Setup: Anaconda 3 (Win10 64), Spyder 4 and Python 3.7. The IPython Graphics setting is default (Inline).I'm still a new to Python but I've looked around and have not found an answer that solves my problem so far. Thanks everyone in advance.
So in this setup, whenever I create a plot using matplotlib, it appears in the plot pane of Spyder. e.g.
import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
df = pd.DataFrame(np.random.randint(0,100,size=(100, 1)), columns=list('A'))
bp = df.boxplot(column = 'A')
creates a boxplot. Now, if I want to add a title to the plot, the code would be
bp.set_title("This Title")
This is where I'm getting some problems. If I run the entire block together
df = pd.DataFrame(np.random.randint(0,100,size=(100, 1)), columns=list('A'))
bp = df.boxplot(column = 'A')
bp.set_title("This Title")
then I get a box plot with "This Title" as the title, showing up in the plot pane,
which is what I want.
However, if I run the above code line by line in the IPython console, the 2nd line will produce a boxplot as expected, but the 3rd line will not have an effect on the image in the plot pane, so the image in the plot pane still do not have a title
Now,if i go to Tools > Preference >IPython Console > Graphics and set the graphics backend to Automatic instead of the default Inline, then when I run the code in the Console line by line, I get an image that pops up in another window, and that it does update/refreshes based on new lines entered into the console. I understand that the inline plots are supposed to be static, but I thought I saw another post where someone said that it is possible to update inline plots? So now my questions are:
Do plots only update/refresh by line codes in the IPython console if the Graphics Backend is not static like inline?
Why do I get different result when I run code blocks vs line by line?
If it is possible to update the inline plots (preferably in the plot pane of Spyder), how do you do it? I've tried various methods to redraw the plots,for example
plt.show()
plt.draw()
bp.get_figure().canvas.draw()
but none of these updates the image in the plot pane. I figured that even if I can't update the image, I should at least be able to redraw it (i.e a 2nd image appears in the plot pane with the update characteristics). But nothing I've tried worked so far. Please advise and thanks again.
(Spyder maintainer here) About your questions:
Do plots only update/refresh by line codes in the IPython console if the Graphics Backend is not static like inline?
Correct.
Why do I get different result when I run code blocks vs line by line?
Because when you run code cells (which is what I think you mean by "code blocks") your plot is shown at the end of that code and hence it takes all modifications you've done to it in intermediate lines.
If it is possible to update the inline plots (preferably in the plot pane of Spyder), how do you do it?
No, it's not possible. As you correctly mentioned above, inline plots are static images, so they can't be modified.
Related
I'm embedding certain graphs into my tkinter GUI, and I noticed I have a lot of "empty space" around the figure.
There are 3 frames in the picture, first containing buttons above the figure, second containing the canvas that contains the figure, and third containing the test.
I've checked the graphs by plotting them outside of TkAgg and there is minimum empty space, so I suspect something goes wrong when I call FigureCanvasTkAgg.
Is it possible to control how much a matplotlib Figure fills Canvas widget?
code below:
canvas=FigureCanvasTkAgg(figure,master=canvasframe)
widget=canvas.get_tk_widget()
widget.pack()
Found the error. Error was in adding subplots to figure, I added with:
.add_subplots(221)
instead of
.add_subplots(121)
Really really basic error.
I have a script which plots some pandas data, and then either shows the plot interactively with plt.show(), or saves it to a file with plt.savefig(args.out).
import matplotlib.pyplot as plt
# set up the dataframe here
ax = df.plot.line(x=0, title=args.title, figsize=(12,8), grid=True, **kwargs)
if (args.out):
vprint("Saving figure to ", args.out, "...")
plt.savefig(args.out)
else:
vprint("Showing interactive plot...")
plt.show()
The question is, does the default matplotlib backend matter for the scenario where I save to a file with savefig? It definitely matters in the other case since it's used to display the interactive plot, but if I call savefig is another backend used entirely?
When showing a figure, the backend obviously matters, because it provides two things:
The renderer to draw the image
The GUI within which the image is shown.
When saving a figure, only the former matters. However, matplotlib provides a multitude of export formats. At the end, the chosen backend will determine what to do when a figure is saved, and in most cases, will use one of the existing non-interactive backends to produce the output file.
Some examples:
TkAgg will use the tkinter GUI to show a figure. For saving a png figure, it will fall back to the basic Agg backend to produce the png file. For saving an svg file, it will fall back to the svg backend, for saving a pdf it will fallback to the pdf backend, etc.
TkCairo, will use the tkinter GUI to show a figure. For saving a png figure, it will fall back to the basic Cairo backend to produce the png file. For the rest, same as above.
Qt5Agg will use the PyQt GUI to show a figure. For png will fall back to Agg. For others same as above.
similar for other backends.
