not the best choice for this. What I would do is use pyplot.draw() instead. You also might want to include a small time delay (e.g., time.sleep(0.05)) in the loop so that you can see the plots happening. If I make these changes to your example it works for me and I see each point appearing one at a time.

Question 35

None of the methods worked for me. But I have found this Real time matplotlib plot is not working while still in a loop

All you need is to add

plt.pause(0.0001)

and then you could see the new plots.

So your code should look like this, and it will work

import matplotlib.pyplot as plt
import numpy as np
plt.ion() ## Note this correction
fig=plt.figure()
plt.axis([0,1000,0,1])

i=0
x=list()
y=list()

while i <1000:
    temp_y=np.random.random();
    x.append(i);
    y.append(temp_y);
    plt.scatter(i,temp_y);
    i+=1;
    plt.show()
    plt.pause(0.0001) #Note this correction

Question 36

The top (and many other) answers were built upon plt.pause(), but that was an old way of animating the plot in matplotlib. It is not only slow, but also causes focus to be grabbed upon each update (I had a hard time stopping the plotting python process).

TL;DR: you may want to use matplotlib.animation (as mentioned in documentation).

After digging around various answers and pieces of code, this in fact proved to be a smooth way of drawing incoming data infinitely for me.

Here is my code for a quick start. It plots current time with a random number in [0, 100) every 200ms infinitely, while also handling auto rescaling of the view:

from datetime import datetime
from matplotlib import pyplot
from matplotlib.animation import FuncAnimation
from random import randrange

x_data, y_data = [], []

figure = pyplot.figure()
line, = pyplot.plot_date(x_data, y_data, '-')

def update(frame):
    x_data.append(datetime.now())
    y_data.append(randrange(0, 100))
    line.set_data(x_data, y_data)
    figure.gca().relim()
    figure.gca().autoscale_view()
    return line,

animation = FuncAnimation(figure, update, interval=200)

pyplot.show()

You can also explore blit for even better performance as in FuncAnimation documentation.

An example from the blit documentation:

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.animation import FuncAnimation

fig, ax = plt.subplots()
xdata, ydata = [], []
ln, = plt.plot([], [], 'ro')

def init():
    ax.set_xlim(0, 2*np.pi)
    ax.set_ylim(-1, 1)
    return ln,

def update(frame):
    xdata.append(frame)
    ydata.append(np.sin(frame))
    ln.set_data(xdata, ydata)
    return ln,

ani = FuncAnimation(fig, update, frames=np.linspace(0, 2*np.pi, 128),
                    init_func=init, blit=True)
plt.show()

Question 37

I know this question is old, but there’s now a package available called drawnow on GitHub as “python-drawnow”. This provides an interface similar to MATLAB’s drawnow — you can easily update a figure.

An example for your use case:

import matplotlib.pyplot as plt
from drawnow import drawnow

def make_fig():
    plt.scatter(x, y)  # I think you meant this

plt.ion()  # enable interactivity
fig = plt.figure()  # make a figure

x = list()
y = list()

for i in range(1000):
    temp_y = np.random.random()
    x.append(i)
    y.append(temp_y)  # or any arbitrary update to your figure's data
    i += 1
    drawnow(make_fig)

python-drawnow is a thin wrapper around plt.draw but provides the ability to confirm (or debug) after figure display.

Question 38

The problem seems to be that you expect plt.show() to show the window and then to return. It does not do that. The program will stop at that point and only resume once you close the window. You should be able to test that: If you close the window and then another window should pop up.

To resolve that problem just call plt.show() once after your loop. Then you get the complete plot. (But not a ‘real-time plotting’)

You can try setting the keyword-argument block like this: plt.show(block=False) once at the beginning and then use .draw() to update.

