The following steps have been tested on a clean install of Mavericks 10.9.4.
All others must fend for themselves to install matplotlib, scipy and their third-party dependencies.
To install, first install the Anaconda Python Distribution, which comes pre-packaged with a bunch of the scientific packages we use all the time, pre-installed.
We recommend creating a “sandbox” where you can install any and all packages without disturbing the rest of the Python distribution. This way, your flotilla environment is completely insulated from everything else, just in case something goes wrong with You can do this with
conda create --yes flotilla_env pip numpy scipy cython matplotlib nose six scikit-learn ipython networkx pandas tornado statsmodels setuptools pytest pyzmq jinja2 pyyaml
You’ve now just created a “virtual environment” called flotilla_env (the first argument). Now activate that environment with,
source activate flotilla_env
Now at the beginning of your terminal prompt, you should see:
(flotilla_env)
Which indicates that you are now in the flotilla_env virtual environment. Now that you’re in the environment, follow along with the non-sandbox installation instructions.
To make sure you have the latest packages, on the command line in your terminal, enter this command:
conda update --yes pip numpy scipy cython matplotlib nose six scikit-learn ipython networkx pandas tornado statsmodels setuptools pytest pyzmq jinja2 pyyaml
Not all packages are available using conda, so we’ll install the rest using pip, which is a Python package installer and installs from PyPI, the Python Package Index.
pip install seaborn fastcluster brewer2mpl pymongo gspread husl semantic_version joblib
Next, to install the latest release of flotilla, do
pip install flotilla
If you want the bleeding-edge master version because you’re a developer or a super-user, (we work really hard to make sure it’s always working but could be buggy!), then install the git master with,
pip install git+git://github.com/yeolab/flotilla.git
This part only needs to be done once
ruby -e "$(curl -fsSL https://raw.github.com/Homebrew/homebrew/go/install)"
brew install freetype
conda install --yes pip numpy scipy cython matplotlib nose six scikit-learn ipython networkx pandas tornado statsmodels setuptools pytest pyzmq jinja2 pyyaml`
conda create --yes -n flotilla_env pip numpy scipy cython matplotlib nose six scikit-learn ipython networkx pandas tornado statsmodels setuptools pytest pyzmq jinja2 pyyaml
source activate flotilla_env
pip install git+https://github.com/YeoLab/flotilla.git
mkdir ~/flotilla_scratch
mkdir ~/flotilla_projects
source deactivate
source activate flotilla_env
ipython notebook --notebook-dir=~/flotilla_scratch
Docker is the preferred method to obtain the most up-to-date version of flotilla. Every change we make to the source code triggers a new build of a virtual machine that contains flotilla and all its dependencies.
Here are instructions to get an active docker image. These instructions have not been tested on Windows or Linux.
Note: On Mac OS X and Windows you will need to start docker through the “boot2docker” application before you can use docker.
Install docker ( ≥ version 1.3) according to the instructions appropriate for your system.
Then start flotilla on the command line (OS X Terminal application):
curl https://raw.githubusercontent.com/YeoLab/flotilla/master/docker/start_docker.py | python
After the ipython notebook interface opens, test the installation with our test dataset by running the following commands in a new notebook:
import flotilla study = flotilla.embark(“http://sauron.ucsd.edu/flotilla_projects/neural_diff_chr22/datapackage.json”) study.interactive_pca()