I am writing this script in Spyder (Python 3.5) and I want it to do this:
1) Plot something
2) Allow me to pick some values from the plot
3) Store those values into a variable
4) Do something with that variable
I have checked this thread: Store mouse click event coordinates with matplotlib and modified the function presented there for my own code. The problem I have is that spyder seems to ignore the interactive plot and runs the whole script at once, without waiting for me to pick any values from the plot. As I am using the values for further calculations, I obviously get an error from this. I have even tried to set an input('Press enter to continue...') after the plot, to see if it made it stop and wait for my pickings, but it does not work either.
When I run the script step by step, it works fine, I get the plot, pick my values, print the variable and find all of them in there and use them afterwards. So the question is: how can I make it work when I run the whole script?
Here is my code:
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.pyplot import plot as plot
def onpick(event):
ymouse = event.ydata
print ('y of mouse: {:.2f}'.format(ymouse))
times.append(ymouse)
if len(times)==5:
f.canvas.mpl_disconnect(cid)
return times
#
t=np.arange(1000)
y=t**3
f=plt.figure(1)
ax=plt.gca()
ax.plot(t,y,picker=5)
times=[]
cid=f.canvas.mpl_connect('button_press_event',onpick)
plt.show()
#Now do something with times
mtimes=np.mean(times)
print(mtimes)
(Spyder maintainer here) I think to solve this problem you need to go to
Preferences > IPython console > Graphics
and turn off the option called Activate support. That will make your script to block the console when a plot is run, so you can capture the mouse clicks you need on it.
The only problem is you need to run
In [1]: %matplotlib qt5
before starting to run your code because Spyder doesn't that for you anymore.
I am plotting mutliple figures using Matplotlib using Python 3.4.
When the multiple figures are open and I close the windows closing the first figure last (ie once all other figures are closed) python does not crash.
If, however, I close the first figure that was plotted first and then close the rest Python crashes.
It seems as though you need to close the windows in such an order that the first window that was opened must be closed last. Has anyone else experienced and is there a solution?
Here is a trivial example code that can be used to verify:
import matplotlib.pyplot as plt
plt.figure(1) # the first figure
plt.plot([1,2,3])
plt.figure(2) # a second figure
plt.plot([4,5,6])
plt.show()
As discussed on the IPython bug tracker this is a bug in a TCL/TK library which is shipped with python 3.4 on windows.
Changing the backend to Qt works around the problem by using a different gui framework.
The way I have managed to solve this issue has been to use Qt4 as the matplotlib backend.
Simply add the following two lines of code after importing matplotlib.
import matplotlib as mpl
mpl.rcParams['backend'] = "qt4agg"
mpl.rcParams['backend.qt4'] = "PySide"
This is what I do on Python 3 and have no plot closing errors
I use this to do effectively what should be done by plt.close('all'):
def closeall(): # this closes all figures in reverse order
l = plt.get_fignums()
l.reverse()
for a in l:
plt.close(a)
Whenever you are plotting multiple figures, do not use plt.show, create your figures separately in a figure instance and add an Axes using add_subplot. Here is an example:
import matplotlib.pyplot as plt
fig1 = plt.figure()
ax1 = fig1.add_subplot(211) # the first subplot in the first figure
ax1.plot([1,2,3])
ax2 = fig1.add_subplot(212) # the second subplot in the first figure
ax2.plot([4,5,6])
plt.suptitle('Easy as 1,2,3')
fig1.show()
fig2 = plt.figure()
ax3 = fig2.add_subplot(211) # the first subplot in the second figure
ax3.plot([4,5,6])
ax4 = fig2.add_subplot(212) # the second subplot in the second figure
ax4.plot([4,5,6])
plt.suptitle('Easy as 1,2,3')
fig2.show()
By doing so you can still use your python shell even when the plots are active. This is the best way to plot multiple plots.
Writing in Python 2.7 using pyQt 4.8.5:
How may I alter the background and graph-area (foreground?) of a Matplotlib widget? I would like to make the background of the graph widget 'light gray' (same as the background colour of the GUI), and I would like to make the graph-area (see below) black.
I'm new to GUI programming with pyQt and would like to achieve this:
my code:
self.ui.graph.axes.clear()
self.ui.graph.axes.hold(True)
self.ui.graph.axes.plot(self.Value,'r-')
self.ui.graph.axes.grid()
self.ui.graph.draw()
This should do it:
ax = self.ui.graph.axes
ax.set_axis_bgcolor('k')
self.ui.graph.set_facecolor('none')
This did only partly work for me. The two first lines worked fine but the last did not work. The error I got was:
AttributeError: 'MatplotlibWidget' object has no attribute 'set_facecolor'
The solution was to add figure to the code:
ax = self.ui.graph.axes
ax.set_axis_bgcolor('k')
self.ui.graph.figure.set_facecolor('none')
An interesting note is that I had been trying to solve this problem by using setPalette() but this didn't work until the facecolor was set to 'none' then suddenly all the changes I had made to the palette showed up.