.. _part_2_grib:

Part 2 - GRIB
#############

.. note::

  In Metview GRIB data is represented as a :class:`Fieldset`.

Setup
*****

Navigate into the *2_grib* folder within Metview.

Customising the plot to bring out information
*********************************************

Right-click on the Metview desktop, **Create new icon** and select :ref:`Contouring <mcont_icon>`. Rename the newly created icon to 'shade' and edit it, setting these parameters:

.. list-table::

  * - **Parameter**
    - **Value**

  * - **Legend**
    - **On**

  * - **Contour Shade**
    - **On**

  * - **Contour Shade Method**
    - **Area Fill**

You can also play with the colour settings. 
When done, click OK to close the editor, then visualise the GRIB file and drop your shade icon into the plot window.
There are loads of options in the Contouring icon, which exposes all of the Magics parameters. 
Supplied in the folder are some additional Contouring icons for you to try dropping into the plot window to see what effect they have.

.. image:: /_static/metview_90_minute_introduction_part_2_grib/grid-shading.png
.. image:: /_static/metview_90_minute_introduction_part_2_grib/gridvals.png


.. list-table::

  * - **Icon**
    - **Description**

  * - **shade30**
    - 30 contour levels, no isolines

  * - **grid_shade**
    - no interpolation, just shade the grid boxes 
    
      (looks best with anti-aliasing turned off)

  * - **marker_shade**
    - plots a coloured marker at each grid point location 
    
      (best for categorical data)

  * - **grid_points**
    - plots a small marker at each grid point location 
    
      (all markers are the same, not dependent on value ; 
      
      good for seeing grid structure)

  * - **grid_vals**
    - plots values at each grid point location 
    
      (zoom into a small area before applying this ; good for inspecting
      
      values ; drop the land_sea_shade icon into the plot window too to 
      
      get better orientation)

Have a look at the supplied Python script *plot_max_value_locations.py* and execute it. 
It shows one way to plot the locations of the highest values, by masking off the lower values as invalid. 
Another way would be a customised Contouring icon that only plotted large values.

.. image:: /_static/metview_90_minute_introduction_part_2_grib/max-vals.png

Examine the GRIB metadata
*************************

Right-click on the GRIB file icon and choose **Examine**. 
This brings up the GRIB examiner tool - have a look!

.. image:: /_static/metview_90_minute_introduction_part_2_grib/mv-grib-examiner.png

If you find this tool useful, you can invoke it directly from the command line:

.. code-block::

  metview -e grib <path/to/grib/file>

If you have time...
*******************

Compute and plot a time series from a point
===========================================

Look at the supplied script *plot_time_series.py* as an example. 
It takes a GRIB file containing temperature data on multiple vertical levels for multiple time steps. 
The script plots a time series for each level for a given grid point.

.. image:: /_static/metview_90_minute_introduction_part_2_grib/grib-time-series.png

Convert a GRIB field into text format
=====================================

Create a new :ref:`Grib to Geopoints <grib_to_geo_icon>` icon. Edit it and drop one of the GRIB files into the **Data** box as shown.

.. image:: /_static/metview_90_minute_introduction_part_2_grib/grib-to-gpts.png

Executing this icon will convert the first GRIB field into :class:`Geopoints` format. 
The result can be directly visualised, examined or saved to disk.

Get a Python xarray from a GRIB file
====================================

The script *grib_xarray.py* shows how to get an xarray from a GRIB file. 
See the :ref:`/examples/ens_mean_spread_xarray.ipynb` Jupyter notebook for a more detailed example using ensemble data.