Question 39

Here is a version that I got to work on my system.

import matplotlib.pyplot as plt
from drawnow import drawnow
import numpy as np

def makeFig():
    plt.scatter(xList,yList) # I think you meant this

plt.ion() # enable interactivity
fig=plt.figure() # make a figure

xList=list()
yList=list()

for i in np.arange(50):
    y=np.random.random()
    xList.append(i)
    yList.append(y)
    drawnow(makeFig)
    #makeFig()      The drawnow(makeFig) command can be replaced
    #plt.draw()     with makeFig(); plt.draw()
    plt.pause(0.001)

The drawnow(makeFig) line can be replaced with a makeFig(); plt.draw() sequence and it still works OK.

Question 40

If you want draw and not freeze your thread as more point are drawn you should use plt.pause() not time.sleep()

im using the following code to plot a series of xy coordinates.

import matplotlib.pyplot as plt 
import math


pi = 3.14159

fig, ax = plt.subplots()

x = []
y = []

def PointsInCircum(r,n=20):
    circle = [(math.cos(2*pi/n*x)*r,math.sin(2*pi/n*x)*r) for x in xrange(0,n+1)]
    return circle

circle_list = PointsInCircum(3, 50)

for t in range(len(circle_list)):
    if t == 0:
        points, = ax.plot(x, y, marker='o', linestyle='--')
        ax.set_xlim(-4, 4) 
        ax.set_ylim(-4, 4) 
    else:
        x_coord, y_coord = circle_list.pop()
        x.append(x_coord)
        y.append(y_coord)
        points.set_data(x, y)
    plt.pause(0.01)

Question 41

Another option is to go with bokeh. IMO, it is a good alternative at least for real-time plots. Here is a bokeh version of the code in the question:

from bokeh.plotting import curdoc, figure
import random
import time

def update():
    global i
    temp_y = random.random()
    r.data_source.stream({'x': [i], 'y': [temp_y]})
    i += 1

i = 0
p = figure()
r = p.circle([], [])
curdoc().add_root(p)
curdoc().add_periodic_callback(update, 100)

and for running it:

pip3 install bokeh
bokeh serve --show test.py

bokeh shows the result in a web browser via websocket communications. It is especially useful when data is generated by remote headless server processes.

Question 42

An example use-case to plot CPU usage in real-time.

import time
import psutil
import matplotlib.pyplot as plt

fig = plt.figure()
ax = fig.add_subplot(111)

i = 0
x, y = [], []

while True:
    x.append(i)
    y.append(psutil.cpu_percent())

    ax.plot(x, y, color='b')

    fig.canvas.draw()

    ax.set_xlim(left=max(0, i - 50), right=i + 50)
    fig.show()
    plt.pause(0.05)
    i += 1

Question 43

I’m trying to display a grayscale image using matplotlib.pyplot.imshow(). My problem is that the grayscale image is displayed as a colormap. I need the grayscale because I want to draw on top of the image with color.

I read in the image and convert to grayscale using PIL’s Image.open().convert(“L”)

image = Image.open(file).convert("L")

Then I convert the image to a matrix so that I can easily do some image processing using

matrix = scipy.misc.fromimage(image, 0)

However, when I do

figure()  
matplotlib.pyplot.imshow(matrix)  
show()

it displays the image using a colormap (i.e. it’s not grayscale).

What am I doing wrong here?

Question 44

The following code will load an image from a file image.png and will display it as grayscale.

import numpy as np
import matplotlib.pyplot as plt
from PIL import Image

fname = 'image.png'
image = Image.open(fname).convert("L")
arr = np.asarray(image)
plt.imshow(arr, cmap='gray', vmin=0, vmax=255)
plt.show()

If you want to display the inverse grayscale, switch the cmap to cmap='gray_r'.

Question 45

Try to use a grayscale colormap?

