Install Instructions#

There are different ways to install MDIO:


We strongly recommend using a virtual environment venv or conda to avoid potential conflicts with other Python packages.

Using pip and virtualenv#

Install the 64-bit version of Python 3 from

Then we can create a venv and install MDIO.

$ python -m venv mdio-venv
$ mdio-venv/Scripts/activate
$ pip install -U multidimio

To check if installation was successful see checking installation.

You can also install some optional dependencies (extras) like this:

$ pip install multidimio[distributed]
$ pip install multidimio[cloud]
$ pip install multidimio[lossy]

distributed installs Dask for parallel, distributed processing.
cloud installs fsspec backed I/O libraries for AWS’ S3, Google’s GCS, and Azure ABS.
lossy will install the ZFPY library for lossy chunk compression.

Using conda#

MDIO can also be installed in a conda environment.


MDIO is hosted in the conda-forge channel. Make sure to always provide the -c conda-forge when running conda install or else it won’t be able to find the package.

We first run the following to create and activate an environment:

$ conda create -n mdio-env
$ conda activate mdio-env

Then we can to install with conda:

$ conda install -c conda-forge multidimio

The above command will install MDIO into your conda environment.


MDIO extras must be installed separately when using conda.

Checking Installation#

After installing MDIO, run the following:

$ python -c "import mdio; print(mdio.__version__)"

You should see the version of MDIO printed to the screen.

Building from Source#

All dependencies of MDIO are Python packages, so the build process is very simple. To install from source, we need to clone the repo first and then install locally via pip.

$ git clone
$ cd mdio-python
$ pip install .

We can also install the extras in a similar way, for example:

$ pip install .[cloud]

If you want an editable version of MDIO then we could install it with the command below. This does allow you to make code changes on the fly.

$ pip install --editable .[cloud]

To check if installation was successful see checking installation.