Handling Visualization types in Geostreams

[aarajh]

Currently, we do not have a dynamic way of adding or managing visualization types in geostreams. Firstly, visualization types are linked with parameter categories, so each category can have only one type of visualization. Secondly, since geostreams now supports nested datapoints, there should be some information stored with each parameter whether it should expect nested datapoints or simple value datapoints. To solve this, one idea is to create the following new tables with the mentioned columns:

• parameter_visualization
  • parameter_id
  • visualization_id
• visualizations
  • id
  • name
  • type (simple or nested)
  • definition

Explanation:

• parameter_visualization table is just to map parameters to their visualizations types.
• visualizations table contains the info regarding the visualization. This table is not entirely necessary for current use case since we can add the type column to parameters table directly. But adding this table can allow us to support custom visualizations in the future where we can add the vega spec to the definition column and use that in the frontend.

[ka7eh]

Is that something that can be handled on the UI side? E.g., put in the project config? It's a good practice to keep the backend agnostic of data representation stuff. It should mainly be responsible for providing raw data and each project can decide how to handle the data.

[ka7eh]

I should have asked this first! Are the mentioned tables for the backend, in the geostreams API? I just realized this discussion is on geodashboard, so maybe I got the idea wrong!