Cube in a Box
The Cube in a Box is a simple way to run the OpenDataCube.
How to use:
If you have make
installed you can use it to save some typing using the instructions a little further down.
All you need to know:
- Set environment variables for
ODC_ACCESS_KEY
andODC_SECRET_KEY
to something valid with your AWS account credentials. - Start a local environment:
docker-compose up
- Set up your local postgres database (after the above has finished) using:
docker-compose exec jupyter datacube -v system init
docker-compose exec jupyter datacube product add /opt/odc/docs/config_samples/dataset_types/ls_usgs.yaml
- Before indexing Landsat 8, you need to grab the pathrows index. Download the file from here and save the zip file to
data/wrs2_descending.zip
- Index a default region with:
docker-compose exec jupyter bash -c "cd /opt/odc/scripts && python3 ./autoIndex.py -p '/opt/odc/data/wrs2_descending.zip' -e '146.30,146.83,-43.54,-43.20'"
- View the Jupyter notebook at http://localhost using the password
secretpassword
- Shutdown your local environment:
docker-compose down
If you have make
:
- Set environment variables for
ODC_ACCESS_KEY
andODC_SECRET_KEY
to something valid with your AWS account credentials. - Start a local environment using
make up
- Set up your local postgres database (after the above has finished) using
make initdb
- Before indexing Landsat 8, you need to grab the pathrows index using
make download-pathrows-file
- Index a default region with
make index
- Edit the Makefile to change the region of interest
- View the Jupyter notebook at http://localhost using the password
secretpassword
Todo:
- Set up notebooks that work on indexed data
Deploying to AWS:
To deploy to AWS, you cam either do it on the command line, with the AWS command line installed or the magic URL below and the AWS console.
Once deployed, if you navigate to the IP of the deployed instance, you can access Jupyter with the password you set in the parameters.json file or in the AWS UI if you used the magic URL.
Magic URL
You need to be logged in to the AWS Console deploy using this URL. Once logged in, click the link, and follow the prompts including settings a bounding box region of interest, EC2 instance type and password for Jupyter.
Command line
- Alter the parameters in the parameters.json file
- Run
make create-infra
- If you want to change the stack, you can do
make update-infra
(although it may be cleaner to delete and re-create the stack)
IMPORTANT NOTES
In your local environment, in order to be able to get data from S3, you need to ensure that the environment variables ODC_ACCESS_KEY
and ODC_SECRET_KEY
are set to something valid.