Installation¶

Supported Hardware¶

Some components of MDTraj are written in C / C++ and make use of SSE intrinsics for maximum performance. The SSE4.1 instruction set (the latest we use) was released in 2007, so most recent x86 machines should have no problem with this requirement. Utilization of these features also requires using a suitably modern compiler. This is only really a question on Linux, where gcc >= 4.3 is a prerequisite.

Install with Conda¶

conda is a python package manager built for scientific python. Unlike easy_install or pip, it handles binaries and binary dependencies, which are critical for most scientific workflows. If you’re a conda user, you can install MDTraj by adding the omnia channel. If you’re not a conda user, you should look into it.

To install MDTraj with conda, use the following commands

$ conda config --add channels http://conda.binstar.org/omnia
$ conda install mdtraj

Install with Pip¶

On a Mac or Linux machine, just run

$ pip install mdtraj

Or, if you want the bleeding-edge source code, use

$ pip install git+git://github.com/rmcgibbo/mdtraj.git

Install from Source¶

Clone the source code repository from github:

$ git clone git://github.com/rmcgibbo/mdtraj.git

If you don’t have git, you can download the source code as a zip file from https://github.com/rmcgibbo/mdtraj/archive/master.zip. Then, in the directory containing the source code, you can install it with.

$ python setup.py install

Avoid Hassles with Anaconda or Canopy¶

The easiest way to get all of the dependencies is to install one of the pre-packaged scientific python distributes like Enthought’s Canopy or Continuum’s Anaconda. These distributions already contain all of the dependences, and are distributed via 1-click installers for Windows, Mac and Linux.

On python26, two backported packages, argparse and importlib are required. You can install them with pip install argparse and pip install importlib.

Manually Installing the Dependencies¶

$ sudo apt-get install libhdf5-serial-dev python-dev python-numpy python-scipy python-nose python-setuptools cython python-numexpr python-tables python-pandas

$ pip install nose numexpr cython tables pandas

Compiling Dependencies from source (no root needed)¶

If you don’t already have a python installation you want to use, you can compile a new one.

$ wget http://www.python.org/ftp/python/2.7.5/Python-2.7.5.tgz
$ tar -xzvf Python-2.7.5.tgz
$ cd Python-2.7.5
$ ./configure --enable-shared --prefix=$HOME/local/python
$ make
$ make install

$ export PATH=$HOME/local/python/bin:$PATH
$ export LD_LIBRARY_PATH=$HOME/local/python/lib:$LD_LIBRARY_PATH

To compile the dependences from source, you need to get libhdf5 and numpy, which can have some BLAS issues. I recommend configuring everything with --prefix so that you don’t get your packages mixed up with system packages.

$ wget http://www.hdfgroup.org/ftp/HDF5/current/src/hdf5-1.8.11.tar.gz
$ tar -xzvf hdf5-1.8.11.tar.gz
$ cd hdf5-1.8.11
$ ./configure --prefix=$HOME/opt/hdf5-1.8.11
$ make
$ make install

$ export LD_LIBRARY_PATH=$HOME/opt/hdf5-1.8.11/lib:$LD_LIBRARY_PATH
$ export PATH=$HOME/opt/hdf5-1.8.11/bin:$PATH

You’ll probably want to add those export statements to your bashrc too.

If you don’t have easy_install or pip yet, you can get them with

$ wget http://pypi.python.org/packages/source/s/setuptools/setuptools-0.6c11.tar.gz
$ tar -xzvf setuptools-0.6c11.tar.gz
$ cd setuptools-0.6c11.tar.gz
$ python setup.py install
$ easy_install pip

Now you’re home free

$ pip install numpy
$ pip install scipy
$ pip install cython
$ pip install numexpr
$ pip install tables
$ pip install nose