search

opengeospatial / ogcapi-discrete-global-grid-systems

https://ogcapi.ogc.org/dggs
Other
20 stars 8 forks source link

Questions re. DGGS Set and Query functions #35

Open recalcitrantsupplant opened 2 years ago

recalcitrantsupplant commented 2 years ago

A short set of questions, happy to split these out as necessary depending on the answers:

DGGS Set Functions:

  1. Sibling - Is this the same as neighbours? I.e. the ZonalIDs immediately arround a given ZonalID, or is it children of the ZonalID’s parent? As a follow on, if they're different, would it be worth having both?
  2. Parent - parent returns all parents currently, would it be useful to have a split between the immediate parent and others, e.g. parent vs. ancestors (all parents)

DGGS Query Functions:

  1. Is ‘overlaps’ valid for two ZonaIDs? (I was initially thinking for rHEALPix, I realise now it is probably valid for H3. Would still be good to confirm.)
  2. withinDistance - is there a useful alternative for this where the distance is measured in ZonalIDs at a specified resolution?
  3. Crosses - is this valid concept for (single) ZonalIDs?
rggibb commented 2 years ago

@recalcitrantsupplant .. Topic 21 Figure 15 gives examples of the use of parent, child and sibling and the use of the levels and inheritID parameters to constrain or extend the result set.

1, siblings are the other children of a zone's parent(s), not the same as neighbours. We could certainly discuss having neighbour or ring .. effectively these are similar to buffer with an appropriately selected distance. But that doesnt prevent us considering them.

  1. Parent by default should only return the parent(s) one level above the source zone, to get all parents you would have to specify parent with a levels value equal or greater than the level of the source zone.
  2. overlaps is certainly valid in general, though you are correct that in rHEALPix it will always return False. That doesn't make it an invalid function.
  3. withinDistance - we would have to specify the level as the unit as well as the number of zones to traverse, but lets look at it.
  4. Crosses, I cant think of a DGGS zone topology for single zones that would return true for crosses, but once again I dont see that as a reason to exclude it as a test.