Download and install¶
This page covers the necessary steps to install Nipype.
1. Install Nipype¶
Getting Nipype to run on your system is rather straightforward. And there are multiple ways to do the installation:
Using docker¶
You can follow the Nipype tutorial
You can pull the
nipype/nipype
image from Docker hub:docker pull nipype/nipype
You may also build custom docker containers with specific versions of software using Neurodocker (see the Neurodocker Tutorial).
Using conda¶
If you have conda, miniconda or anaconda on your system, than installing Nipype can be done with just the following command:
conda install --channel conda-forge nipype
It is possible to list all of the versions of nipype available on your platform with:
conda search nipype --channel conda-forge
For more information, please see https://github.com/conda-forge/nipype-feedstock.
Using Pypi¶
The installation process is similar to other Python packages.
If you already have a Python environment set up, you can do:
pip install nipype
If you want to install all the optional features of nipype
, use the following command:
pip install nipype[all]
While all
installs everything, one can also install select components as listed below:
'doc': ['Sphinx>=1.4', 'matplotlib', 'pydotplus', 'pydot>=1.2.3'],
'tests': ['pytest-cov', 'codecov'],
'nipy': ['nitime', 'nilearn', 'dipy', 'nipy', 'matplotlib'],
'profiler': ['psutil'],
'duecredit': ['duecredit'],
'xvfbwrapper': ['xvfbwrapper'],
Debian and Ubuntu¶
Add the NeuroDebian repository and install the python-nipype
package using apt-get
or your favorite package manager:
apt-get install python-nipype
Using Github¶
To make sure that you really have the newest version of Nipype on your system, you can run the pip
command with a flag that points to the github repo:
pip install git+https://github.com/nipy/nipype#egg=nipype
Mac OS X¶
The easiest way to get nipype running on Mac OS X is to install Miniconda and follow the instructions above. If you have a non-conda environment you can install nipype by typing:
pip install nipype
Note that the above procedure may require the availability of gcc on your system path to compile the traits package.
From source¶
The most recent release is found here: https://github.com/nipy/nipype/releases/latest
For previous versions: prior downloads
If you downloaded the source distribution named something like
nipype-x.y.tar.gz
, then unpack the tarball, change into thenipype-x.y
directory and install nipype using:pip install .
Note: Depending on permissions you may need to use sudo
.
2. Interface Dependencies¶
Nipype provides wrappers around many neuroimaging tools and contains some algorithms. These tools will need to be installed for Nipype to run. You can create containers with different versions of these tools installed using Neurodocker (see the Neurodocker Tutorial).
3. Testing the install¶
The best way to test the install is checking nipype’s version and then running the tests:
python -c "import nipype; print(nipype.__version__)"
python -c "import nipype; nipype.test(doctests=False)"
The test will create a lot of output, but if all goes well you will see at the end something like this:
----------------------------------------------------------------------
2091 passed, 68 skipped, 7 xfailed, 1 warnings in 236.94 seconds
The number of tests and time will vary depending on which interfaces you have installed on your system.
Don’t worry if some modules are being skipped or marked as xfailed. As long as no main modules cause any problems, you’re fine. The number of tests and time will vary depending on which interfaces you have installed on your system. But if you receive an OK, errors=0 and failures=0 then everything is ready.