E.g. something like

imshow(..., cmap=pyplot.cm.binary)

For a list of colormaps, see http://scipy-cookbook.readthedocs.org/items/Matplotlib_Show_colormaps.html

Question 46

import matplotlib.pyplot as plt

You can also run once in your code

plt.gray()

This will show the images in grayscale as default

im = array(Image.open('I_am_batman.jpg').convert('L'))
plt.imshow(im)
plt.show()

Question 47

I would use the get_cmap method. Ex.:

import matplotlib.pyplot as plt

plt.imshow(matrix, cmap=plt.get_cmap('gray'))

Question 48

@unutbu’s answer is quite close to the right answer.

By default, plt.imshow() will try to scale your (MxN) array data to 0.0~1.0. And then map to 0~255. For most natural taken images, this is fine, you won’t see a different. But if you have narrow range of pixel value image, say the min pixel is 156 and the max pixel is 234. The gray image will looks totally wrong. The right way to show an image in gray is

from matplotlib.colors import NoNorm
...
plt.imshow(img,cmap='gray',norm=NoNorm())
...

Let’s see an example:

this is the origianl image: original

this is using defaul norm setting,which is None: wrong pic

this is using NoNorm setting,which is NoNorm(): right pic

Question 49

try this:

import pylab
from scipy import misc

pylab.imshow(misc.lena(),cmap=pylab.gray())
pylab.show()

Question 50

Use no interpolation and set to gray.

import matplotlib.pyplot as plt
plt.imshow(img[:,:,1], cmap='gray',interpolation='none')

Question 51

I’m new to Python and Matplotlib, I would like to simply apply colormap to an image and write the resulting image, without using axes, labels, titles or anything usually automatically added by matplotlib. Here is what I did:

def make_image(inputname,outputname):
    data = mpimg.imread(inputname)[:,:,0]
    fig = plt.imshow(data)
    fig.set_cmap('hot')
    fig.axes.get_xaxis().set_visible(False)
    fig.axes.get_yaxis().set_visible(False)
    plt.savefig(outputname)

It successfully removes the axis of the figure, but the figure saved presents a white padding and a frame around the actual image. How can I remove them (at least the white padding)? Thanks

Question 52

I think that the command axis('off') takes care of one of the problems more succinctly than changing each axis and the border separately. It still leaves the white space around the border however. Adding bbox_inches='tight' to the savefig command almost gets you there, you can see in the example below that the white space left is much smaller, but still present.

Note that newer versions of matplotlib may require bbox_inches=0 instead of the string 'tight' (via @episodeyang and @kadrach)

from numpy import random
import matplotlib.pyplot as plt

data = random.random((5,5))
img = plt.imshow(data, interpolation='nearest')
img.set_cmap('hot')
plt.axis('off')
plt.savefig("test.png", bbox_inches='tight')

enter image description here

Question 53

I learned this trick from matehat, here:

import matplotlib.pyplot as plt
import numpy as np

def make_image(data, outputname, size=(1, 1), dpi=80):
    fig = plt.figure()
    fig.set_size_inches(size)
    ax = plt.Axes(fig, [0., 0., 1., 1.])
    ax.set_axis_off()
    fig.add_axes(ax)
    plt.set_cmap('hot')
    ax.imshow(data, aspect='equal')
    plt.savefig(outputname, dpi=dpi)

# data = mpimg.imread(inputname)[:,:,0]
data = np.arange(1,10).reshape((3, 3))

make_image(data, '/tmp/out.png')

yields

enter image description here

Question 54

Possible simplest solution:

I simply combined the method described in the question and the method from the answer by Hooked.

fig = plt.imshow(my_data)
plt.axis('off')
fig.axes.get_xaxis().set_visible(False)
fig.axes.get_yaxis().set_visible(False)
plt.savefig('pict.png', bbox_inches='tight', pad_inches = 0)

After this code there is no whitespaces and no frame.

No whitespaces, axes or frame

Question 55

No one mentioned imsave yet, which makes this a one-liner:

import matplotlib.pyplot as plt
import numpy as np

data = np.arange(10000).reshape((100, 100))
plt.imsave("/tmp/foo.png", data, format="png", cmap="hot")

It directly stores the image as it is, i.e. does not add any axes or border/padding.

