Osdyn dependent packages installed from yaml osdyn environment
..............................................................

You can also create an environment identical to the one of the conda environments available.

As osdyn regularly checks the latest available versions of the different packages, different environments have been used.

These tested python environments (yaml configuration files) are listed below according to osdyn versions:

- osdyn2019.10.yml
- osdyn2020.01.yml
- osdyn2020.02_nospyder.yml
- osdyn2020.06.yml
- osdyn2020.12.yml
- osdyn2021.02.yml
- osdyn2021.12.yml
- osdyn-latest.yml

available under `requirements` directory.

Use the following command to create an environment from YAML file

.. code-block:: bash

  conda env create --name myenv --file osdyn.yml

Add packages unvailable from conda
++++++++++++++++++++++++++++++++++

A few packages are not available from conda, so install them into osdyn environment from `pip install` command.
.. (but never mix pip and conda installation !)

Load osdyn environment and then add packages from their respective cloned or remote projects.

.. code-block:: bash

  source PATH/Anaconda3-2019.10/etc/profile.d/conda.csh [or sh...]
  conda activate osdyn  [or myname]

.. As long as sphinx_rtd_theme is not in 1.5.1 in anaconda cloud, remove this package from the conda list and add sphinx_rtd_theme through ::

..  pip install sphinx-rtd-theme

eo_harvest library
^^^^^^^^^^^^^^^^^^

For user-friendly downloads from CMEMS and other databases, feel free to add eo_harvest.

The package can be downloaded at `ifremer eo_harvest <https://gitlab.ifremer.fr/naiad/eo_harvest in the environment through>`__ and then add it to osdyn environment ::

  cd path_to_eo_harvest
  python setup.py install



.. geopandas (pour notebook hvplot hors osdyn) USELESS for osdyn I think
.. - geo-py:
..
..  - `download <https://pypi.org/project/geo-py/>`__
..  - tar xvf geo-py-0.4.tar.gz
..  - cd geo-py-0.4
..  - python setup.py install

.. .. warning:: geo-py has been replaced by pyproj

datarmor usage
^^^^^^^^^^^^^^

To speed up access to your dataset of interest (or a subset), the user is encouraged to use intake datarmor catalogues. To do so, install dask-hpconfig package::

  python -m pip install git+https://extranetlogin@gitlab.ifremer.fr/iaocea/dask-hpcconfig.git


osdyn library
^^^^^^^^^^^^^

First, get the package. At the moment, you need to register to the project and then 

.. code-block:: bash

  git clone https://gitlab.ifremer.fr/vg056b7/osdyn.git

Second, compile fortran tools::

  cd osdyn/lib/fortran
  make _interp_.so > & outmake_osdyn202102

Third, install osdyn library into your python environment::

  python -m pip install .

.. You need a **fortran compiler** to install osdyn from sources.

.. Run the installation command::

..    $ python setup.py install --prefix=/usr/local

.. Or run only the build command to compile fortran source command::

..    $ python setup.py build