-
Notifications
You must be signed in to change notification settings - Fork 25
A Hitchhiker's Guide to Installing candis on Mac OS X
From Zero to Hero - Achilles Rasquinha
To the best of my knowledge, this will be the only page you'd ever require to install candis right from scratch. This script also believes to be a foolproof setup to have candis onto your system. In case if you run into some problems, feel free to raise an issue here.
NOTE: To see if your Mac OS X version has been tested for the following script, check here. You're free to revise this page in order to add your Mac OS X version if you're successful with this script too (psst, helps others).
First, open your Terminal (Finder > Go (Menu Bar) > Utilities > Terminal)
$ curl -sL git.io/install-candis | python
- Install Xcode command-line tools
$ xcode-select --install
When you execute the above command, a pop-up window will appear. Click on Install and then accept the terms and conditions (by clicking on Agree). Once downloaded and installed, click on Done.
- Install Homebrew - Mac OS X's package manager
(Requires Ruby installed. To our luck, Ruby comes packaged with most Mac OS X systems)
On OS X El Capitan, Yosemite, Mavericks, and macOS Sierra, Ruby 2.0 is included. OS X Mountain Lion, Lion, and Snow Leopard ship with Ruby 1.8.7 - Ruby Lang Documentation
(Type ruby --version
to check whether Ruby is available on your system)
$ /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
- Check whether Homebrew has been installed correctly
$ brew doctor
You should then see - Your system is ready to brew! on your terminal screen
- Update Homebrew and its formulae
$ brew update
- Install git (Using Homebrew)
$ brew install git
- ...and check if git has been correctly installed.
$ git --version
You should then see the version information only.
- Clone candis (Optional - Add the
--recursive
flag to clone candis-res, a repository with sample genomic data) using git.
$ git clone --recursive https://github.com/HelikarLab/candis.git
This just goes ahead and pulls the source files from the remote git repository.
- Ensure you're pointing to candis's root directory
$ cd candis
Add candis to your $PYTHONPATH
Environment Variable as follows:
$ # Add Environment Variable to ~/.bash_profile
$ echo "export PYTHONPATH=path/to/candis/candis:$PYTHONPATH" >> ~/.bash_profile
$ # Refresh ~/.bash_profile
$ source ~/.bash_profile
where path/to/candis
is the absolute path to your candis source directory. Note that $PYTHONPATH
points to candis's Python source directory (path/to/candis/candis
) and not its root source directory (path/to/candis
).
candis's API (CLI as well as the RIA server) is built using Python (3.6+); while from Data Manipulation to Model Deployment uses Python wrappers for R and WEKA (which requires Java). In addition, candis requires graphviz installed (to generate Graph Visualisations for models like Decision Trees, etc.).
- Install Python 3 (Using Homebrew)
(Comes withpip
- Python's Package Manager installed)
$ brew install python3
- Install R (Using Homebrew)
$ brew tap homebrew/science # Everything Science
$ brew tap caskroom/cask # Homebrew's "drag-and-drop" alternative
$ brew cask install xquartz # X Window System (Using Homebrew Cask)
$ brew install r # Install R
- Check if R is correctly installed
$ R --version
- Install WEKA (Using Homebrew Cask)
$ brew cask install weka
Homebrew Cask automatically handles the required dependencies for you.
- Install Graphviz (Using Homebrew)
$ brew install graphviz
- Run the
R/setup.R
script within the candis source directory as follows:
$ cd R && Rscript setup.R && cd ..
We recommend you to install candis's Python Dependencies within a virtualenv
(A virtualenv
goes ahead and creates an isolated Python environment for you).
- Install and Create a
virtualenv
named venv
$ pip3 install virtualenv # Installs a Python Virtual Environment
$ virtualenv venv -p python3 # Creates a Python Virtual Environment named `venv`
$ source venv/bin/activate # Goes ahead and activates the Virtual Environment for you.
ISSUE: candis uses python-weka-wrapper (which then uses the python-javabridge
library for running and accessing the Java Virtual Machine) to utilize everything (almost) WEKA has to offer. A quick warning is that candis uses a bleeding-edge version of python-javabridge
(Check out python-javabridge
's Issue #111).
- First, install
numpy
(a dependency ofpython-javabridge
)
(venv) $ pip install numpy
- Now, simply install candis's Python Dependencies
(venv) $ pip install -r requirements.txt
(venv) $ pip install -r requirements-dev.txt
- Finally, candis uses
matplotlib
to render neat graphs. Hence, changematplotlib
's backend toTkAgg
$ echo "backend: TkAgg" >> ~/.matplotlib/matplotlibrc
candis comes with out-of-the-box precompiled JavaScript and CSS files. However if you're interested in building one yourself, you'd then require Node.js and SASS as additional dependencies.
- Install Node.js (Using Homebrew)
(Comes withnpm
- Node's Package Manager installed).
$ brew install node
- Install SASS (as a Ruby Gem)
$ gem install sass
candis comes with a candis-res directory that provides you some sample genomic data to start with. Simply:
$ mkdir CRES && cp candis-res/CRES/* CRES
Run candis as a module
(venv) $ python -m candis
Open your favourite browser, type localhost:5000 on the address bar and press Enter (Return). You should then see something like this.
NOTE: If you're attempting to revise this page after successfully installing and running candis, kindly add the required details in the following format only.
Name | Version | By |
---|---|---|
Mac Sierra | 10.12.5 | Achilles Rasquinha @achillesrasquinha, [email protected] |