[Patch v1] addr:interpolation

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

[Patch v1] addr:interpolation

Gerd Petermann
Hi all,

attached is a 1st try to implement the evaluation of the
addr:interpolation tag.
In some areas this is heavily used, so it really improves address search.

addr-interpolation-v1.patch

The implementation is straight forward:
before style evaluation
- check for tag addr:interpolation, make sure the value is one of "odd","even", or "all"
- if the way has two or more nodes with addr:housenumber=* and addr:street=*,
the way is a candidate for the interpolation and a tag mkgmap:node-ids is added to
the way to be able to find them later

after style evaluation:
- for each pair of housenumber nodes :
    calculate the positions of the interpolated housenumbers and
  generate a node containing the housenumber and streetname

I am not sure what to do with partly invalid data, e.g.
a way with addr:interpolation=odd in combination with more than two nodes, the last
having an even number. The current implementation will generate odd nodes for
all valid pairs in that way.

Effect: When searching for an address, it is more likely that you
will not be pointed to a much higher or lower number.
I did not notice an increase of img size.
I assume in a few areas the additional calulations may require more CPU
or memory, but the effect should be rather small.

Gerd
Reply | Threaded
Open this post in threaded view
|

Re: [Patch v1] addr:interpolation

Gerd Petermann
Hi all,

I found a few more special cases:
http://www.openstreetmap.org/way/33747687

The way has three nodes, but the one in the middle
doesn't have any tags. The patch is not handling this
and will generate nodes at a wrong position.
I am not sure if the way is correct or not?

Another possible problem:
The wiki says"
If there is a house number on a single node or single building polygon and that house number also appears as the result of an interpolation, software should handle this case gracefully and favour the individually tagged house number as the real position. "
The patch doesn't implement it this way, but I'll try to code it tomorrow.

Gerd




> Date: Sun, 7 Dec 2014 06:53:45 -0700
> From: [hidden email]

> To: [hidden email]
> Subject: [mkgmap-dev] [Patch v1] addr:interpolation
>
> Hi all,
>
> attached is a 1st try to implement the evaluation of the
> addr:interpolation tag.
> In some areas this is heavily used, so it really improves address search.
>
> addr-interpolation-v1.patch
> <http://gis.19327.n5.nabble.com/file/n5826523/addr-interpolation-v1.patch>
>
> The implementation is straight forward:
> before style evaluation
> - check for tag addr:interpolation, make sure the value is one of
> "odd","even", or "all"
> - if the way has two or more nodes with addr:housenumber=* and
> addr:street=*,
> the way is a candidate for the interpolation and a tag mkgmap:node-ids is
> added to
> the way to be able to find them later
>
> after style evaluation:
> - for each pair of housenumber nodes :
> calculate the positions of the interpolated housenumbers and
> generate a node containing the housenumber and streetname
>
> I am not sure what to do with partly invalid data, e.g.
> a way with addr:interpolation=odd in combination with more than two nodes,
> the last
> having an even number. The current implementation will generate odd nodes
> for
> all valid pairs in that way.
>
> Effect: When searching for an address, it is more likely that you
> will not be pointed to a much higher or lower number.
> I did not notice an increase of img size.
> I assume in a few areas the additional calulations may require more CPU
> or memory, but the effect should be rather small.
>
> Gerd
>
>
>
> --
> View this message in context: http://gis.19327.n5.nabble.com/Patch-v1-addr-interpolation-tp5826523.html
> Sent from the Mkgmap Development mailing list archive at Nabble.com.
> _______________________________________________
> mkgmap-dev mailing list
> [hidden email]
> http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

_______________________________________________
mkgmap-dev mailing list
[hidden email]
http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev