GeoSPARQL Docs update request

Hi Jena Team,

I just discovered that SPARQL queries using geosparql spatial relations like

`FILTER(geof:sfWithin(?geom1, ?geom2))`

do not hit the spatial index. But property functions like

`?geom geo:sfWithin ?envelope .`,

do utilize the spatial index.

The docs state that the Index is optional for geosparql spatial relations which kind of makes it sound like they would use the index if it is provided.

I think that a disclaimer to this effect would be great to include on the geosparql docs page.

A second point:

There also seems to be a massive difference in performance with the spatial index for rectangular polygons and free form polygons.

Would love to see a write-up of why that is and perhaps a warning or something to explain that this is the case. (if it is!?). Noting that this hasn't been an issue until recently when we have started to use Fuseki against ~ 600GB sized datasets.

I note that there is clearly lots of work going on in this space for Jena 6. So this might all be changing soon anyway.

Thanks

Start by locating the GeoSPARQL docs page and checking how it describes optional spatial indexes; compare the documented FILTER and property-function examples. Confirm the claimed performance difference between rectangular and free-form polygons, then document the index limitation and performance caveat while accounting for possible Jena 6 changes. Done means the page clearly states the behavior and limitations.

Written by the indexing model from the issue text.