that idea, needs a bit more clarification On Sat, Jan 6, 2024 at 3:01 PM Claude Warren <cla...@xenei.com> wrote:
> Perhaps we should think about a Geometry SPARQL extension. > > On Sat, Jan 6, 2024 at 1:51 PM Marco Neumann <marco.neum...@gmail.com> > wrote: > > > It used to be all on Google Code. I believe Google folded that site > > a few years ago. I will have a look through my backups. But to be frank > it > > would be straightforward to implement nowadays and the old code would be > > pre Jena 3. > > > > You might be able to get at some of the pre 2007 stuff by way of > > archive.org > > https://web.archive.org/web/20110710100953/http://geosparql.org/ > > > > Keep in mind I came with a background in 3D queries back in 1996. So the > > Jena 2D spatial stuff was quick for me to implement. The most challenging > > part was to hookup the functions to the RDF query engine back in 2002 > with > > what was available in RDQL (a precursor to SPARQL) in Jena. This part is > > well documented nowadays with what is sometimes dubbed as Jena property > > functions or magic properties in SPARQL. > > > > I am currently working again with a 3D extension to SPARQL but this is > > limited to a custom implementation (and not in line with OGC GeoSPARQL) > in > > a client project. > > > > Let's discuss in more detail, you might be flying in no time. > > > > Marco > > > > > > > > On Sat, Jan 6, 2024 at 12:00 PM Claude Warren <cla...@xenei.com> wrote: > > > > > Marco, > > > > > > Is your earlier code available in a repository somewhere? I'd love to > > get > > > a head start on the functions. > > > > > > Claude > > > > > > On Sat, Jan 6, 2024 at 12:43 PM Marco Neumann <marco.neum...@gmail.com > > > > > wrote: > > > > > > > yes, a lot of the geosparql functions will push you towards geo. > > > > > > > > It might be better to just reuse the existing low level functions. My > > > first > > > > spatial implementation for Jena in 2006 made use of the JTS Topology > > > Suite. > > > > > > > > JTS is still used in Jena today to run some of the low level > evaluation > > > > mentioned above but can also conveniently be used for X,Y data. I > > believe > > > > this might just do the job for your use case. > > > > > > > > Marco > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > . > > > > > > > > On Sat, Jan 6, 2024 at 11:15 AM Claude Warren <cla...@xenei.com> > > wrote: > > > > > > > > > Background: I am building an "autonomous chair side table" for my > > > wife. > > > > > The goal is to have the table carry her tea from the kitchen to > > > whatever > > > > > room she is in. As part of this project I am building a reasoning > > > system > > > > > that builds a map based on various inputs (that I call scanners). > I > > > > store > > > > > all the data in an RDF graph, because all data is a graph and I > need > > to > > > > be > > > > > able to annotate it in ways I have not discovered yet. > > > > > > > > > > So basically I get sensor data that tells me an obstacle appears > at a > > > > > location. That goes into the system initially as a point, however > > > > > processing may group it with other points using a convex hull > > > calculation > > > > > to convert a point cloud to an obstacle. > > > > > > > > > > I need to be able to determine if a path of a specific with > (modeled > > > as a > > > > > line with a rounded-end buffer) can traverse the space between 2 > > points > > > > > without touching the obstacles. > > > > > > > > > > So the functions I have been playing with are: > > > > > > > > > > - Geof.SF_OVERLAPS > > > > > - Geof.SF_INTERSECTS > > > > > - Geof.SF_TOUCHES > > > > > - Geof.DISTANCE_NAME > > > > > - SpatialExtension.NEARBY > > > > > > > > > > I may just revert to writing a couple of filter functions to do > what > > I > > > > > want, but I was trying to learn the existing library. > > > > > > > > > > Claude > > > > > > > > > > On Sat, Jan 6, 2024 at 10:47 AM Marco Neumann < > > marco.neum...@gmail.com > > > > > > > > > wrote: > > > > > > > > > > > In terms of geo you would talk about some sort of equidistant > > > > projection > > > > > > that is available via the CRS/SRS.This is currently implemented > > with > > > > > Apache > > > > > > SIS in Jena. > > > > > > > > > > > > But it sounds like geo isn't really your thing here, so in > > conclusion > > > > > > geosparql isn't either as it's all about geo data. > > > > > > > > > > > > What type of functions (access methods) would you like to perform > > on > > > > the > > > > > > data? > > > > > > > > > > > > > > > > > > > > > > > > On Sat, Jan 6, 2024 at 8:24 AM Claude Warren <cla...@xenei.com> > > > wrote: > > > > > > > > > > > > > Yes just x and y on a flat surface. > > > > > > > > > > > > > > On Sat 6 Jan 2024, 00:45 Marco Neumann, < > marco.neum...@gmail.com > > > > > > > > wrote: > > > > > > > > > > > > > > > Can you give an example? you just want to use x and y instead > > of > > > > > WGS84 > > > > > > > > coordinates? > > > > > > > > > > > > > > > > On Fri, Jan 5, 2024 at 11:28 PM Claude Warren < > > cla...@xenei.com> > > > > > > wrote: > > > > > > > > > > > > > > > > > I want to use the geospatial capabilities of Jena but > without > > > the > > > > > > > global > > > > > > > > > shape. I am looking for a planar geometry (flat, like a > > small > > > > > scale > > > > > > > > map). > > > > > > > > > Is there a way to specify this? > > > > > > > > > > > > > > > > > > Claude > > > > > > > > > > > > > > > > > > -- > > > > > > > > > LinkedIn: http://www.linkedin.com/in/claudewarren > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > > > > > > > > > > > > > > > > > > --- > > > > > > > > Marco Neumann > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > > > > > > > > > > > > > > > --- > > > > > > Marco Neumann > > > > > > > > > > > > > > > > > > > > > -- > > > > > LinkedIn: http://www.linkedin.com/in/claudewarren > > > > > > > > > > > > > > > > > -- > > > > > > > > > > > > --- > > > > Marco Neumann > > > > > > > > > > > > > -- > > > LinkedIn: http://www.linkedin.com/in/claudewarren > > > > > > > > > -- > > > > > > --- > > Marco Neumann > > > > > -- > LinkedIn: http://www.linkedin.com/in/claudewarren > -- --- Marco Neumann