[mapserver-users] Re: Clustering and tile clipping

[Wim Vanbelle]

Does anyone have any working examples of how they implemented clustering
without the tile clipping?

Thanks.

On 11 January 2012 10:00, Wim Vanbelle <wimvanbelle at gmail.com> wrote:

> Hi,
>
> I'm very happy that mapserver is now capable of server side clustering.
> However I wonder if it's possible to completely get rid of the tile
> clipping.
>
> Currently I've clustered a point layer with 44k points:
>
>             CLUSTER
>                         MAXDISTANCE 100
>                         REGION "ellipse"
>
>                 END
>
> It is displayed with a simple image:
>                        STYLE
>                                 SYMBOL          "small_shaded.png"
>                                 SIZE            21
>                                 OUTLINECOLOR    0 0 0
>                         END
>
> However, when viewing this on google maps, the images are often clipped at
> the tile edges.
>
> I figured that I could solve this with mapcache:
>
>  <tileset name="ld">
>     <source>ld</source>
>     <cache>disk</cache>
>     <grid>g2</grid>
>     <format>PNG</format>
>     <metatile>2 2</metatile>
>     <metabuffer>30</metabuffer>
>     <expires>3600</expires>
>     <auto_expire>2592000</auto_expire>
>   </tileset>
>
> i've tried higher values for metatile and it helps a bit, but the problem
> simply shifts location. Also with higher values (eg 5 5 metatile) mapserver
> doesn't return the tiles on time for mapcache and I get gaps in my results.
>
> Would preseeding the data help with the timeouts? Any other solutions that
> I may try? The metadata parameter for tile_edge buffer only works with
> labels as far as i know.
>
> Thanks in advance, this issue has been troubling me for some time now :).
>
> Best regards,
> Wim
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/mapserver-users/attachments/20120116/810e2bc3/attachment.htm>