Ubuntu Developer Setup#
Base OS: Ubuntu 20.04 LTS
This guide will setup an ODC core development environment and includes:
Anaconda python using conda environments to isolate the odc development environment
installation of required software and useful developer manuals for those libraries
Postgres database installation with a local user configuration
Integration tests to confirm both successful development setup and for ongoing testing
Build configuration for local ODC documentation
GDAL, HDF5, and netCDF4:
sudo apt-get install libgdal-dev libhdf5-serial-dev libnetcdf-dev
Install the latest Postgres version available <https://packages.ubuntu.com/search?keywords=postgresql> for your Ubuntu distribution, eg:
sudo apt-get install postgresql-12 # Optionally, Postgis too sudo apt-get install postgresql-12-postgis-3
Ubuntu’s official repositories usually ship older versions of Postgres. You can alternatively get the most recent version from the official PostgreSQL repository.
Optional packages (useful utilities, docs):
sudo apt-get install libhdf5-doc netcdf-doc libgdal-doc sudo apt-get install hdf5-tools netcdf-bin gdal-bin pgadmin3
Python and packages#
Python 3.8+ is required.
Add conda-forge to package channels:
conda config --append channels conda-forge
Conda environments are recommended for use in isolating your ODC development environment from your system installation and other Python environments.
Install required Python packages and create a conda environment named
conda create --name odc_env python=3.8 datacube
odc_env conda environment:
conda activate odc_env
Find out more about conda environments here.
Install other packages:
conda install jupyter matplotlib scipy pytest-cov hypothesis
Postgres database configuration#
This configuration supports local development using your login name.
If this is a new installation of Postgres on your system it is probably wise to set the postgres user password. As the local “postgres” Linux user, we are allowed to connect and manipulate the server using the psql command.
In a terminal, type:
sudo -u postgres psql postgres
Set a password for the “postgres” database role using the command:
and set the password when prompted. The password text will be hidden from the console for security purposes.
Type Control+D or q to exit the posgreSQL prompt.
By default in Ubuntu, Postgresql is configured to use
ident sameuser authentication for any connections from the same machine which is useful for development. Check out the excellent Postgresql documentation for more information, but essentially this means that if your Ubuntu username is
foo and you add
foo as a Postgresql user then you can connect to a database without requiring a password for many functions.
Since the only user who can connect to a fresh install is the postgres user, here is how to create yourself a database account (which is in this case also a database superuser) with the same name as your login name and then create a password for the user:
sudo -u postgres createuser --superuser $USER sudo -u postgres psql postgres=# \password $USER
Now we can create an
agdcintegration database for testing:
Connecting to your own database to try out some SQL should now be as easy as:
psql -d agdcintegration
Open Data Cube source and development configuration#
Download the latest version of the software from the repository
git clone https://github.com/opendatacube/datacube-core cd datacube-core
We need to specify the database user and password for the ODC integration testing. To do this:
cp integration_tests/agdcintegration.conf ~/.datacube_integration.conf
Then edit the
~/.datacube_integration.conf with a text editor and add the following lines replacing
<foo> with your username and
<foobar> with the database user password you set above (not the postgres one, your
[datacube] db_hostname: localhost db_database: agdcintegration db_username: <foo> db_password: <foobar>
Verify it all works#
Run the integration tests:
cd datacube-core ./check-code.sh integration_tests
Build the documentation:
cd datacube-core/docs pip install -r requirements.txt make html
_build/html/index.html in your browser to view the Documentation.