Wow, I'm really impressed by your geohacking skills. Thanks a lot.
Yes, there is a GRASS::v.clean function that cuts lines.
Also, GRASS already includes allocation algorithms.
We could also imagine performing shortest path calculations for extra
vertexes that aren't initally on the graph: shop locations, particular
events, houses...
For this, spatstat::project2segment might be the appropriate function.
Based on your initial code, I think we could write a function that would
combine a linear network with extra vertexes into a single graph. The
vertex labels could discriminate the intersection nodes from the thematic
ones.
2012/8/31 Barry Rowlingson <b.rowling...@lancaster.ac.uk>
> On Fri, Aug 31, 2012 at 11:58 AM, Rowlingson, Barry
> <b.rowling...@lancaster.ac.uk> wrote:
>
> > If I didn't have to be on a train in three hours I'd code this up...
>
> Oh who am I kidding:
>
> buildTopo <- function(lines){
> require(rgeos)
> require(igraph)
>
> g = gIntersection(lines,lines)
> edges = do.call(rbind,lapply(g@lines
> [[1]]@Lines,function(ls){as.vector(t(ls@coords))}))
> lengths = sqrt((edges[,1]-edges[,3])^2+(edges[,2]-edges[,4])^2)
>
> froms = paste(edges[,1],edges[,2])
> tos = paste(edges[,3],edges[,4])
>
> graph = graph.edgelist(cbind(froms,tos),directed=FALSE)
> E(graph)$weight=lengths
> return(graph)
>
> }
>
> > m = readOGR(".","Line")
> > gg = buildTopo(m)
>
> compute shortest distance between two vertices (note the -1 is because
> I'm still on old igraph with 0-indexing):
>
> > E(gg)[get.shortest.paths(gg,V(gg)[1],V(gg)[14])[[1]]-1]
>
> Edge sequence:
>
> [0] 0.81711193 47.57711634 -- -0.28546154 48.4287593
> [1] 2.2151139 46.49728033 -- 0.81711193 47.57711634
> [9] 1.57160852 45.36348513 -- 2.2151139 46.49728033
> [10] 1.39973834 45.06066625 -- 1.57160852 45.36348513
> [11] 1.29755246 44.88062446 -- 1.39973834 45.06066625
> [12] 0.91544563 43.67971729 -- 1.29755246 44.88062446
> [13] 0.88815229 43.43407719 -- 0.91544563 43.67971729
>
> I think this could be improved by adding the coordinates to the nodes.
> But anyway, 90% done.
>
> Barry
>
[[alternative HTML version deleted]]
_______________________________________________
R-sig-Geo mailing list
R-sig-Geo@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