Question 56

You can also specify the extent of the figure to the bbox_inches argument. This would get rid of the white padding around the figure.

def make_image(inputname,outputname):
    data = mpimg.imread(inputname)[:,:,0]
    fig = plt.imshow(data)
    fig.set_cmap('hot')
    ax = fig.gca()
    ax.set_axis_off()
    ax.autoscale(False)
    extent = ax.get_window_extent().transformed(plt.gcf().dpi_scale_trans.inverted())
    plt.savefig(outputname, bbox_inches=extent)

Question 57

This should remove all padding and borders:

from matplotlib import pyplot as plt

fig = plt.figure()
fig.patch.set_visible(False)

ax = fig.add_subplot(111)

plt.axis('off')
plt.imshow(data)

extent = ax.get_window_extent().transformed(fig.dpi_scale_trans.inverted())
plt.savefig("../images/test.png", bbox_inches=extent)

Question 58

I found that it is all documented…

https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.axes.Axes.axis.html#matplotlib.axes.Axes.axis

My code…. “bcK” is a 512×512 image

plt.figure()
plt.imshow(bck)
plt.axis("off")   # turns off axes
plt.axis("tight")  # gets rid of white border
plt.axis("image")  # square up the image instead of filling the "figure" space
plt.show()

Question 59

I liked ubuntu’s answer, but it was not showing explicitly how to set the size for non-square images out-of-the-box, so I modified it for easy copy-paste:

import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import numpy as np

def save_image_fix_dpi(data, dpi=100):
    shape=np.shape(data)[0:2][::-1]
    size = [float(i)/dpi for i in shape]

    fig = plt.figure()
    fig.set_size_inches(size)
    ax = plt.Axes(fig,[0,0,1,1])
    ax.set_axis_off()
    fig.add_axes(ax)
    ax.imshow(data)
    fig.savefig('out.png', dpi=dpi)
    plt.show()

Saving images without border is easy whatever dpi you choose if pixel_size/dpi=size is kept.

data = mpimg.imread('test.png')
save_image_fix_dpi(data, dpi=100)

However displaying is spooky. If you choose small dpi, your image size can be larger than your screen and you get border during display. Nevertheless, this does not affect saving.

So for

save_image_fix_dpi(data, dpi=20)

The display becomes bordered (but saving works):

Question 60

The upvoted answer does not work anymore. To get it to work you need to manually add an axis set to [0, 0, 1, 1], or remove the patch under figure.

import matplotlib.pyplot as plt

fig = plt.figure(figsize=(5, 5), dpi=20)
ax = plt.Axes(fig, [0., 0., 1., 1.])
fig.add_axes(ax)
plt.imshow([[0, 1], [0.5, 0]], interpolation="nearest")
plt.axis('off')                                # same as: ax.set_axis_off()

plt.savefig("test.png")

Alternatively, you could just remove the patch. You don’t need to add a subplot in order to remove the paddings. This is simplified from Vlady’s answer below

fig = plt.figure(figsize=(5, 5))
fig.patch.set_visible(False)                   # turn off the patch

plt.imshow([[0, 1], [0.5, 0]], interpolation="nearest")
plt.axis('off')

plt.savefig("test.png", cmap='hot')

This is tested with version 3.0.3 on 2019/06/19. Image see bellow:

A much simpler thing to do is to use pyplot.imsave. For details, see luator’s answer bellow

Question 61

First, for certain image formats (i.e. TIFF) you can actually save the colormap in the header and most viewers will show your data with the colormap.

