conda-recipes

A selection of recipes for the conda build system (from the anaconda Python distro) that we use in our [lab]. Most of these packages build binaries for external programs or install libraries for external programs, rather than specifically building Python packages. This is why conda is excellent for this task - to help install programs with very long dependency chains. As such, it's largely equal to a cross-platform homebrew rather than just a Python package manger.

Installing conda

To install conda, see https://store.continuum.io/static/img/Anaconda-Quickstart.pdf. You can install miniconda if you want a smaller distribution.

Installing packages

conda is a binary package manger. As such, we have built binary versions of most/all of the packages in this repository for the following platforms:

• OS X (x86_64) [tested on 10.9]
• Linux (x86_64) [tested on CentOS 6.5 and Ubuntu 14.04]

We do not (at present) support Windows.

Package repository

Packages build from these recipes are available from the binstar repository: https://binstar.org/faircloth-lab. To install packages from this repository using conda, edit your .condarc file to look like:

# channel locations. These override conda defaults, i.e., conda will
# search *only* the channels listed here, in the order given. Use "default"
# to automatically include all default channels.

channels:
  - defaults
  - http://conda.binstar.org/faircloth-lab

Then install a given package using:

conda install <package name>

You can search for specific packages using:

conda search <package name>

Package structure

Within your conda environment, the directory structure should look like:

Examples/
Launcher.app
bin/
conda-meta/
docs/
include/
jar/ # contains all java *.jar files
lib/
libexec/
python.app/
share/ # contains additional support files
test/ # contains text code for certain programs (e.g. mafft)

Build requirements

Ordinarily, you do not need to build packages from this repo. We have done this for you and the binary versions are available from https://binstar.org/faircloth-lab (see above). However, if you are interested in how we build each package or want to clone/contribute to these build files, build instructions are below.

OSX

• anaconda
• OSX Command Line Tools
• cmake
• JAVA (1.7)

Linux

• anaconda
• Developer tools (yum groupinstall "Development Tools")
• cmake
• JAVA (1.7)

Building

To build an individual package, use:

conda build <package name>

More info

See http://docs.continuum.io/conda/build.html for information on how to make a recipe, or just look at the at the main conda-recipes repo, particularly the sample recipe.