datacube-ows 1.9.15

Features

• Leverages the power of the Open Data Cube, including support for COGs on S3.

• Fully supports WMS v1.3.0. Partial support (GetMap requests only) for v1.1.1.

• Supports WMTS 1.0.0

• Supports WCS versions 1.0.0, 2.0.0 and 2.1.0.

• Richly featured styling engine for serving data visualisations via WMS and WMTS.

System Architecture

Known CRS Limitations

1. ODC datasets with WKT-format CRSs will not work with OWS - data from such datasets will never be displayed. OWS currently only works with EPSG format CRSs.

2. Datasets that straddle the anti-meridian or the north or south polar region will cause issues with the legacy postgres driver.

These are fundamental limitation of the way OWS works with the postgres ODC index driver. These limitations are addressed in v1.9.0, but only for the new ODC postgis index driver.

Note for Developers

This repository is currently running the TypeChangeBot as part of an academic research project. Developers making changes to type hints will be prompted to fill out a short survey about the reasons for the annotation changes. Participation is voluntary. More information about the project can be found here

Community

This project welcomes community participation.

Join the ODC Discord if you need help setting up or using this project, or the Open Data Cube more generally.

Please help us to keep the Open Data Cube community open and inclusive by reading and following our Code of Conduct.

Setup

Datacube_ows (and datacube_core itself) has many complex dependencies on particular versions of geospatial libraries. Dependency conflicts are almost unavoidable in environments that also contain other large complex geospatial software packages. We therefore strongly recommend some kind of containerised solution and we supply scripts for building appropriate Docker containers.

Linting

ruff check
isort --check --diff **/*.py

Configuration and Environment

The configuration file format for OWS is fully documented here.

And example configuration file datacube_ows/ows_cfg_example.py is also provided, but may not be as up-to-date as the formal documentation.

Environment variables that directly or indirectly affect the running of OWS are documented here.

Docker-Compose

docker compose up

docker compose up --build

docker compose -f docker-compose.yaml -f docker-compose.db.yaml up

docker compose -f docker-compose.yaml -f docker-compose.override.yaml -f docker-compose.db.yaml up

docker compose -f docker-compose.yaml -f docker-compose.db.yaml up

# Enable pydev for pycharm (needs rebuild to install python libs)
# hot reload is not supported, so we need to set FLASK_DEV to production
export PYDEV_DEBUG=yes
export FLASK_DEV=production
docker compose -f docker-compose.yaml -f docker-compose.db.yaml up --build

cp .env_simple .env # for a single ows config file setup
cp .env_ows_root .env # for multi-file ows config with ows_root_cfg.py
docker compose up

Docker

To run the standard Docker image, create a docker volume containing your ows config files and use something like:

docker build --tag=name_of_built_container .

docker run --rm \
      -e DATACUBE_OWS_CFG=datacube_ows.config.test_cfg.ows_cfg   # Location of config object
      -e AWS_NO_SIGN_REQUEST=yes                                 # Allowing access to AWS S3 buckets
      -e AWS_DEFAULT_REGION=ap-southeast-2 \                     # AWS Default Region (supply even if NOT accessing files on S3! See Issue #151)
      -e SENTRY_DSN=https://key@sentry.local/projid \            # Key for Sentry logging (optional)
      \ # Database connection URL: postgresql://<username>:<password>@<hostname>:<port>/<database>
      -e ODC_DEFAULT_DB_URL=postgresql://myuser:mypassword@172.17.0.1:35434/mydb \
      -e PYTHONPATH=/code                                        # The default PATH is under env, change this to target /code
      -p 8080:8000 \                                             # Publish the gunicorn port (8000) on the Docker
      \                                                          # container at port 8008 on the host machine.
      --mount source=test_cfg,target=/code/datacube_ows/config \ # Mount the docker volume where the config lives
      name_of_built_container

The image is based on the standard ODC container and an external database

Installation with Conda

The following instructions are for installing on a clean Linux system.

