I am trying to use the graph-tool to perform a huge automated series of path analyses as part of a commute-justice project I am working on.  I am very new to graph-tool and somewhat new to Python.

I have converted an ESRI shapefile of street centerlines in Los Angeles County into a GT file using this short script:  https://github.com/CordThomas/shp2gt - i think it creates a valid graph-tool network.    Prior to that conversion I made sure the OpenStreetMap centerlines were split at all intersections.

Running the code below gives me INF as the output to the dist_map returned from the shortest_path request.
What diagnostics might i try to check the traversal of my network?   Running this on the Shapefile used to create the GT, I am able to use QGIS to produce the shortest_path - see screenshot attached.

Screen Shot 2018-11-29 at 1.54.38 PM (2).png

I'd be glad to post my GT file somewhere if others wanted to poke at this

Thank you in advance,

Cord Thomas

Code:

from graph_tool import util as gutil

from graph_tool import topology as gtopo

from graph_tool.all import *


g = Graph(directed=False)

g.load('/ds/data/spatial/Network/streets/tl_2016_06000_roads_la_clipped_extended_split.gt')


vertex_geoid = g.vertex_properties["geoid"]

vertices = gutil.find_vertex(g, vertex_geoid, '060372760005003')

start_vertex = vertices[0]

vertices = gutil.find_vertex(g, vertex_geoid, '060377019023041')

end_vertex = vertices[0]


dist_map = gtopo.shortest_distance(g, source=start_vertex, target=end_vertex, weights=g.edge_properties['weight_dist'])


dist_map


returns inf


output of some data for verification:


>>> start_vertex

<Vertex object with index '7783' at 0x7f91c35693f0>

>>> end_vertex

<Vertex object with index '11265' at 0x7f91c35696f0>

>>> g.num_edges()

560517

>>> g.num_vertices()

549021