[Routing] Updated pyroute - loads data as required, can now do long-distance routes

[OJ W]

I'm just wondering if it could be dodgy data on the tile data server...

the original import of uk-planet.osm was done from late May data.
This weekend I imported the whole planet.osm, but using the old node
data from May, which means that lots of roads are missing, and the
data isn't really usable.  I'm starting another import now, of latest
planet.osm data, which should be better (shout if anyone wants
notification when it's finished).

If you're interested in that specific road, look in the cache
directory for the relevant tile ('view image' on a slippy map to see
which one it is), and see if the road exists, if all the nodes exist,
and what it's tagged as.

Another way of viewing those data tiles is using the latest pyrender
(so new it's not even in SVN), which uses the same tile data server:

http://dev.openstreetmap.org/~ojw/files/dev/pyrender_layers/

By default pyrender has its own cache directory for data tiles, but
you should be able to hack it to share the same cache directory as
your pyroutelib2 installation (see line: directory = 'cache in
tiledata.py).  Run server.py and browse the map on
http://localhost:1280/




On Sun, Jun 15, 2008 at 7:08 PM, Mark Williams
<mark.666 at blueyonder.co.uk> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> OJ W wrote:
>> I'm just testing an updated version of pyroute (proper announcement
>> later), but it looks like it's now capable of routing across hundreds
>> of miles
>>
>> software is here:
>>
>> http://svn.openstreetmap.org/applications/routing/pyroutelib2/
>>
>> it relies on a tile data server
>>
>> http://wiki.openstreetmap.org/index.php/Tile_data_server
>>
>> which currently has UK data only.
>>
>> To run it, use a command-line like:
>>
>> python routeAsGpx.py 52.03818 -0.76171 50.90855 -1.41319 car > output.gpx
>>
>
> I haven't had a lot of joy with this - I appreciate it's only in
> development, so hopefully you've already fixed it, but :
>
>
> # The following works:
> python routeAsGpx.py 51.4861044 0.3355225 51.4926519 0.3240215 car
> # This doesn't:
> python routeAsGpx.py 51.4861044 0.3355225 51.4929613 0.3237939 car
> python routeAsGpx.py 51.4926519 0.3240215 51.4929613 0.3237939 car
>
> These basically cover 1 mile or so of fairly simple roads.
> For some reason it barfs at a section of road (the final line above)
> with a footway left from point 1, and a service road off to left of
> point 2, connected by an unbroken (1 segment!) primary road.
>
> There are issues; some twit turned the service road into an un-named
> residential at some point, & I've just fixed it again. The footway
> doesn't have foot=yes on it (but does now).
> I don't see why these should break routing past them? The major road is
> a long unbroken way which the first example copes with nicely.
>
> Mark
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFIVVqpJfMmcSPNh94RAhdeAJ4xkEE2bb2GyueKb4rUovLnUgIo8gCdEq+H
> T+G1PpJmwityRP5GQDDb030=
> =sLYd
> -----END PGP SIGNATURE-----
>
>
> _______________________________________________
> Routing mailing list
> Routing at openstreetmap.org
> http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/routing
>