Plan to incorporate the results in the new portal

[ocefpaf]

@kwilcox and @vembus now that the new SECOORA portal is online I should work to get the skill results displayed in there too.

@kwilcox I am not sure what would be the best approach, but I am sure that the current notebooks are no good for a production site. Here is my naive plan and I would love to get your feedback:

• Convert the notebooks to scripts
• Put those scripts in a docker image that someone can run periodically to obtain the paired series (Model-Observations)
• Save the series and skill in a digestible format for the site. JSON?

[kwilcox]

SECOORA is in the process of putting together "external software contribution requirements" for this exact purpose... being able to easily integrate applications and products that were developed outside of the core DMAC team.

While the requirements are not complete, it boils down to:

1. A self-contained docker or docker-compose based application that SECOORA can run on their own infrastructure.
2. Documentation on how SECOORA can maintain and update the libraries used in the application (patching security fixes, etc.)
3. Well-defined HTTP endpoints that SECOORA can use to determine if the application is in a healthy state: "is this application working correctly?"
4. If the application should integrate with the SECOORA DMAC systems (catalog, search, portal, etc.), the application will need to produce SECOORA compliant output such as CF NetCDF files, GIS (shapefile, spatialite, geojson, etc.) files, or provide a service endpoint to retrieve data in an agreed upon format.

I don't know the full details of the series and skill output. Can it be represented in a CF-NetCDF file? If not we will have to agree upon a custom JSON format that the DMAC system can ingest regularly.

[ocefpaf]

Can it be represented in a CF-NetCDF file?

The right answer is probably yes. The skill per-se is not a quantity that exists in the CF document, but the data processed for the calculation already are CF-1.5 (time-series, 2D iso-surfaces, and trajectories).

If not we will have to agree upon a custom JSON format that the DMAC system can ingest regularly.

I can provide both the data (CF-netCDF) and the skill in a light weight JSON.

@kwilcox Do you think we can setup a call to talk about the details?

PS: Meanwhile I will work on a docker image for this.

[ocefpaf]

This is stuck until we solve #242 and merge #243.