Installation of lorax under linux can be done either directly
(via the combination of your distribution's package manager together
with virtualenv and pip) or indirectly via Anaconda Python and packages in the
bioconda channel.  Which path you choose depends on your OS (Anaconda is not
available for BSD) as well as your preferences.  If you are a casual user of
lorax, installation from Anaconda Python will get you up and running more
quickly.  Follow the directions in  INSTALL.anaconda.txt.


Prerequisites for Direct Installs
---------------------------------
Using a direct install gives you the most control, especially in compilation
of binary dependencies.  The more that you plan to use lorax with RAxML, the
more advisable it is to build RAxML from source, because RAxML sees major
benefits from compilation with hardware-specific options such as AVX.
The following prerequisites require compiled binaries that must be on
your PATH:

package   version
--------- ---------
biopython v1.69
HMMer     v3.1beta2
RAxML     8.2.9
redis     3.2.0

Note that the FastTree binary is built by lorax at setup time; this
design choice was dictated by distrust of FastTree being built correctly
(with OpenMP and double-precision calculations enabled) on most platforms,
a problem made worse by the lack of configuration files for FastTree.

Note also that some distributions put the redis-server binary in
/usr/sbin, and you may need to prepend this directory to your path
in order for the redis process to start.

If you use the Gentoo source distribution, you may wish to use the ebuilds in
the ebuilds/ directory are used for installation of lorax under Gentoo.  They
are intended to go in a local overlay under dev-python/.
