Download PyCogent.app. This native OSX app comes bundled with all required dependencies and is a download, decompress and go experience! It also implements the novel script form system that controls command line scripts via a form based input mechanism.
One way to install PyCogent is to install the QIIME virtual machine using VirtualBox. The installation instructions can be found here. As QIIME does not yet support the latest version of matplotlib, some PyCogent drawing capabilities are broken.
Please, note that this is the only installation method supported for Windows and that natively Windows does not support gz files properly so to uncompress a gz file in Windows use 7-zip.
For the list of dependencies see the Required software list.
The following assumes you have easy_install on your system (this comes standard with new Mac’s for instance), that you have administrator privileges and that you’re connected to the internet. See below if you don’t have easy_install.
The key steps for the minimal install are:
Download the requirements file.
Install pip
$ sudo easy_install -U pip
Use pip to download, build and install PyCogent plus the numpy dependency.
$ DONT_USE_PYREX=1 sudo pip install -r path/to/cogent-requirements.txt
Note
The DONT_USE_PYREX=1 statement is required if you have Pyrex installed due to a conflict between setuptools and later versions of Pyrex. If you don’t have Pyrex, this will still work.
If the above fails to download PyCogent you can download the tarball to your hard drive and replace the first line of the requirements file with the full path to the tarball, e.g. /Users/my_user_name/Downloads/cogent-1.4.tgz.
Add the following lines to the requirements file
MySQL-python>=1.2.2
SQLAlchemy>=0.5
Note
The MySQL-python module requires that you have MySQL installed.
Just replace the first line of the requirements file with https://pycogent.svn.sourceforge.net/svnroot/pycogent/trunk.
You need to install matplotlib (version 0.99+) to use the drawing code. However, compiling matplotlib can be a challenge. We therefore suggest you obtain a prebuilt binary for your platform from the matplotlib project page rather than modify the requirements file. For OSX, we suggest reading the following instructions on how to compiling matplotlib.
If your system doesn’t have easy_install, you can execute the following:
$ sudo curl http://peak.telecommunity.com/dist/ez_setup.py | python
or, if you are on a linux system that has a package manager, you may only need to do something like:
$ sudo apt-get install python-setuptools
Use the approach to getting easy_install that best suites your system, then follow the (above) instructions for the pip based installation.