Hi Peter, I did some Java-hacking last night and my osmosis doesn't ignore the visibility anymore. I'll do some testing, because I guess I can import AOIs of the planetfile into the api-schema by filtering the full-planet using --bp and not populating the currenttables. This can be a basis for some analyses or a nice test-environment.
I think it's a clean behaviour if --read-xml file =- write-xml file=- doesn't change anything, or is this nonsense? Because I'm not really a Java-developer, I think my changes are not very nice implemented and may break other features (just added "boolean visible" at each entity, node, way, ... in the osmosis-source and did some cleaning). If this feature is to be implemented, I can share my work, but I'm not sure if it's less work if Brett just implements it (if he likes to) without my dirty code. The simple Schema is missing the visibility-column anyway - may be I'll do some (dirty) patch for this as well. Marco Am 09.09.2010 16:09, schrieb Peter Körner: > Hi Marco, > > I feel the need to add a disclaimer: you got further then I have > tested until now. All I tested with were simple hand-written osm files > containing three or four nodes and a single way, not even getting near > to a planet extract. > > Relations will be simply written to the db, there's no special > handling for them but they wont do any harm either. > > The visibility-flag is a problem I'm aware of. It won't be hard to add > another column to the ways table, but it's a bigger issue to get it > through the osmosis pipeline. > > I was not able to take a deep look on the planet history file, so I > don't have a clue if it countains a visible-flag but I'd guess so. > > The default xml reader tasks ignore the flag and the default domain > objects (Node, Way, & Co.) don't have a field for it. But I already > implemented my own domain class - MinorWay - which adds a new property > to the way (the minor version number), so that it might be possible to > to this also with the Node class. I would need to extend the xml > reader tasks and offer my own --read-history-xml. It would then still > be possible to use the default --read-xml tasks but then the > visibility will get thrown away. > > Peter > > Am 07.09.2010 13:41, schrieb Marco Lechner - FOSSGIS e.V.: >> Hi Peter, hi everybody, >> >> I did some tests with the history-plugin and was able to import an area >> of interest (AOI) of the full-planet.dump by filtering it through a >> bounding-polygon. It worked without problems (I had to remove the >> relation s of course, because this is not supported yet). >> >> How can I find out if an object (point, way) is "visible" = marked as >> deleted? By importing an AOI of a planet-file this information is not >> necessary, because just the currently visible objects are imported (only >> these are part of the planet-file), but if it comes to the history-dump >> there might be nodes that were visible and have been deleted in a more >> recent version. How could we handle this information? I guess this >> information is not available in the history-dump, is it? What happens to >> "deleted" objects while creating the full-history.dump? Are they >> ignored? Is just the "invisible" version ignored? Is the "invisible" >> version in the history-dump but you can not reproduce the information? >> >> Marco _______________________________________________ osmosis-dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/osmosis-dev
