Importing Python Modules

Hello mates! This is a new Julia user. I was caught by a trouble to import a Py module “graphviz” to Julia.

I’ve checked that I am referring to the correct path of Py


and I’ve added my target module by Conda

using Conda

and I confirm in the library folder that it does exist in


But once I import the module

using PyCall

it comes up with error

ERROR: PyError (PyImport_ImportModule

The Python package graphviz could not be imported by pyimport. Usually this means
that you did not install graphviz in the Python version being used by PyCall.

PyCall is currently configured to use the Python version at:


and you should use whatever mechanism you usually use (apt-get, pip, conda,
etcetera) to install the Python package containing the graphviz module.

One alternative is to re-configure PyCall to use a different Python
version on your system: set ENV[“PYTHON”] to the path/name of the python
executable you want to use, run“PyCall”), and re-launch Julia.

Another alternative is to configure PyCall to use a Julia-specific Python
distribution via the Conda.jl package (which installs a private Anaconda
Python distribution), which has the advantage that packages can be installed
and kept up-to-date via Julia. As explained in the PyCall documentation,
set ENV[“PYTHON”]=“”, run“PyCall”), and re-launch Julia. Then,
To install the graphviz module, you can use pyimport_conda("graphviz", PKG),
where PKG is the Anaconda package that contains the module graphviz,
or alternatively you can use the Conda package directly (via
using Conda followed by Conda.add etcetera).

) <class ‘ModuleNotFoundError’>
ModuleNotFoundError(“No module named ‘graphviz’”)

[1] pyimport(name::String)
@ PyCall ~/.julia/packages/PyCall/ilqDX/src/PyCall.jl:558
[2] top-level scope
@ REPL[2]:1

It says the module is not found. What’s the problem?


It might be simpler for you to look into PythonCall.jl, which is newer. It downloads Python for your if needed, and takes care of dependencies using Conda (actually compatible micromamba clone).

1 Like

Not a direct answer to your question - but there is a Julia wrapper around GraphViz, you can install it from the REPL by typing ] to enter the package manager, and then type add GraphViz to add it. You type the backspace key on an empty line in the package manager to get out.

You may find the Julia-native library to be a lot more stable in your code.

In general, you may find that simply doing a web search for a library will wind up giving you the name of a Julia package for it, the community has a very diverse set of add-on packages.


Note that by default, the Conda package installs its own conda distribution, independent of the one you’ve apparently installed into /Users/whao/anaconda3, unless you configured Conda.jl explicitly to use that environment.

But more importantly, you are installing the wrong package.

I confirm in the library folder that it does exist in /Users/whao/anaconda3/lib/graphviz

That only confirms that the C library is installed, not the Python module. As explained in this Stackoverflow question, the graphviz anaconda package is only the C library and executable, not the Python wrapper module.

If you want the Python wrapper module then you need to install python-graphviz (anaconda) or pygraphviz (conda-forge).


Thank you Palli, I’ve settled it!

Thank you frylock, I gotta understand more about that!

Thank you stevengj, I follow the SF comment and solve it.

Also, depending on what you are doing with GraphViz, the file format is actually quite readable. The API is nice to have in Python, Julia, whatever … but GraphViz is even a great command-line tool.

You can try it out online without even installing the package and there are lots of examples of how to write input files (*.dot files). Once you get familiar with all of the neat options, you can work some real magic with it in Julia.


Cool, I’ve got the point, thanks!