-
Notifications
You must be signed in to change notification settings - Fork 120
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add label collision rank for pois and other layer #988
Comments
Can we use |
For pois and places that could work. But we already have Looking at Tangram this is called "priority" so I propose |
Subtask of #1351. |
There's a ton of client side logic to set these for POIs layer in particular, moving into v1.7 milestone. |
We can sorta use the sort_rank, except the landuse labels should (a) generally collide out the road labels / that can change by zoom and (b) we need to set them for places and pois layers, so let's do some research... Generally the collision should be driven by min_zoom (or min_label) and then tie break with the relative ranking of the features. @zerebubuth can you generate a CSV that has the following columns, filled out for all the kinds?
|
Here's a CSV with all the kinds with their A few notes on the data:
|
@zerebubuth how should we arrange the |
Right, it's another way of achieving the same thing. Using a function for
`priority` is more inline with our usual pattern for this case anyway
though (vs. one time tile load modifications, which are useful for more
significant transformation of features, or synthesizing entirely new
features, etc.).
…On Thu, Jan 10, 2019 at 8:17 PM Varun ***@***.***> wrote:
Yeah ES doesn't use the scripts property, but I'm not sure how that would
be used here.
If I am not mistaken this is useful for one time evaluation of properties
at tile load time, instead of style building time per feature. (Referring
to @zerebubuth <https://github.com/zerebubuth>'s comment above)
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#988 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AABBXRDBi23bKnSGusQltLsHouU9HXneks5vB-YIgaJpZM4JsNQZ>
.
|
From #988 (comment): See |
Done (left 100)
Done (left 100)
Done (left 100)
Skip... someone should control this with their own combo of min_zoom and probably setting collision order to value in the
Left space for this... needs population_rank logic.
Left space for this... needs shield logic.
Skipped, the min_zoom already encodes this (and when it doesn't the min_zoom should be updated)
Generally implemented.
Skipped, the min_zoom already carries this.
|
I've taken the order I had before, added kind_detail, and explained some of the
(The following table is from all_the_kinds_simplified.csv.zip, which is from the bigger all_the_kinds.csv.zip which also includes the min_zoom and min_label).
|
Just confirming: JS supports functions for |
- floating point values for priority will be used soon once tilezen/vector-datasource#988 lands in tilzen tiles
Yep!
…On Fri, Jan 18, 2019 at 8:09 PM Varun ***@***.***> wrote:
Just confirming:
#988 (comment)
<#988 (comment)>
JS supports functions for priority for both text and point styles right?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#988 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AABBXYd2ZUhsa-Lmu_M1lnt9Ldq5WC23ks5vEnBJgaJpZM4JsNQZ>
.
|
- floating point values for priority will be used soon once tilezen/vector-datasource#988 lands in tilzen tiles
* floating point values for priority style param - floating point values for priority will be used soon once tilezen/vector-datasource#988 lands in tilzen tiles * Add dummy floats for priority to test - TODO: replace with collision data function when tilezen tiles support this
We need to add in the population ranks for places layer See: #988 (comment) for population breaks. While the Here we don't see Los Angeles label (which does label in previous and next zooms), and is a regression over v1.6 with client-side population graded priority ranks. In the case of San Jose south of San Francisco it seems to also be because we're not carrying the NE min_zoom onto the OSM feature, which means park labels then push what was visible out of the way. |
When two different kinds features of equal min_zoom collide, which should win?
I have a big spreadsheet of relative ranks between
kind
&kind_detail
values (as part of the category work). This would remove a good bit of hack logic from the stylesheet, and provide much more fidelity than we have currently there.Since the discussion is so long, the related PR is: #1793.
The text was updated successfully, but these errors were encountered: