Open Map Tiles Style Sheet

[mdoakes42]

I am trying to utilize the OpenMapTiles Style Sheets in Geoserver. Most of the styles are looking for name:latin and/or name:nonlatin. "text-field": "{name:latin} {name:nonlatin}", However, When I generate the vector tiles using the OsmMapTiles profile (or the standard/default profile).. Neither appears to have those fields in the mbtiles and thus I'm getting errors about these fields not existing. Anyone experience this issue and have a work around? Are these fields part of the standard schema? If so why aren't thye exported in the mbtiles? Thanks

[mdoakes42]

Also wondering why the field brunnel does not show up in the "water" layer. This is in the standard schema. I see a note about brunnel not being included in the transportation layer.. But it should still be included here? Any reason why this might happen. I'm using the docker container openmaptiles/planetiler-openmaptiles:latest to run this. Thanks

[msbarry]

Do you have any specific example areas/features? There is code in the openmaptiles profile to set brunnel and also name:latin and name:nonlatin - so you might have found a bug preventing them from getting set in certain situations and we'd need more details to debug.

[mdoakes42]

Thank you, I'll take a look. Similarly, There are fields that are defined in the openmaptiles schema (and included in their styles.json files) but are not in the output.mbtiles file. A couple examples

POI Layer - The rank field is not included
Boundary Layer - claimed_by is missing.

Is this a bug? or is this intentional? Thanks

[mdoakes42]

I think I might see whats happening. When I did a test with "monaco" compared with "egypt" the outputed fields were different. It looks like I only get the fields available from what the source data contains. (IE, There is NOT a standard schema output.. It is more of what is contained in the source data set).

As an example when I dumped egypt I got the aerodrome_label layer. (that wasn't even included in the monaco one).

Also, Starting with boundary for egypt I see extra fields such as adm0_r, disputed_name, adm0_l
Building layer also has a "hide_3d" in egypt which monaco doesn't have.

My guess is if I dump the "planet" file I will see ALL the fields/schema. I think this might be why I was having an issue with fields missing. It was because they didn't exist in the small source that I was querying.

Is this something you can confirm? Thanks

[msbarry]

Oh are you talking about the tilejson metadata? That only contains the attributes present on features in the actual tile output.

[mdoakes42]

brunnel also shows up now with the egypt process.

I think this was just a matter of understanding how the schema's work and that it could be different depending on what fields are available in that source pbf.

[mdoakes42]

I did a full planet dump and the "rank" field is missing in the POI layer. Any ideas why? I'm running this by using docker

docker run -e JAVA_TOOL_OPTIONS="-Xmx64g" -v "$(pwd)/data":/data openmaptiles/planetiler-openmaptiles:latest --osm-path=/data/sources/planet.osm.pbf.............

[msbarry]

Ah I see what's happening. The tilejson attribute metadata is computed during source feature processing, but rank gets added during tile post-processing. So rank is in the output, but it's missing from the tilejson metadata. I opened #749 to move computing the tilejson metadata to when the tiles are emitted instead to account for attributes added/removed during post-processing.

[mdoakes42]

Thank you Michael

[msbarry]

Fixing in #752