Hi,
Given that the modules are unsupported and not working properly anyways, I’d >rather fix things to follow GDAL’s approach.
Any objection?
No objection. If two open source GIS software under the OSGeo umbrella do something in the same way we can start calling it as a de facto standard.
Even has been committing a lot to GPKG lately but mostly to the vector side. It might still be good to check if GDAL still follows the same approach as you are planning to follow.
-Jukka Rahkonen-
Andrea Aime wrote:
[Geoserver-devel] Geopackage tiles generation issues
···
Hi,
I am looking into why the raster geopackages GeoServer generates via the geopackage module are not
being read correctly by gdal and found a few issues, some straight bugs, other more like odd design decisions.
The straight bug is what’s causing GDAL not to read the files correctly, the current code, shared with mbtiles
generation, forces the request to be cast on a GWC managed gridset (and actually works only if the requested
layers are cached, which is a bit of a nonsense to me given this is a WMS request…), and uses that gridset
coordinates for tiles.
Now the problem is that in the geopackage metadata for the layer and the gridset we are declaring the requested bbox, not the
ones used for tiles generation, resulting in tile coordinates being off the expected grid.
Now, geopackage natively has the layer bounds that are separate from the tile matrix ones, but the current geotools code
forces them to be the same… that’s a limitation I’m fixing.
However, the main issue remains… if someone asks for a layer in a certain SRS, why try to coax it into a tile matrix managed
by GWC?
I believe the intention was to try and leverage GWC cache, but still, while I understand taking the opportunity if possible, in
my opinion the request of the user should be respected, and if the request is not a match either by SRS or BBOX to a tileset,
GeoServer should not force a different response.
From the same line of thought, a WMS request should not fail if it’s made against a layer that is not cached by GWC, like it
does today, with a strange error message too, e.g. asking for UTM 32N, which is not tile cached:
results in:
java.lang.IllegalArgumentException: Thread 19276 Unknown layer nurc:mosaic. Check the logfiles, it may not have loaded properly
The second thing that I find confusing is, if I asked for a result image that’s 768 pixels wide, why am I getting back a geopackage that’s
over 100000 pixels wide? That’s… insane, especially for a WMS request (large jobs should be deferred to WPS asynch requests).
Looking at the code, there is logic that generates the min and max zoom levels for the tile matrix, and it does something rather surprising…
it picks the zoom level closest to the requested one as the min zoom level, and then starts adding zoom levels until the matrix
has (stricly) more than 256 tiles in width… in my case, that means 512 tiles, or a total width of 512*256=131072 pixels!
Trying GDAL on the command line, translating from a geotiff to a geopackage, it makes what looks like a more sensible choice instead: it uses the
requested zoom level as the maximum one, and then adds overviews to it, in the cases I’ve tried, until a single tile is left (basically, a good
output with overviews). With our case we can also generate raster geopackages out of vector layers, and creating the overviews might end up
hitting badly setup styles displaying too much data, but still, that would be a problem also in random WMS requests, so I’m not too concerned…
Also, by default, GDAL generates a gridset that’s an exact match for the source data, meaning a custom one. That also makes sense to me, if someone
wants something other than the BBOX/SRS requested there are format options to choose both the target gridset and min/max zoom levels.
Given that the modules are unsupported and not working properly anyways, I’d rather fix things to follow GDAL’s approach.
Any objection?
Cheers
Andrea
–
==
GeoServer Professional Services from the experts! Visit
http://goo.gl/it488V for more information.
==
Ing. Andrea Aime
@geowolf
Technical Lead
GeoSolutions S.A.S.
Via di Montramito 3/A
55054 Massarosa (LU)
phone: +39 0584 962313
fax: +39 0584 1660272
mob: +39 339 8844549
http://twitter.com/geosolutions_it
AVVERTENZE AI SENSI DEL D.Lgs. 196/2003
Le informazioni contenute in questo messaggio di posta elettronica e/o nel/i file/s allegato/i sono da considerarsi strettamente riservate. Il loro utilizzo è consentito esclusivamente al destinatario del messaggio, per le finalità indicate nel messaggio stesso. Qualora riceviate questo messaggio senza esserne il destinatario, Vi preghiamo cortesemente di darcene notizia via e-mail e di procedere alla distruzione del messaggio stesso, cancellandolo dal Vostro sistema. Conservare il messaggio stesso, divulgarlo anche in parte, distribuirlo ad altri soggetti, copiarlo, od utilizzarlo per finalità diverse, costituisce comportamento contrario ai principi dettati dal D.Lgs. 196/2003.
The information in this message and/or attachments, is intended solely for the attention and use of the named addressee(s) and may be confidential or proprietary in nature or covered by the provisions of privacy act (Legislative Decree June, 30 2003, no.196 - Italy’s New Data Protection Code).Any use not in accord with its purpose, any disclosure, reproduction, copying, distribution, or either dissemination, either whole or partial, is strictly forbidden except previous formal approval of the named addressee(s). If you are not the intended recipient, please contact immediately the sender by telephone, fax or e-mail and delete the information in this message that has been received in error. The sender does not give any warranty or accept liability as the content, accuracy or completeness of sent messages and accepts no responsibility for changes made after they were sent or for other risks which arise as a result of e-mail transmission, viruses, etc.
