[HOT] preset2style.py (converting JOSM presets into style usable with osm2pgsql)
dane at dbsgeo.com
Mon Nov 28 06:45:33 GMT 2011
When I was deployed with Nicolas in Haiti after the 2010 quake I wrote a quick and dirty python script to automate conversion from JOSM XML preset files into something usable by osm2pgsql. It's called 'preset2style.py' and lives at https://github.com/hotosm/scripts.
I never imagined it would still be in use, but it is. COSMHA  workflows in Haiti still actively use it. So, just thought I would send out a quick background to the list on this script in case anyone might benefit from it or have ideas to improve it.
Basically, as you all know, the humanitarian data model  uses custom tags which are often very long. A frequent task for visualizing osm data with custom tags is to import it into PostGIS (postgres database) so that the data can be easily viewed in QuantumGIS and filtered, since shapefiles will truncate long column names.
Osm2pgsql is not the easiest tool to use for beginners but is robust (it is used to import data for rendering at tile.osm.org) and has a windows binary thanks to Jon Burgess (that I was able to package into an easy installer  for haiti deploys). The problem however is that, by default, osm2pgsql only supports a limited set of tags to import into postgis and needs to be extended to notice the HOT tags. Because JOSM preset files are in active use by HOT to guide data collection, it makes sense to automatically pull the tags out of a given preset file, merge these with the default set of osm2pgsql tags, and generate a new style to get the right tags for import. That is basically what this simple script does, not much more.
As I mentioned above, this was a very quick and dirty solution. Lots of ways to improve it. I'm personally following the osmium project  because I think that looks like a great library for flexible osm data conversion. But until we can replace the JOSM->preset2style.py->osm2pgsql workflow with something better (that works on windows), I just wanted to get these details out there :)
More information about the HOT