Weacast

# Forecast data loaders (opens new window)

Forecast model plugins are modules that need to be either integrated in a monolithic Weacast application or deployed as independent micro-service like applications to be interconnected, see the architecture section.

Because the most consuming part of a Weacast application is usually the gathering and processing of forecast model data, the weacast-loader (opens new window) module provides you with a set of download services available as Docker containers out-of-the-box. These services perform the same data processing workflow than the ARPEGE, AROME and GFS plugins, based on Krawler (opens new window). This means that your Weacast application doesn't need to integrate these plugins anymore, the running download services will feed the database as usual in the background.

The actual download services are built from generic Docker images containing required dependencies and functions used to generate a Krawler download job (opens new window) tailored to your forecast model and elements: weacast-gfs (opens new window) and weacast-arpege (opens new window) (which also covers weacast-arome (opens new window) because the interface is similar).

You can use the provided Docker images with default model setups, for each model a different setup generates a different tag, e.g.:

  • weacast/weacast-arpege:europe-latest
  • weacast/weacast-arpege:europe-1.4.0
  • weacast/weacast-arpege:world-latest
  • weacast/weacast-arpege:world-1.4.0

You can also build the Docker images containing the different configurations of your ARPEGE, AROME and GFS services:

# Manually
docker build -t weacast/weacast-arpege-world -f dockerfile.arpege-world .
docker build -t weacast/weacast-gfs-world -f dockerfile.gfs-world .
...
# Using Docker compose file
docker-compose build weacast-arpege-world weacast-gfs-world ...

Then you have to run it on your infrastructure:

# Stop/remove previous instances if any
docker-compose stop weacast-arpege-world weacast-gfs-world ...
docker-compose rm weacast-arpege-world weacast-gfs-world ...
# Launch new ones
docker-compose up -d weacast-arpege-world weacast-gfs-world ...

The loaders can archive all data using the COG (opens new window) file format in a (compatible) AWS S3 bucket if you define the following environment variables:

  • S3_ACCESS_KEY (access key - required)
  • S3_SECRET_ACCESS_KEY (secret access key - required)
  • S3_BUCKET (bucket name - required)
  • S3_ENDPOINT (endpoint URL - optional)

← Probe plugin →