OHM Planet index is outdated

[1ec5]

The OHM Planet index says it’s still querying an extract from January 17, whereas a much newer extract from March 20 is available on the osm2rdf site. The OSM Planet index is working off the latest extract, so I wonder if something has gotten stuck?

[hannahbast]

@1ec5 Thanks a lot for the reminder. We were busy implementing new GeoSPARQL features and forgot to update OHM. Actually, I just see that the dumps are updated daily, so we could update daily as well. For now, I have updated to the the 21.03.2024 version.

[1ec5]

Thank you!

[hannahbast]

@1ec5 I just updated to the latest dump from https://planet.openhistoricalmap.org/?prefix=planet, dated today. Can you please check whether everything works fine from your side?

And do you know of an easy programmatic way to get the latest version of the dataset? If we had that, we would be happy to create a cronjob that updates https://qlever.cs.uni-freiburg.de/ohm-planet daily (it only takes half an hour)

[1ec5]

And do you know of an easy programmatic way to get the latest version of the dataset?

As far as I can tell, https://planet.openhistoricalmap.org/planet/state.txt returns the URL of the latest planet file as plain text. Would that be adequate?

The data looks much fresher at a glance. Nodes from within the past day are now being returned (though a few are reported as coming from the future for some reason). With these regular updates, QLever will be quite useful for detecting broken boundary relations and mistagged features based on area. Thank you!

[hannahbast]

@1ec5 Great, https://qlever.cs.uni-freiburg.de/ohm-planet should now update daily, please let me know when it doesn't. Note that the first "Map view" after a rebuild takes a minute or so because the viewer needs to fetch all the geometries once.

[1ec5]

Awesome, thank you!