admin boundaries not rendering

[jeffreyameyer]

Could be user / tagging errors, but it's unclear why some admin boundaries are not rendering in vector tiles.

Example area: there are about 3 different boundary sets around Maine that should be showing up, but aren't. I grabbed one border way, moved it slightly & closed it for testing here: https://openhistoricalmap.org/way/198293513#map=9/45.8977/-69.0765&layers=H

You can see it here on the vector tiles: https://openhistoricalmap.github.io/openhistoricaltiles/mbgl-control-timeslider/demo/#5.307/45.24678/-69.70763/1797,-4000-2018

┆Created By: Dan Rademacher

[danrademacher]

This appears related to #51, where @daelba noted some specific issues around which admin_level tags we do or don't render. There @mojodna linked to https://github.com/OpenHistoricalMap/openmaptiles/blob/3710bbcddc5dea932302c8049cb73969891916b7/layers/boundary/boundary.yaml#L6, which makes it sound like we're styling only levels 4 and 6. Since Jeff's boundary has no level at all I would expect it to not render.

Daniel proposes some options for how we treat levels over in #51. To me this is another vote in favor of renewed more systematic focus on the stylesheet, which we quickly adapted from OpenMapTiles, but now we're finding the gaps!

[jeffreyameyer]

Ha! I did have an admin_level - 4 to be sure, but it still wasn't rendering. A different boundary that had no admin_level was rendering, so I tried to see if it would work without it...

On Mon, Sep 9, 2019 at 4:41 PM Dan Rademacher notifications@github.com wrote:

This appears related to #51 https://github.com/OpenHistoricalMap/openhistoricaltiles/issues/51, where @daelba https://github.com/daelba noted some specific issues around which admin_level tags we do or don't render. There @mojodna https://github.com/mojodna linked to https://github.com/OpenHistoricalMap/openmaptiles/blob/3710bbcddc5dea932302c8049cb73969891916b7/layers/boundary/boundary.yaml#L6, which makes it sound like we're styling only levels 4 and 6. Since Jeff's boundary has no level at all

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/OpenHistoricalMap/openhistoricaltiles/issues/60?email_source=notifications&email_token=AALM4ESHEGLI7ANZO6CMGJ3QI3NJRA5CNFSM4IU5XDNKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD6JLJKI#issuecomment-529708201, or mute the thread https://github.com/notifications/unsubscribe-auth/AALM4EWCRAODU3RIDW2VDUDQI3NJRANCNFSM4IU5XDNA .

-- Jeff Meyer 206-676-2347 osm: Open Historical Map (OHM) http://wiki.openstreetmap.org/wiki/Open_Historical_Map / my OSM user page http://www.openstreetmap.org/user/jeffmeyer t: @OpenHistMap

[gregallensworth]

Checking the osm_border_polygon table, I do not see it in the tables.

> select osm_id, name from osm_border_polygon where osm_id = 198293513;
(0 rows)

> select osm_id, name from osm_border_polygon where name like 'test%';
(0 rows)

So this is not a problem in the vector tiles nor the style, but the ingestion process.

Looking at the mapping.yaml which defines what features go where, I see this as the filters. I do not see an additional restriction by admin_level or other tags, but I do see a filter by relation_type.

    mapping:
      boundary:
      - administrative
    relation_types:
    - boundary

This looks as if it would ingest only relations but not ways such as these features?

Other boundaries which are present, such as Maine District of Massachusetts, look as if they are fairly old and so may have been imported into the database through other means?

[danrademacher]

Chatted a bit about this with @mojodna while we were discussing other stuff, and he suggested that the best way forward is simply for @jeffreyameyer to make in JOSM a relation with this one way as a member, and then tag the relation. The assumption is that most borders are too complex to be single ways.

Per API, https://api.openstreetmap.org/api/0.6/capabilities, and https://help.openstreetmap.org/questions/365/is-there-a-limit-on-the-number-of-nodes-in-a-way ways are limited to 2000 nodes, but relations can have an unlimited numbers of members.

So easiest way forward here IMO is to stick with the idea that borders must be relations

[jeffreyameyer]

Ahhh... I've tried with relations, as well. No luck there, either.

On Fri, Sep 13, 2019 at 6:16 PM Dan Rademacher notifications@github.com wrote:

Assigned #60 https://github.com/OpenHistoricalMap/openhistoricaltiles/issues/60 to @jeffreyameyer https://github.com/jeffreyameyer.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/OpenHistoricalMap/openhistoricaltiles/issues/60?email_source=notifications&email_token=AALM4ESSJ2OBLNZ7D7CCKJLQJQ3OZA5CNFSM4IU5XDNKYY3PNVWWK3TUL52HS4DFWZEXG43VMVCXMZLOORHG65DJMZUWGYLUNFXW5KTDN5WW2ZLOORPWSZGOTT6XZYY#event-2633858275, or mute the thread https://github.com/notifications/unsubscribe-auth/AALM4ERENR343O6DD4AAAU3QJQ3OZANCNFSM4IU5XDNA .

-- Jeff Meyer 206-676-2347 osm: Open Historical Map (OHM) http://wiki.openstreetmap.org/wiki/Open_Historical_Map / my OSM user page http://www.openstreetmap.org/user/jeffmeyer t: @OpenHistMap

[danrademacher]

Hmm, OK, so much for that. Can you share a link to a relation that we're also failing to render?

[jeffreyameyer]

Here are a few:

https://www.openhistoricalmap.org/relation/2660772#map=7/45.314/-68.988&layers=H

admin_level = 4

https://www.openhistoricalmap.org/relation/2660784 - admin_level = 4

https://www.openhistoricalmap.org/relation/2660783 - admin_level = 4

On Tue, Sep 17, 2019 at 10:04 AM Dan Rademacher notifications@github.com wrote:

Hmm, OK, so much for that. Can you share a link to a relation that we're also failing to render?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/OpenHistoricalMap/openhistoricaltiles/issues/60?email_source=notifications&email_token=AALM4EQRZEOBEOSUMU5KLJTQKEE3HA5CNFSM4IU5XDNKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD65GZDI#issuecomment-532311181, or mute the thread https://github.com/notifications/unsubscribe-auth/AALM4EVFEUNAV2PO4IJZNZ3QKEE3HANCNFSM4IU5XDNA .

-- Jeff Meyer 206-676-2347 osm: Open Historical Map (OHM) http://wiki.openstreetmap.org/wiki/Open_Historical_Map / my OSM user page http://www.openstreetmap.org/user/jeffmeyer t: @OpenHistMap

[danrademacher]

So overall this seems similar to #59 where we need to change how data is ingested into our vector tile database and the reconsume the whole planet to get missing stuff. Best to line up a number of these and do fewer full planet wipe/reingest.