• Create and activate a Python 3.10 Conda environment:

  conda create -n ows -c conda-forge python=3.10 datacube pre_commit postgis
  conda activate ows

• Install the latest release using pip install:

  pip install datacube-ows[all]

• Initialise and run PostgreSQL:

  pgdata=$(pwd)/.dbdata
  initdb -D ${pgdata} --auth-host=md5 --encoding=UTF8 --username=ubuntu
  pg_ctl -D ${pgdata} -l "${pgdata}/pg.log" start # if this step fails, check log in ${pgdata}/pg.log
  
  createdb ows -U ubuntu

• Enable the PostGIS extension:

  psql -d ows
  create extension postgis;
  \q

• Initialise the Datacube and OWS schemas:

  export ODC_DEFAULT_DB_URL=postgresql:///ows
  datacube system init
  
  # to create schema, tables and materialised views used by datacube-ows.
  
  export DATACUBE_OWS_CFG=datacube_ows.ows_cfg_example.ows_cfg
  datacube-ows-update ubuntu --schema
  
  # If you are not using the `default` ODC environment, you can specify the environment to create the schema in:
  
  datacube-ows-update -E myenv --write-role ubuntu --schema

• Create a configuration file for your service, and all data products you wish to publish in it. Detailed documentation of the configuration format can be found here.

• Set environment variables as required. Environment variables that directly or indirectly affect the running of OWS are documented here.

• Run datacube-ows-update (in the Datacube virtual environment).

• When additional datasets are added to the datacube, the following steps will need to be run:

  # Update the materialised views (postgis index driver only - can be skipped for the postgis index driver):
  datacube-ows-update --views
  # Update the range tables (both index drivers)
  datacube-ows-update

• If you are accessing data on AWS S3 and running datacube_ows on Ubuntu you may encounter errors with GetMap similar to: Unexpected server error: '/vsis3/bucket/path/image.tif' not recognized as a supported file format.. If this occurs run the following commands:

  mkdir -p /etc/pki/tls/certs
  ln -s /etc/ssl/certs/ca-certificates.crt /etc/pki/tls/certs/ca-bundle.crt

• Launch the flask app using your favorite WSGI server. We recommend using Gunicorn with either Nginx or a load balancer.

The following approaches have also been tested:

Flask Dev Server

• Good for initial dev work and testing. Not (remotely) suitable for production deployments.

• cd to the directory containing this README file.

• Set the FLASK_APP environment variable:

  export FLASK_APP=datacube_ows/ogc.py

• Run the Flask dev server:

  flask run

• If you want the dev server to listen to external requests (i.e. requests from other computers), use the –host option:

  flask run --host=0.0.0.0

Local Postgres database

1. create an empty database and db_user

2. run datacube system init after creating a datacube config file

3. A product added to your datacube datacube product add url some examples are here: https://github.com/GeoscienceAustralia/dea-config/tree/master/products

4. Index datasets into your product for example refer to https://datacube-ows.readthedocs.io/en/latest/usage.html

   aws s3 ls s3://deafrica-data/jaxa/alos_palsar_mosaic/2017/ --recursive \
   | grep yaml | awk '{print $4}' \
   | xargs -n1 -I {} datacube dataset add s3://deafrica-data/{}

5. Write an ows config file to identify the products you want available in ows, see example here: https://github.com/opendatacube/datacube-ows/blob/master/datacube_ows/ows_cfg_example.py

6. Run datacube-ows-update --schema --read-role <db_read_role> --write-role <db_write_role> as a database superuser role to create ows specific tables and views

7. Run datacube-ows-update as db_write_role to populate ows extent tables.

Apache2 mod_wsgi

Getting things working with Apache2 mod_wsgi is not trivial and probably not the best approach in most circumstances, but it may make sense for you.

If you use the pip install approach described above, your OS’s pre-packaged python3 apache2-mod-wsgi package should suffice.

• Activate the wsgi module:

cd /etc/apache2/mods-enabled
ln -s ../mods-available/wsgi.load .
ln -s ../mods-available/wsgi.conf .

• Add the following to your Apache config (inside the appropriate VirtualHost section):

  WSGIDaemonProcess datacube_ows processes=20 threads=1 user=uuu group=ggg maximum-requests=10000
  WSGIScriptAlias /datacube_ows /path/to/source_code/datacube-ows/datacube_ows/wsgi.py
  <Location /datacube_ows>
          WSGIProcessGroup datacube_ows
  </Location>
  <Directory /path/to/source_code/datacube-ows/datacube_ows>
          <Files wsgi.py>
                  AllowOverride None
                  Require all granted
          </Files>
  </Directory>

  Note that uuu and ggg above are the user and group of the owner of the Conda virtual environment.

• Copy datacube_ows/wsgi.py to datacube_odc/local_wsgi.py and edit to suit your system.

• Update the url in the configuration

Credits

This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.