Matplotlib: textcolor doesn't show any colors

Question

I'm using matplotlib 3.4.3, ubuntu 20.04, and Python 3.8 .I'm trying to change the label color with multicolors, but it shows black color without giving any error. I have installed all packages and tried differents examples without any success. Also, I have multiple labels that need different colors which makes it difficult to arrange them manually.

sudo apt-get install texlive-latex-base texlive-fonts-recommended texlive-fonts-extra texlive-latex-extra

import matplotlib.pyplot as plt


plt.rcParams['text.usetex'] = True
plt.rcParams['text.latex.preamble'] =r"\usepackage{xcolor} "
_, ax = plt.subplots()

plt.plot([0, 1], [0, 1], 'r')
plt.plot([0, 1], [0, 2], 'b')


plt.ylabel(r"\color{blue}{y} "+r"\textcolor{red}{label} ")

plt.savefig('test.pdf')
plt.show()

possibly helpful: https://matplotlib.org/stable/gallery/text_labels_and_annotations/rainbow_text.html#sphx-glr-gallery-text-labels-and-annotations-rainbow-text-py — tmdavison, Oct 18 '21 at 09:43
Thanks, I have tried and it is on the link shown in the questions and it works but it is not matched what I need because every point in the loop needs to have a label with different colors. — I_Al-thamary, Oct 18 '21 at 09:46
@tmdavison how can make it colors the `plt.ylabel` because it shows the text in the figures not in the label. — I_Al-thamary, Oct 18 '21 at 10:02
It seems to me that the _y_ label is not rendered by TeX, look at the typeface that was used. — gboffi, Oct 18 '21 at 10:57
it is latex; it's using the sans-serif font, while the tick labels are using the serif font. try adding some math-mode text in the y label to check (e.g. put something between `$$`) — tmdavison, Oct 18 '21 at 12:08
but this seems to be a known issue: https://github.com/matplotlib/matplotlib/issues/6724 — tmdavison, Oct 18 '21 at 12:12

score 4 · Answer 1 · answered Oct 18 '21 at 12:18

4

This appears to be a known issue with many of the matplotlib backends. For example, see here: https://github.com/matplotlib/matplotlib/issues/6724

One potential solution, as suggested here, is to save the figure in ps format, and then convert to pdf later.

For example:

import matplotlib.pyplot as plt


plt.rcParams['text.usetex'] = True
plt.rcParams['text.latex.preamble'] =r"\usepackage{xcolor} "
_, ax = plt.subplots()

plt.plot([0, 1], [0, 1], 'r')
plt.plot([0, 1], [0, 2], 'b')


plt.ylabel(r"\color{blue}{y} "+r"\textcolor{red}{label} ")

plt.savefig('test.ps')

and then, in a terminal

ps2pdf test.ps

Below is a screenshot of the resulting test.pdf file:

Note: tested on python 3.7.10, matplotlib 3.4.2, MacOS 10.15.7

answered Oct 18 '21 at 12:18

tmdavison

64,360
12
187
165

you can also use the pgf backend: `plt.savefig("test.pdf", use='pgf')` works just fine – Stef Oct 18 '21 at 13:04
1

Thanks for your answers but it shows black. `plt.savefig("test.pdf", use='pgf')` – I_Al-thamary Oct 18 '21 at 14:55
The codes in the link above use `ps`, and I have more than `34 figures`. It will be difficult to covert them. – I_Al-thamary Oct 18 '21 at 14:58
Is there any way to convert all of them to pdf. – I_Al-thamary Oct 18 '21 at 15:00
2

In a bash terminal: `for i in *.ps; do ps2pdf $i ; done` – tmdavison Oct 18 '21 at 18:37
@tmdavison Thanks, I will try it. – I_Al-thamary Oct 18 '21 at 20:44

score 1 · Accepted Answer · answered Oct 19 '21 at 05:51

The answers by tmdavison only support the ps format. So, I tried to solve the problem to accept any format. To show how I solve this problem, I provide a dataset that makes the whole idea sample to understand. In my case, I have more than 34 figures with different points, and every point needs to be described. I use this code, and I added the x=min(x-axis) and y=median(y-axis).

Here is the content of the emp.csv

import pandas as pd
from matplotlib import pyplot as plt
from matplotlib.transforms import Affine2D
import os
cwd = os.path.dirname(__file__)

def rainbow_text(x, y, strings, colors, orientation='vertical',
                 ax=None, **kwargs):
    if ax is None:
        ax = plt.gca()
    t = ax.transData
    canvas = ax.figure.canvas

    assert orientation in ['horizontal', 'vertical']
    if orientation == 'vertical':
        kwargs.update(rotation=90, verticalalignment='bottom')

    for s, c in zip(strings, colors):
        text = ax.text(x, y, s + " ", color=c, transform=t, **kwargs)

        # Need to draw to update the text position.
        text.draw(canvas.get_renderer())
        ex = text.get_window_extent()
        if orientation == 'horizontal':
            t = text.get_transform() + Affine2D().translate(ex.width, 0)
        else:
            t = text.get_transform() + Affine2D().translate(0, ex.height)
df = pd.read_csv('emp.csv', error_bad_lines=False)
colors_ = ['black', 'red', 'black', 'red']
i=0
for row in df.itertuples(index=True, name='Pandas'):

    plt.scatter(getattr(row, "sal"), getattr(row, "inc"), color = 'b', s=10)
    word = ['First name=', str(getattr(row, "first_name")), 'Last name=', str(getattr(row, "last_name"))]
    rainbow_text(df["sal"].min()-8.3, df["inc"].median()-1.2, word, colors_, size=5)
    word = ['age=', str(getattr(row, "age")), 'gender=', str(getattr(row, "gender"))]
    rainbow_text(df["sal"].min()-7.8, df["inc"].median()-1.2, word, colors_, size=5)
    plt.savefig(os.path.join(cwd, 'fig/Test_fig_' + str(i) + '.pdf'), format='pdf', dpi=600, bbox_inches='tight')
    i += 1
    plt.show()

Matplotlib: textcolor doesn't show any colors

2 Answers2