logo separator

[mkgmap-dev] overlapping highways and route restrictions

From Gerd Petermann gpetermann_muenchen at hotmail.com on Wed Apr 4 14:55:38 BST 2018

Hi Felix,

Maybe we have different ideas what merging of labels means.
With r4149 I've implemented this:
If way w1 has e.g. mkgmap:label:1 = "Main Street" and mkgmap:label:2 = "N 232" (a ref)
and way w2 has only mkgmap:label:1 = "Main Street" mkgmap adds mkgmap:label:2  from the first way.
Also if w2 has e.g. mkgmap:label:1 = "Main Place" the merged result would be
mkgmap:label:1 = "Main Street"
mkgmap:label:2 = "N 232"
mkgmap:label:3 = "Main Place"
I did not yet test if that really works with address search, also there is no method to detect which way is w1 and which is w2.
Probably order of appearance.

Another option would be to somehow concatenate labels, esp. those from ref tags, but I have no idea how to code that
as every style might use differents ways to combine refs.

Gerd

________________________________________
Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk> im Auftrag von Felix Hartmann <extremecarver at gmail.com>
Gesendet: Mittwoch, 4. April 2018 12:46:44
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] overlapping highways and route restrictions

okay that's fine. I think the result will lead to nearly no more actual problem cases...

On 4 April 2018 at 12:44, Gerd Petermann <gpetermann_muenchen at hotmail.com<mailto:gpetermann_muenchen at hotmail.com>> wrote:
Hi Felix,

you may try r4147. I plan to add code to merge different labels of overlapping ways, I think I can't do much more.

Gerd

________________________________________
Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk<mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk>> im Auftrag von Felix Hartmann <extremecarver at gmail.com<mailto:extremecarver at gmail.com>>
Gesendet: Mittwoch, 4. April 2018 11:44:05
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] overlapping highways and route restrictions

Yes I think that could be problematic. I'm not sure how it will end up. I do not create routable lines from the relations file directly, I use the set command and then in the lines file the additional ways are created - so that if one line has 1 or X route relations, there will be no difference. However that assumes they are all added to the same line I guess. So if different routes are copied onto different overlapping ways this principle would go wrong - it won't be a problem if the underlying overlaying ways are moved/merged before the handling of the relations however.

I have so far seen mostly cases where two overlaying ways were both part of the same route relation - not of different route relations.

On 4 April 2018 at 11:10, Gerd Petermann <gpetermann_muenchen at hotmail.com<mailto:gpetermann_muenchen at hotmail.com><mailto:gpetermann_muenchen at hotmail.com<mailto:gpetermann_muenchen at hotmail.com>>> wrote:
Hi Felix,

well, you started to mention route relations:
"The case of overlapping ways being part of different route relations on the other hand I've seen quite often - in that case I think it's best to just add all route relations to one way, and remove the other one."
My understanding is that the code in mkgmap doesn't have to care about route relations (e.g. type=route, route=bicycle), this is done in the style. With your style those might cause more trouble because you add more routable lines for the members but the current overlap remover will only remove segments with more or less identical attributes.
What you suggest requires a completely different approach, right?

Gerd

________________________________________
Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk<mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk><mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk<mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk>>> im Auftrag von Felix Hartmann <extremecarver at gmail.com<mailto:extremecarver at gmail.com><mailto:extremecarver at gmail.com<mailto:extremecarver at gmail.com>>>
Gesendet: Mittwoch, 4. April 2018 10:58:43
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] overlapping highways and route restrictions

yes I know - but even though both ways you linked have different turn restriction ID - the content of the turn restriction is identical (only straight on).

On 4 April 2018 at 10:54, Gerd Petermann <gpetermann_muenchen at hotmail.com<mailto:gpetermann_muenchen at hotmail.com><mailto:gpetermann_muenchen at hotmail.com<mailto:gpetermann_muenchen at hotmail.com>><mailto:gpetermann_muenchen at hotmail.com<mailto:gpetermann_muenchen at hotmail.com><mailto:gpetermann_muenchen at hotmail.com<mailto:gpetermann_muenchen at hotmail.com>>>> wrote:
Hi Felix,

yes, 1) would handle the case in my example.
Just to make sure: I meant turn restrictions, not route relations.

