Objective JH-2: Expand “Open in QGIS on JupyterHub” work

[batpad]

Motivation

We currently can launch QGIS (in a Linux desktop environment in your browser) on the JupyterHub, and via URL instruct it to open specific vector layers. This is the start of ‘Open in QGIS’ work, and should be expanded.

Proposal

The expansion is in two ways:

• Learn what kinds of datasets would be most important to open, and add methods to open those kinds of datasets
• Work with the VEDA frontend team to figure out appropriate next steps for linking to specific datasets from the frontend to “Open in QGIS” in the hub

We will also perform additional maintenance and stabilization work on this integration to make sure users have a smooth experience throughout.

Owners

• @batpad (Development Seed)

Success criteria

• [ ] Add ability to open more types of datasets than standard vector formats from a public URL
• [ ] Demo how a VEDA dataset can be opened in QGIS on the hub
• [ ] In collaboration with VEDA UI team, make specific plan for integration into VEDA UI frontend

[batpad]

Spoke to @wildintellect a bit about this and what data / tile formats we should support.

For now, I think we want to support 3 types of datasets:

• Vector data, eg. GeoJSON, fgb - this is currently supported already
• Raster data pointing to a GeoTIFF
• XYZ tile layers

@sunu can you add some pointers of where in the code we should look to add methods to support different types, and then I guess this should be pretty easy to test.

[sunu]

@batpad The code that opens remote vector datasets as a layer is here. Right now, it should support any remote source that can be opened with vsicurl.