.. _quick_start:


Quick Start
===========


.. image:: https://travis-ci.org/MPAS-Dev/geometric_features.svg?branch=master
   :target: https://travis-ci.org/MPAS-Dev/geometric_features
   :alt: Build Status


This repository houses definitions of geometric features. These features
can include regions, transects, and points, described in geojson format.
For example, here are some regions for Antarctica.

.. image:: https://cloud.githubusercontent.com/assets/4179064/12921663/93282b64-cf4e-11e5-9260-a78dfadc4459.png
   :target: https://cloud.githubusercontent.com/assets/4179064/12921663/93282b64-cf4e-11e5-9260-a78dfadc4459.png
   :alt: alt text


The python ``geometric_features`` package can be used to help maintain and use
this repository. Several example scripts that make use of the package can be
found in the ``examples`` directory.  Each of the classes and functions that make
up the package have extensive documentation.  More user-level documentation
will be added shortly.

To use geometric features, you can install it in a conda environment:

.. code-block:: bash

   conda create -n geometric_features -c conda-forge python=3.7 geometric_features
   conda activate geometric_features

By default, ``geometric_features`` will download the data it needs on the fly.
Since some of the features are quite large, this can be convenient if disk
space is at a premium.

Some systems do not support downloading the data (e.g. because of firewalls or
compute nodes that don't have internet access.  In such cases, it is convenient
to install ``geometric_features`` including all the data:

.. code-block:: bash

   conda create -n geometric_features -c conda-forge python=3.7 \
       "geometric_features=*=*with_data*"
   conda activate geometric_features

This syntax is admittedly a bit clunky but it selects for a special build of
the conda package with the data included.

To develop ``geometric_features`` (e.g. to add new features), it is highly
recommended that you use an ``anaconda`` python environment.  Here is how to
create and activate an environment with all of the required dependencies:

.. code-block:: bash

   conda create -n geometric_features -c conda-forge python=3.7 numpy matplotlib \
       cartopy shapely requests progressbar2
   conda activate geometric_features

A typical workflow will look like:


* Create a ``GeometricFeatures`` object and point it to a location where you have
  stored (or would like to store) geometry data.

  * ``gf = GeometricFeatures(localCache='./geometric_data')``

* Read in one or more ``FeatureCollection``\ s from the ``geojson`` files in the
  ``geometric_data`` directory.

  * ``fcArctic = gf.read('ocean', 'region', featureNames=['Arctic Ocean'])``
  * ``fcAtlantic = gf.read('ocean', 'region', tags=['Atlantic_Basin'])``

* Edit features:

  * Merge, combine, tag, mask out or simplify the features, see below.
  * Use the ``shapely`` package to edit the geometry in more sophisticated ways

* Visualize features:

  * ``fc.plot(projection='cyl')``

* Split feature collection back into individual features for inclusion in the
  repo:

  * ``gf.split(fc)``

Available functionality includes:


* ``fc.merge(other)`` - Merge two feature collection together.
* ``fc.combine()`` - Combine features into a single feature.
* ``fc.difference()`` - Mask features using shapes in a second feature file.
* ``fc.fix_antimeridian()`` - Split a feature at the antimeridian (+/- 180 longitude). The resulting feature has all points between -180 and 180 lon.
* ``fc.set_group_name()`` - Set the "groupName" property of the ``FeatureCollection``
* ``fc.tag()`` - Add one or more tags to the "tag" property of each feature in a collection.  This can be useful for reading back a collection of features with that tag.

**IMPORTANT:** Always use the ``gf.split(fc)`` script when placing features into
the ``geometric_data`` directory. This will help everyone maintain the
repository, and allow tools to parse them cleanly.

Many of this functionality can also be accessed with a command-line interface:

.. code-block:: bash

   merge_features
   combine_features
   difference_features
   set_group_name
   split_features
   simplify_features
   tag_features
   plot_features

Use the ``-h`` flag to find out more.

An example workflow to read in, plot and write out a set of features is

.. code-block:: python

   #!/usr/bin/env python

   from geometric_features import GeometricFeatures
   import matplotlib.pyplot as plt

   # create a GeometricFeatures object that points to a local cache of geometric
   # data and knows which branch of geometric_feature to use to download
   # missing data
   gf = GeometricFeatures(cacheLocation='./geometric_data')

   # read in a FeatureCollection containing all ocean regions in the Atlantic
   # basin
   fcAtlantic = gf.read(componentName='ocean', objectType='region',
                        tags=['Atlantic_Basin'])

   fcAtlantic.plot('cyl')
   plt.title('Atlantic merged')

   # combine them all into a single feature
   fcAtlantic = fcAtlantic.combine(featureName='Atlantic_Basin')
   fcAtlantic.plot('cyl')
   plt.title('Atlantic combined')

   # make another feature containing the regions in Filchner-Ronne Ice Shelf
   fcFilchnerRonne = gf.read(componentName='iceshelves', objectType='region',
                             featureNames=['Filchner_1', 'Filchner_2',
                                           'Filchner_3', 'Ronne_1', 'Ronne_2'])
   fcFilchnerRonne.plot('southpole')
   plt.title('Filchner-Ronne')


   # make one more collection of all the IMBIE basins in West Antarctica
   fcWestAntarctica = gf.read(componentName='landice', objectType='region',
                              tags=['WestAntarcticaIMBIE'])

   fcWestAntarctica.plot('southpole')
   plt.title('West Antarctica')

   fcWestAntarctica.to_geojson('west_antarctica.geojson')
   plt.show()