I have an issue where some mathtext formatting is making some labels take up more vertical space than others, which causes them to not line up when placed in two columns in the legend. This is particularly important because the rows are also used to indicate related data.
Here is an example:
import matplotlib as mpl import matplotlib.pyplot as plt import matplotlib.mathtext as mathtext mpl.rc("font", family="Times New Roman",weight='normal') plt.rcParams.update({'mathtext.default': 'regular' }) plt.plot(1,1, label='A') plt.plot(2,2, label='B') plt.plot(3,3, label='C') plt.plot(4,4,label='$A_{x}^{y}$') plt.plot(5,5,label='$B_{x}^{y}$') plt.plot(6,6,label='$C_{x}^{y}$') plt.legend(fontsize='xx-large', ncol=2) plt.show()
This generates a figure like so:
For a while, I was able to "fake it" a bit by adding some empty subscripts and superscripts, however this only works when the plot is exported to pdf. It does not appear to work when exporting to png. How can I spread out the first column of labels so that they line up with the second column?
To change the position of a legend in Matplotlib, you can use the plt. legend() function. The default location is “best” – which is where Matplotlib automatically finds a location for the legend based on where it avoids covering any data points.
MATLAB has introduced native support for multiple columns in legend from version 2018a. Just add 'NumColumns',desired_number at the end of the legend() command. Additionally, the orientation of the legend entries can be changed from top-to-bottom to left-to-right.
Change order of items in the legend The above order of elements in the legend region can be changed by the gca method that uses another sub-method called get_legend_handles_labels method. These handles and labels lists are passed as parameters to legend method with order of indexes.
You could set the handleheight
keyword argument to a number which is just large enough that the height of the handle is larger than the space taken by the font. This makes the text appear aligned. Doing so may require to set the labelspacing
to a small number, in order not to make the legend appear too big.
plt.legend(fontsize='xx-large', ncol=2,handleheight=2.4, labelspacing=0.05)
The drawback of this method, as can be seen in the picture, is that the lines shift upwards compared to the text's baseline. It would probably depent on the usage case if this is acceptable or not.
In case it is not, one needs to dig a little deeper. The following subclasses HandlerLine2D
(which is the Handler for Lines) in order to set a slightly different position to the lines. Depending on the total legend size, font size etc. one would need to adapt the number xx
in the SymHandler
class.
from matplotlib.legend_handler import HandlerLine2D import matplotlib.lines class SymHandler(HandlerLine2D): def create_artists(self, legend, orig_handle,xdescent, ydescent, width, height, fontsize, trans): xx= 0.6*height return super(SymHandler, self).create_artists(legend, orig_handle,xdescent, xx, width, height, fontsize, trans) leg = plt.legend(handler_map={matplotlib.lines.Line2D: SymHandler()}, fontsize='xx-large', ncol=2,handleheight=2.4, labelspacing=0.05)
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With