[Qgis-developer] Server features
Bernhard Ströbl
bernhard.stroebl at jena.de
Thu Nov 15 00:54:11 PST 2012
Hi Andreas,
what you describe is where I use tiles. My city map is a QGIS project
with around 25 layers, lots of labels (street names) and elements. I use
mapcache (standalone) for that and of course it would be nice to have
that integrated in QGIS server but it was not hard to configure it by
itself. Creating the tiles is slower than direct rendering from QGIS
server but accessing is a lot faster than QGIS server (and less CPU).
eSo caching is an option if data changes do not need to be visible
immediately. If mapcache could be integrated in QGIs Server we could
benefit from faster service and have legend and feature info, too.
So +1 for that.
What I really miss in QGIS server compared to UNM mapserver is the
possibility to keep a certain space around the edge of the generated
output free of labels (UMN Mapserver has had that already in V 5.? or
even earlier). This is a must-have if we want to integrate mapcache
because otherwise there will be broken labels and tiles will not fit to
each other labelwise.
just my 2 ct
Bernhard
Am 15.11.2012 08:57, schrieb Andreas Neumann:
> Hi,
>
> For 90% of my projects performance is really good. For the remaining
> 10%, rather complext projects, I would say it is good enough, but it
> could be "snappier".
>
> I am specifically talking about the "Leitungskataster" project, the
> project displaying all supply lines and waste-water lines, containing
> cadastral data, electricity, gaz, water, communication, waste-water and
> heating.
>
> These complex projects typically have:
>
> * a lot of layers (around 70 layers)
> * a lot of rules in the layer symbology
> * automatically placed labels along lines
> * dash patterns - almost all the lines are dashed
> * SVG symbols
>
> I will have to do my own homework. I will probably have to reduce the
> number of layers by using views.
>
> And to be honest, I haven't set up the same project on UMN mapserver to
> compare whether it is really faster.
>
> I don't know yet where the bottlenecks are (too many layers, too many
> rules, too many rules, too many labels?).
>
> Thanks,
> Andreas
>
>
> Am 14.11.2012 21:13, schrieb Marco Hugentobler:
>> Hi all
>>
>> Performance is a complex topic, it essentially depends what you are
>> testing ( few complex polygons, many small polygons, type of raster,
>> lots of labels, ...).
>> I'm regularly profiling some of my important server projects (e.g. WMS
>> of cadastral survey). One thing that takes up a lot of render time is
>> the drawing of the labels (only the rendering, not the placement).
>> I think this is because they are rendered using QPainterPath and not
>> with QPainter->drawText (but it might be difficult to render the buffer
>> using drawText).
>>
>> And don't forget that normally the transfer of the image over the
>> internet takes more time than the WMS rendering itself. So don't forget
>> to use jpg for rasters and the improved png8 for vectors. It can improve
>> perceived performance by a factor of 3 or 4.
>>
>> Regards,
>> Marco
>>
>> On 14.11.2012 19:57, Andreas Neumann wrote:
>>> Lets say that speed of QGIS Server is good enough for most of my
>>> projects. But it could be faster for more complex rendering and labeling.
>>>
>>> UMN feels faster than QGIS Server - but I don't have time for double
>>> configurations for desktop and web. Plus - it is so much easier to
>>> configure QGIS projects compared to editing mapfiles.
>>>
>>>
>>>
>>> "G. Allegri"<giohappy at gmail.com> schrieb:
>>>
>>> The one thing were UMN Mapserver is a much better is
>>> performance. This
>>> is where we should really catch up.
>>>
>>>
>>> I would put this to the higher position! I don't know where the
>>> effort should be put, because I suppose it's something deep in the
>>> rendering engine of QGis. Do you see bottlenecks at the server level?
>>>
>>> I would also add a point to the whishlist: Qgis Server Python API ;)
>>>
>>> giovanni
>>>
>>> 2012/11/14 kimaidou<kimaidou at gmail.com<mailto:kimaidou at gmail.com>>
>>>
>>> Hi
>>>
>>> 2012/11/14 Vincent Picavet<vincent.ml at oslandia.com
>>> <mailto:vincent.ml at oslandia.com>>
>>>
>>> Hi,
>>>
>>> > * templates for HTML popups (already done this in
>>> Lizmap, but why not
>>> > having it in Qgis)
>>> You mean accessing the templates through QGIS server ?
>>> Because html popups
>>> already are in QGIS :)
>>>
>>>
>>>
>>> I mean through Qgis Server (in respect to this thread title ;)
>>> Just exposing methods for the server must do it, as Andreas
>>> proposed.
