Just downloaded to take a look, Most seem right. In order to QA something like this, taking random samples won't help since any number of errors would be small.
Does the script currently check for a byte-to-byte match of the name=* text? I think that should be the determining faction of adding a node to the relation. The remaining un-merged entries can be determined manually. On 10/11/18 10:44 am, Andrew Davidson wrote: > If you want to start working on something I have had a go at matching > place nodes: > > http://overpass-turbo.eu/s/Dy3 > > with the corresponding admin_level 10 boundaries: > > https://github.com/FrakGart/TestAdmin10Labels > > These have been matched by finding the best matching place node by > name within a bbox 150% the size of the admin boundary. In the cases > where there are more than one perfect match the first one the program > encountered is used. > > Needs reviewing to see if the most appropriate node has been selected. > Particularly where a town/city node has been selected: do we need to > add a suburb node? (or do we use the same node for city and suburb eg: > Brisbane) > > This is v1.0, I was planning to see if I could flag if the selected > node is inside/outside the boundaries (eg: Mount Isa) and also look > for duplicate place nodes (ie: same name appearing on different values > of place). > > > On 7/11/18 2:57 pm, Joel H. wrote: >> How are we with our import plan on the wiki, Do we have any blockers >> stopping the PSMA import? >> >> I think we should start setting dates state-by-state for import. >> >> >> _______________________________________________ >> Talk-au mailing list >> Talk-au@openstreetmap.org >> https://lists.openstreetmap.org/listinfo/talk-au >> > > _______________________________________________ > Talk-au mailing list > Talk-au@openstreetmap.org > https://lists.openstreetmap.org/listinfo/talk-au _______________________________________________ Talk-au mailing list Talk-au@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-au
