Install matplotlib In A Docker Container

matplotlib is an excellent library for creating graphs and visualizations in Python. For example, I used it to generate the performance graphs in my merging article, and internally, we use it now and again for visualizing any metrics we produce. It is a bit hard to install inside a docker container, though.

Installing Matplotlib in Alpine Linux

On Alpine, or an Alpine-based docker image, it's possible to install matplotlib; however it will involve compiling it from source as pip does not provide any pre-compiled binaries -- this will take quite a bit of time. If you don't mind compiling from source, you will need to have its dependencies in place to make this work:

FROM python:3.6-alpine
 RUN apk add g++ jpeg-dev zlib-dev libjpeg make
 RUN pip3 install matplotlib

Installing Matplotlib in Ubuntu

On Ubuntu, or a Ubuntu-based docker image, the process is much simpler:

FROM ubuntu:20.10
 RUN apt-get update && apt-get install -y python3 python3-pip
 RUN pip3 install matplotlib

In either case, after you've installed it, you can quickly generate great graphs and visualizations:

import numpy as np
from scipy.interpolate import splprep, splev

import matplotlib.pyplot as plt
from matplotlib.path import Path
from matplotlib.patches import PathPatch

N = 400
t = np.linspace(0, 3 * np.pi, N)
r = 0.5 + np.cos(t)
x, y = r * np.cos(t), r * np.sin(t)
fig, ax = plt.subplots()
ax.plot(x, y)
plt.xlabel("X value")
plt.ylabel("Y value")
plt.savefig('1.png')

Appendix: Alpine vs Ubuntu Pip Install

Why is the Ubuntu process fast and simple and the Alpine process slow? The reason is glibc. The pip wheels for matplotlib are compiled c/c++ programs that dynamically link to glibc and Alpine does not have glibc.

Alpine tries to stay small and so uses musl-libc instead. Unfortunately, this means compiling from source on Alpine, which can be a lengthy process.

Pip looks first for a wheel with the correct binaries, if it can't find one, it tries to compile the binaries from the c/c++ source and links them against musl. In many cases, this won't even work unless you have the python headers from python3-dev or build tools like make.

Now the silver lining, as others have mentioned, there are apk packages with the proper binaries provided by the community, using these will save you the (sometimes lengthy) process of building the binaries.