>>>
>>> We should have a look at PyWPS too, which could be an easy
>>> way to provide WPS
>>> services around QGIS API.
>>>
>>>
>>>
>>> Yes, pyWPS could help provide this feature, but we still need
>>> a way so that QGIS Server can use python plugins / processes,
>>> etc.. I remember discussion about this during last hackfests,
>>> but I do not know if someone gave it a try.
>>>
>>> Anyway, +10000 for focusing on performance first, which is the
>>> key "feature" for a Server.
>>>
>>> Michael
>>>
>>>
>>> Vincent
>>>
>>> >
>>> > Good evening
>>> > Michael
>>> >
>>> >
>>> > 2012/11/14 G. Allegri<giohappy at gmail.com
>>> <mailto:giohappy at gmail.com>>
>>> >
>>> > > > - integrated tiling
>>> > >>
>>> > >> Why not integrating mapcache directly to mutualize
>>> effort on this part ?
>>> > >> Mapcache can be used standalone, and could read its
>>> parameters directly
>>> > >> from a
>>> > >> qgis configuration file. No need to reinvent the
>>> wheel here, or would
>>> > >> there be a
>>> > >> good reason to it ?
>>> > >
>>> > > I think that a tile generator from the QGis Desktop
>>> side would suffice.
>>> > > Any other TMS/WMS-T server could use the tile structure.
>>> > >
>>> > >> > - mask layers
>>> > >>
>>> > >> That would be great to have in QGIS indeed. It would
>>> mean having for
>>> > >> each layer an attached mask layer, which is not
>>> displayed but only used
>>> > >> to define
>>> > >> hidden places.
>>> > >>
>>> > >> > - symbol placement
>>> > >> > - label offsets
>>> > >>
>>> > >> That's for qgis symbology, not directly a qgis server
>>> issue. Label
>>> > >> offset in
>>> > >> mapserver definitly looks cool !
>>> > >
>>> > > I agree Vincent.
>>> > >
>>> > >> Vincent
>>> > >> _______________________________________________
>>> > >> Qgis-developer mailing list
>>> > >> Qgis-developer at lists.osgeo.org
>>> <mailto:Qgis-developer at lists.osgeo.org>
>>> > >> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>>> > >
>>> > > _______________________________________________
>>> > > Qgis-developer mailing list
>>> > > Qgis-developer at lists.osgeo.org
>>> <mailto:Qgis-developer at lists.osgeo.org>
>>> > > http://lists.osgeo.org/mailman/listinfo/qgis-developer
>>> _______________________________________________
>>> Qgis-developer mailing list
>>> Qgis-developer at lists.osgeo.org
>>> <mailto:Qgis-developer at lists.osgeo.org>
>>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>>>
>>>
>>>
>>> _______________________________________________
>>> Qgis-developer mailing list
>>> Qgis-developer at lists.osgeo.org
>>> <mailto:Qgis-developer at lists.osgeo.org>
>>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>>>
>>>
>>>
>>> ------------------------------------------------------------------------
>>>
>>> Qgis-developer mailing list
>>> Qgis-developer at lists.osgeo.org
>>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>>>
>>>
>>> --
>>> Diese Nachricht wurde von meinem Android-Mobiltelefon mit K-9 Mail
>>> gesendet.
>>>
>>>
>>> _______________________________________________
>>> Qgis-developer mailing list
>>> Qgis-developer at lists.osgeo.org
>>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>>
>>
>>
>>
>> _______________________________________________
>> Qgis-developer mailing list
>> Qgis-developer at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>>
>
> _______________________________________________
> Qgis-developer mailing list
> Qgis-developer at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
>
> __________ Information from ESET Mail Security, version of virus signature database 7693 (20121114) __________
>
> The message was checked by ESET Mail Security.
> http://www.eset.com
>
>
--
Bernhard Ströbl
Anwendungsbetreuer GIS
Kommunale Immobilien Jena
Am Anger 26
07743 Jena
Tel.: 03641 49- 5190
E-Mail: bernhard.stroebl at jena.de
Internet: www.kij.de
Kommunale Immobilien Jena
Eigenbetrieb der Stadt Jena
Werkleiter: Dr. Götz Blankenburg
__________ Information from ESET Mail Security, version of virus signature database 7693 (20121114) __________
The message was checked by ESET Mail Security.
http://www.eset.com
More information about the Qgis-developer
mailing list