Skip to content

wekeo/learn-olci

Repository files navigation

Learn OLCI

Binder WEkEO


Python License: MIT

The learn-olci module consists of a collection of python-based Jupyter-notebooks design to demonstrate the capability of the Ocean and Land Colour Instrument (OLCI), carried by the Sentinel-3 satellites, and to help users begin to work with its data at level-1B and level-2.

For any questions about this repository, please contact ops@eumetsat.int.

License

This code is licensed under an MIT license. See file LICENSE.txt for details on the usage and distribution terms. No dependencies are distributed as part of this package. Copyright EUMETSAT 2023.

All product names, logos, and brands are property of their respective owners. All company, product and service names used in this website are for identification purposes only.

Authors

Please see the AUTHORS.txt file for more information on contributors.

Prerequisites

You will require Jupyter Notebook to run this code. We recommend that you install the latest Anaconda Python distribution for your operating system. Anaconda Python distributions include Jupyter Notebook.

Dependencies

item version licence package info
python 3.9.13 PSF https://docs.python.org/3/license.html
xarray 0.21.1 Apache-2.0 https://anaconda.org/conda-forge/xarray
netcdf4 1.5.8 MIT https://anaconda.org/conda-forge/netcdf4
shapely 1.8.0 BSD-3 https://anaconda.org/conda-forge/shapely
matplotlib 3.5.1 PSFL https://matplotlib.org/stable/users/project/license.html
cartopy 0.20.2 LGPL-3 https://scitools.org.uk/cartopy/docs/latest/copyright.html
notebook 6.4.12 BSD-3 https://anaconda.org/conda-forge/notebook
jupyter_contrib_nbextensions 0.5.1 BSD-3 https://anaconda.org/conda-forge/jupyter_contrib_nbextensions
ipywidgets 7.6.5 BSD-3 https://anaconda.org/conda-forge/ipywidgets
scikit-image 0.19.1 BSD-3 https://anaconda.org/conda-forge/scikit-image
plotly 5.6.0 MIT https://anaconda.org/conda-forge/plotly
bokeh 2.4.2 BSD-3 https://anaconda.org/conda-forge/bokeh
hda 0.3.7 Apache-2.0 https://pypi.org/project/hda/
eumartools 0.0.1 MIT https://anaconda.org/cmts/eumartools
ipykernel 6.4.1 BSD-3 https://anaconda.org/conda-forge/ipykernel
cmocean 2.0 MIT https://anaconda.org/conda-forge/cmocean
eumdac 1.0.0 MIT https://anaconda.org/eumetsat/eumdac

Installation

The simplest and best way to install these packages is via Git. Users can clone this repository by running the following commands from either their terminal (on Linux/OSx), or from the Anaconda prompt.

You can usually find your terminal in the start menu of most Linux distributions and in the Applications/Utilities folder on OSx. Alternatively, you should be able to find/open your Anaconda prompt from your start menu (or dock, or via running the Anaconda Navigator). Once you have opened a terminal/prompt, you should navigate to the directory where you want to put the code. Once you are in the correct directory, you should run the following command;

git clone --recurse-submodules --remote-submodules https://gitlab.eumetsat.int/eumetlab/oceans/ocean-training/sensors/learn-olci.git

This will make a local copy of all the relevant files.

Note: If you find that you are missing packages, you should check that you ran git clone with both the --recurse-submodules and --remote-submodules options.

Note: if you are using an older version of git, you may find that your submodules are empty. In this case, you need to remove the folder and re-run the line above with --recursive added to the end

Usage

This collection supports Python 3.9. Although many options are possible, the authors highly recommend that users install the appropriate Anaconda package for their operating system. In order to ensure that you have all the required dependencies, we recommend that you build a suitable Python environment, as discussed below.

Python environments

Python allows users to create specific environments that suit their applications. This tutorials included in this collection require a number of non-standard packages - e.g. those that are not included by default in Anaconda. In this directory, users will find a environment.yaml file which can be used to construct an environment that will install all the required packages.

To construct the environment, you should open either terminal (Linux/OSx) or an Anaconda prompt window and navigate to repository folder you downloaded in the Installation section above. In this folder there is a file called environment.yml. This contains all the information we need to install the relevant packages.

To create the environment, run:

conda env create -f environment.yml

This will create a Python environment called cmts_learn_olci. The environment won't be activated by default. To activate it, run:

conda activate cmts_learn_olci

Now you are ready to go!

Note: remember that you may need to reactivate the environment in every new window instance

Running Jupyter Notebook

This module is based around a series of Jupyter Notebooks. These support high-level interactive learning by allowing us to combine code, text description and data visualisations. If you have not worked with Jupyter Notebooks before, please look at the Introduction to Python and Project Jupyter module to get a short introduction to their usage and benefits.

To to run Jupyter Notebook, open a terminal or Anaconda prompt and make sure you have activated the correct environment. Again, navigate to the repository folder.

If you are running this code for the first time in this environment, you need to enable two extensions to Jupyter by running the following commands.

jupyter nbextension enable --py widgetsnbextension
jupyter nbextension enable exercise2/main

Note: you can also enable these in the Nbextensions tab of the Jupyter browser window

Now you can run Jupyter using:

jupyter notebook or jupyter-notebook, depending on your operating system.

This should open Jupyter Notebooks in a browser window. On occasion, Jupyter may not be able to open a window and will give you a URL to past in your browser. Please do so, if required.

Note: Jupyter Notebook is not able to find modules that are 'above' it in a directory tree, and you will unable to navigate to these. So make sure you run the line above from the correct directory!

Now you can run the notebooks! We recommend you start with the Index module.

Collaborating, contributing and issues

If you would like to collaborate on a part of this code base or contribute to it please contact us on copernicus.training@eumetsat.int. If you are have issues and need help, or you have found something that doesn't work, then please contact us at ops@eumetsat.int. We welcome your feedback!



Overview for advanced users

Installation:

git clone --recurse-submodules --remote-submodules https://gitlab.eumetsat.int/eumetlab/oceans/ocean-training/sensors/learn-olci.git

Create and set environment

conda env create -f environment.yml
conda activate cmts_learn_olci

WEkEO SPECIFIC

ipython kernel install --user --name=cmts_learn_olci

Activate extensions (1st run in environment, only)

jupyter nbextension enable --py widgetsnbextension
jupyter nbextension enable exercise2/main

Run

jupyter notebook or jupyter-notebook