1.2.1. Prerequisites

Note

It is highly suggested to install UV-CDAT and other vacumm dependencies using conda.

1.2.1.1. List of packages

1.2.1.1.1. CDAT

CDAT is a meta-package containing:

  • a version of python;
  • standard and specialized python packages like numpy, scipy and matplotlib;
  • python packages made by UV-CDAT providers, and specialized for manipulation and visualization of ocean/atmosphere data;
  • numerous other libraries and tools.

The packages installed by CDAT and that are needed are the following:

- python
- numpy
- libgfortran
- cdms2
- cdutil
- genutil
- matplotlib
- basemap
- basemap-data-hires
- scipy

1.2.1.1.2. Other packages

Python

When you have installed CDAT or the , here is a list of necessary or useful packages that can be install using pip or easy_install:

Necessary or useful packages
Package Necessary? Description
configobj (4.7.2) Yes Manipulate advanced configuration files
pytz Only for time zone support Work with time zones
paramiko Only for remote Use SSH2 protocol.
seawater Yes for advanced diags. Sea water properties
xlwt Only for xls_style(). Write excel files (better: xlutils)
cmocean Only if you want cmocean colormaps Make these colormaps available like cmocean_thermal
PIL (1.1.7) Only sometimes for add_logo(). Manipulate images
sphinx_fortran (1.0) Doc Extension to sphinx: Documenting fortran codes
sphinxcontrib-cheeseshop (0.2) Doc Extension to sphinx: Linking to Cheese Shop (Python Package Index) packages
sphinxcontrib-programoutput (0.8) Doc Extension to sphinx: Include program output

Documentation generation

These utilities are needed to fully compile the documentation.

Graphviz
Used to create hierarchical diagrams of class inheritance during the generation of the documentation by the sphinx extension sphinx.ext.graphviz. See for instance module core_plot. The program dot may be also needed. You can install both with the pydot package.
dvipng
Used to compile latex formula of the documentation.
enumitem
This package is needed to compile the PDF version of the documentation.

1.2.1.2. Install UV-CDAT

1.2.1.2.1. From conda

Using the conda installer:

$ conda install -c uvcdat uvcdat

1.2.1.2.2. From sources

To install UV-CDAT, follow the official instructions. It can be installed on both linux and mac, by compiling sources or using availables binaries.

Note

If you compile it from sources, you need cmake and git, and it is highly suggested to have your own version of Qt4. Packets are generally availables on all plateforms.

If you use UV-CDAT in operational jobs, you should install a separate version. One way to manage several versions is to use environment modules (see Using environment modules).

1.2.1.3. Setup the environment

Once the installation is done, set environment variables (bash):

$ export PATH=/path/to/uvcdat/bin:$PATH

# For install from sources
$ export LD_LIBRARY_PATH=/path/to/uvcdat/Externals/lib:/path/to/uvcdat/lib:$LD_LIBRARY_PATH
$ export C_INCLUDE_PATH=/path/to/uvcdat/Externals/include:$C_INCLUDE_PATH

1.2.1.4. Check the installation

Then check:

$ python -c "import cdms2"

1.2.1.5. Install other packages

Using conda:

$ conda install -c conda-forge configobj PIL paramiko xlutils seawater pytz cmocean

Using pip:

$ pip install configobj PIL paramiko xlutils seawater pytz
$ pip install sphinxcontrib-cheeseshop sphinxcontrib-programoutput sphinx-fortran