[Feat] wxGUI/datacatalog: Adding space time datasets (STDS)

[veroandreo]

It would be very nice to also have the STDS (raster, vector and raster3d) listed in the data catalogue. Moreover, all maps registered in STDS could be listed within them, as to avoid scrolling over a looong list of maps, especially when they are registered within time series.

t.listwith different types could be used for listing the STDS and t.rast.list, t.vect.list and t.rast3d.list could be used to list the maps within them.

We should somehow check for maps within time series to avoid listing them outside the STDS as well. If the maps are then unregistered, they should be listed as regular maps again.

[mlennert]

I'm not a big user of time series, but can we assume that maps in time series are generally not used as individual maps outside these series ? Because if they are, it might be a bit surprising to the user that the map "disappears" just because she added it to a time series.

But I leave that decision to those who have more experience with time series.

[ninsbl]

It seems the import of any temporal module takes ~3 seconds. So, if we have to run t.list for every mapset that can be quite slow.

I assume it will be a bigger / more invasive job to do anything about it (too much for a GSoC project?)...

As of now, repeatedly calling t.list should thus be probably avoided. That said, t.list can return almost all metadata of STDS that t.info can.
So, one approach can be to create a temporary mapset, add all other mapsets and just run one t.list instance to collect STDS per project...
An alternative can be to use the TGIS API... For the GUI it may be an advantage to avoid depending on importing ctypes though!?!
A third alternative could be query the TGIS DB without a full import grass.temporal as tgis

[echoix]

It seems the import of any temporal module takes ~3 seconds. So, if we have to run t.list for every mapset that can be quite slow.

I observed that last year or so too. I tried profiling: the ctype file import is quite long and constant overhead. And then, when stepping through the whole initialisation, it was going everywhere for just about anything , even for the simplest of functions.