Features 🎉 :

New dataset: the Facebook High Resolution Population Density dataset!
Check out this screenshot of the data over Paris where you can clearly see the different "arrondisements" as the population density changes from one neighbourhood to the next:

Improvements 🚀 :

Updated handling of the Nightlight-hd dataset in accordance with Blackmarble's new data format. The source COGs will now contain raw data values from 0 to 255, and the tiler applies the inferno colour map when the tiles are requested.
Added of a localstack container when running the API locally. This container, upon startup, runs the dataset metadata generator lambda and stores the resulting dataset-metadata.json file in an S3 bucket running locally (in localstack). This makes is possible "test" datasets locally, without deploying them to production. To do this, upload the dataset being tested to a "staging" folder in s3://covid-eo-data/, create/update the metadata file in covid_api/db/static/datasets/ with the "staging" location of this dataset and run the API + dashboard locally. The local instance of the dashboard will contain the staged dataset, without it being displayed in production!
Removed boto* packages from the Lambda docker image build process. The boto* packages are available to the lambda by default at runtime. Omitting them from the lambda container image reduces the deployment package size ensuring the it stays below the size limit.

Included water_pzd in the exclusive_with field of several datasets
Fixed a bug that was causing the dataset metadata generator lambda to omit datasets with no time dimension, rather than include them with an empty domain field