Compiling and Installing in a Linux environment
----------------------------------------------------


.. NOTE:: If you are using a pre-built virtual computer, you can skip this page.

This is the second of four final steps required to get the CODAS
software installed on the computer built using earlier sections of
this documentation.  This page is specific to an Ubuntu environment,
with Debian packages used to install Matplotlib etc (not Anaconda).

This assumes you have (from the previous section)
code in ``/home/adcpcode/programs``


Now it is time to compile the CODAS binaries.  This step also creates CODAS
libraries which are used to make the Python code work.

Using the example above, compile and install ``codas3`` first::

    cd /home/adcpcode/programs
    cd codas3
    ./waf configure  --prefix=/usr/local
    ./waf build
    sudo ./waf install
    cd ..


You can check if everything will be OK by typing::

    codas_prefix


It should come back with ``/usr/local``.  You **MUST** have
``codas_prefix`` on your path at this stage.  It is assumed
that ``/usr/local/bin`` is on your ``$PATH``.  If it is not,
add it now, and in the future.


.. NOTE:: You must compile the ``codas3`` library and binaries before
       trying to compile the Python extension code.

At this stage, you can compile the CODAS Python extension code, in
``pycurrents``, which includes the Python CODAS reader and various
tools we use for Python-based CODAS processing::

    cd pycurrents
    python3 ./runsetup.py --sudo
    cd ..

Last, install ``uhdas`` and ``onship``::

    cd uhdas
    python3 runsetup.py --sudo
    cd ..

    cd onship
    python3 setup.py build
    sudo python3 setup.py install
    cd ..


Now that you have actually compiled and installed CODAS C and Python extension
code, it is time for the next step: getting the
`non-Mercurial components <zip_codas.html>`_.