Gerd

________________________________________
Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk<mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk><mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk<mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk>><mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk<mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk><mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk<mailto:mkgmap-dev-bounces at lists.mkgmap.org.uk>>>> im Auftrag von Felix Hartmann <extremecarver at gmail.com<mailto:extremecarver at gmail.com><mailto:extremecarver at gmail.com<mailto:extremecarver at gmail.com>><mailto:extremecarver at gmail.com<mailto:extremecarver at gmail.com><mailto:extremecarver at gmail.com<mailto:extremecarver at gmail.com>>>>
Gesendet: Mittwoch, 4. April 2018 10:48:20
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] overlapping highways and route restrictions

For my maps 1) is enough - and warning not even needed. I think it's hard to find out what the outcome should be.
The case of overlapping ways being part of different route relations on the other hand I've seen quite often - in that case I think it's best to just add all route relations to one way, and remove the other one.

The example you found here - I think is quite possible to solve - both restriction relations are identical - so one way including the restriction relation can be removed, the route relations copied over. If the restriction is having different rules - then however I don't think we can solve it correctly (well an exception is if we have overlapping ways in OSM which are oneway and opposite to each other - this is a rare case where overlapping ways are not to be removed I guess).



On 4 April 2018 at 10:39, Gerd Petermann <GPetermann_muenchen at hotmail.com<mailto:GPetermann_muenchen at hotmail.com><mailto:GPetermann_muenchen at hotmail.com<mailto:GPetermann_muenchen at hotmail.com>><mailto:GPetermann_muenchen at hotmail.com<mailto:GPetermann_muenchen at hotmail.com><mailto:GPetermann_muenchen at hotmail.com<mailto:GPetermann_muenchen at hotmail.com>>><mailto:GPetermann_muenchen at hotmail.com<mailto:GPetermann_muenchen at hotmail.com><mailto:GPetermann_muenchen at hotmail.com<mailto:GPetermann_muenchen at hotmail.com>><mailto:GPetermann_muenchen at hotmail.com<mailto:GPetermann_muenchen at hotmail.com><mailto:GPetermann_muenchen at hotmail.com<mailto:GPetermann_muenchen at hotmail.com>>>>> wrote:
Hi all,

please help, I just try to make up my mind what mkgmap should do when it finds overlapping road segments and
one (or both) of the overlaps is a part of a (valid) restriction relation.  This doesn't happen very often, but it is possible.

The attached example contains these two overlapping ways, and both are members of (different) restriction relations:
https://www.openstreetmap.org/way/48218016
https://www.openstreetmap.org/way/48218008

(I did not try what the trunk version produces for this mess)

I see different possibilities:
1) ignore route restrictions when removing overlaps, remove those which are invalid after overlapping segments were removed and log a warning
2) ignore overlaps when the ways are members of restriction relations
3) complex: remove overlaps but try to "repair" the restriction

I am currently trying to implement 3) but it looks too complicated for such a rare case and in the end we have a clear case of wrong input data here.
What do you think?

Gerd


_______________________________________________
mkgmap-dev mailing list
mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>>><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>>>>
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev



--
Felix Hartman - Openmtbmap.org & VeloMap.org
Schusterbergweg 32/8
6020 Innsbruck
Austria - Österreich
_______________________________________________
mkgmap-dev mailing list
mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>>>
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev



--
Felix Hartman - Openmtbmap.org & VeloMap.org
Schusterbergweg 32/8
6020 Innsbruck
Austria - Österreich
_______________________________________________
mkgmap-dev mailing list
mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk><mailto:mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>>
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev



--
Felix Hartman - Openmtbmap.org & VeloMap.org
Schusterbergweg 32/8
6020 Innsbruck
Austria - Österreich
_______________________________________________
mkgmap-dev mailing list
mkgmap-dev at lists.mkgmap.org.uk<mailto:mkgmap-dev at lists.mkgmap.org.uk>
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev



--
Felix Hartman - Openmtbmap.org & VeloMap.org
Schusterbergweg 32/8
6020 Innsbruck
Austria - Österreich


More information about the mkgmap-dev mailing list