Tiers for landuse

[zerebubuth]

Adds a property representing the "tier" from #473; an integer from 1-6.

Connects to #473.

@rmarianski could you review, please?

[rmarianski]

👍

[nvkelso]

1. park or protected land is an outmoded v0.10 reference, it should be removed I think?
2. Make sure !United States Forest Service matches Olga's change from last week?

[nvkelso]

Looks like this is written more for tier than scalerank?

What about adding an additional column for scalerank for the different area brackets (more on a 1 to 10 scale?) so features within the same kind would get different scalerank values based on their areas (my additional proposal), but all be in the same tier based on the kind (as currently written in the PR).

In the stylesheet it would be nice to be able to filter generically on tier and scalerank for most landuse features.

[zerebubuth]

@rmarianski I've changed the code somewhat to allow CSVs to have multiple output columns. Please could you take a look and let me know what you think?

@nvkelso Please look at the test and let me know if that's what you wanted from tier and scalerank.

[rmarianski]

👍

[rmarianski]

range defaults to 0, but is this equivalent?

for (is_target, key, typ), val in zip(row, cols):

[zerebubuth]

Yes! That's much better, thanks. Fixed in 1cf7672.

[nvkelso]

Tests are looking good. But one question, what is the 2nd test on

vector-datasource/integration-test/473-landuse-scalerank.py

Lines 13 to 18 in 57aeb9c

	{ 'kind': 'national_park', 'id': -921675, 'tier': 1,
	'scalerank': 8 })
	assert_has_feature(
	7, 28, 57, 'landuse',
	{ 'kind': 'national_park', 'id': -921675,
	'tier': type(None), 'scalerank': type(None) })

meant to do? Test that the feature not be included at the lower zoom or that it not receive the tier and scalerank?

That US National Forest thing is definitely not a national_park ;)

[nvkelso]

Debug (colors, filters) Tangram file:

• https://raw.githubusercontent.com/tangrams/bubble-wrap/fc6c44dff24abd42488fc8d3947241d0fef766ce/bubble-wrap.yaml

Prettified (same as debug but with pretty colors) Tangram file:

• https://raw.githubusercontent.com/tangrams/bubble-wrap/2976ad86757d128a39bcfda321c3bc52a81b49a1/bubble-wrap.yaml

[zerebubuth]

But one question, what is the 2nd test meant to do? Test that the feature not be included at the lower zoom or that it not receive the tier and scalerank?

The latter - I thought that's what this issue was about: Providing a tier and scalerank on features that we were being opinionated about for our styles, but not changing the actual min_zoom?

Therefore the features which we don't want to display at a particular zoom are still present in the tiles if someone else wants to show them earlier. I wasn't entirely sure based on the description in #473. It would be redundant if min_zoom == scalerank, so I thought it must have a different meaning.

Please let me know if you wanted to alter the actual min_zoom values, as that would be a completely different implementation.

That US National Forest thing is definitely not a national_park ;)

Tracking in #987.

[zerebubuth]

Following on from discussions about protect_class, I'm having trouble understanding what the desired behaviour is. It would be really helpful if we could pick out some concrete examples of things having the different properties which should and should not get assigned to the different tiers.

[zerebubuth]

Closing this as per discussion. There will be a new PR implementing this by changing the min zoom rules in the YAML and applying the area filters and tier values there.