For saving an actual matplotlib image, which can be useful for adding annotations or other data to images, I’ve used the following solution:

fig, ax = plt.subplots(figsize=inches)
ax.matshow(data)  # or you can use also imshow
# add annotations or anything else
# The code below essentially moves your plot so that the upper
# left hand corner coincides with the upper left hand corner
# of the artist
fig.subplots_adjust(left=0, right=1, top=1, bottom=0, wspace=0, hspace=0)
# now generate a Bbox instance that is the same size as your
# single axis size (this bbox will only encompass your figure)
bbox = matplotlib.transforms.Bbox(((0, 0), inches))
# now you can save only the part of the figure with data
fig.savefig(savename, bbox_inches=bbox, **kwargs)

Question 62

Thanks for the awesome answers from everyone …I had exactly the same problem with wanting to plot just an image with no extra padding/space etc, so was super happy to find everyone’s ideas here.

Apart from image with no padding, I also wanted to be able to easily add annotations etc, beyond just a simple image plot.

So what I ended up doing was combining David’s answer with csnemes’ to make a simple wrapper at the figure creation time. When you use that, you don’t need any changes later with imsave() or anything else:

def get_img_figure(image, dpi):
    """
    Create a matplotlib (figure,axes) for an image (numpy array) setup so that
        a) axes will span the entire figure (when saved no whitespace)
        b) when saved the figure will have the same x/y resolution as the array, 
           with the dpi value you pass in.

    Arguments:
        image -- numpy 2d array
        dpi -- dpi value that the figure should use

    Returns: (figure, ax) tuple from plt.subplots
    """

    # get required figure size in inches (reversed row/column order)
    inches = image.shape[1]/dpi, image.shape[0]/dpi

    # make figure with that size and a single axes
    fig, ax = plt.subplots(figsize=inches, dpi=dpi)

    # move axes to span entire figure area
    fig.subplots_adjust(left=0, right=1, top=1, bottom=0, wspace=0, hspace=0)

    return fig, ax

Question 63

How can I set the y axis range of the second subplot to e.g. [0,1000] ? The FFT plot of my data (a column in a text file) results in a (inf.?) spike so that the actual data is not visible.

pylab.ylim([0,1000])

has no effect, unfortunately. This is the whole script:

# based on http://www.swharden.com/blog/2009-01-21-signal-filtering-with-python/
import numpy, scipy, pylab, random

xs = []
rawsignal = []
with open("test.dat", 'r') as f:
      for line in f:
            if line[0] != '#' and len(line) > 0:
                xs.append( int( line.split()[0] ) )
                rawsignal.append( int( line.split()[1] ) )

h, w = 3, 1
pylab.figure(figsize=(12,9))
pylab.subplots_adjust(hspace=.7)

pylab.subplot(h,w,1)
pylab.title("Signal")
pylab.plot(xs,rawsignal)

pylab.subplot(h,w,2)
pylab.title("FFT")
fft = scipy.fft(rawsignal)
#~ pylab.axis([None,None,0,1000])
pylab.ylim([0,1000])
pylab.plot(abs(fft))

pylab.savefig("SIG.png",dpi=200)
pylab.show()

Other improvements are also appreciated!

Question 64

You have pylab.ylim:

pylab.ylim([0,1000])

Note: The command has to be executed after the plot!

问题：Matplotlib不同大小的子图

回答 0

回答 1

回答 2

回答 3

问题：Matplotlib 2个子图，1个颜色条

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

回答 8

回答 9

问题：使用matplotlib在单个图表上绘制两个直方图

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

回答 8

回答 9

回答 10

回答 11

问题：如何使用matplotlib在while循环中实时绘制？

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

回答 8

回答 9

回答 10

回答 11

问题：使用matplotlib将图像显示为灰度

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

问题：Matplotlib图：删除轴，图例和空白

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

回答 8

回答 9

回答 10

问题：Python，Matplotlib，子图：如何设置轴范围？

回答 0

回答 1

回答 2

回答 3

回答 4

问题：在PyPlot中反转Y轴

回答 0

回答 1

回答 2

回答 3

回答 4

回答 5

回答 6

回答 7

回答 8

问题：为什么很多示例在Matplotlib / pyplot / python中使用`fig，ax = plt.subplots（）`

回答 0