Multi-WMS limitations/assumptions

[jkburges]

Dumping ground for @jkburges' and @dnahodil's notes about limitations and assumptions of the multi-wms implementation. These may need to be addressed in the future.

• a group of related ncwms layers must all have the same time samples

[dnahodil]

It is currently assumed that all the layers are of the same type (ie. GeoServer vs ncWMS)

[dnahodil]

It is assumed that GeoServer layer filters are configured such that they will work for any map or data layer which is part of the same DataCollection

[dnahodil]

We should try this with multiple GeoServer layers. I've only been testing with the one example which is multiple ncWMS.

[dnahodil]

Decision: the style panel range controls (changes the max/min for the colour scale) and style picker will be per layer. As the cited example case of many ncWMS layers effectively means switching layers is switching variables of that dataset, then having a number range stay between those different attributes does not make sense. E.g. the range for a temperature is going to be very different from a range for salinity. Similarly, the palette you'd use to visualise a temperature would be very different to the one you'd use to visualise temperature anomaly data.

[dnahodil]

We've gone back to using layer names (in the layer picker) which is something we've managed to avoid in the Portal for a while now. Layer names are cryptic and unhelpful for people who are not ocean scientists. It would be nice if we could have something human-readable, instead.

[dnahodil]

It is assumed that layers cover roughly the same geographical area. This is because the geographic bounds are stored in the metadata record which is shared for all layers. If this isn't the case then the 'zoom to layer' button will give odd/incorrect results.

[pblain]

In backlog