-
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
Drop bicycle network and shield text to match style. #1707
Conversation
Drop RCN from 12 down to 0, LCN from 15 down to 0, and any network on track paths from 12 down to zero. These properties inhibit some degree of merging at mid zooms. Context: #1331 (comment)
Follow-up: On |
Wow, that’s great!
Can you stat here the tile so I can take a look, please?
…On Mon, Nov 19, 2018 at 07:52 Matt Amos ***@***.***> wrote:
Follow-up: On 12/1206/1540, which is basically Brooklyn, the number of
minor_road features is 38% lower after the change. This means a lot more
merging is going on. The layer as a whole is 8% smaller.
—
You are receiving this because your review was requested.
Reply to this email directly, view it on GitHub
<#1707 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AA0EOy9YYEWv5M51Sp3JFnhIUTssvRqPks5uwtPNgaJpZM4YpB5h>
.
|
For the
Which is what we expected: A big reduction in the number of
Note that nothing has really changed for any highway kind features. The changes in the size of the properties is likely to be due to re-arrangement of the string table, leading to small changes to the length of the varints used to encode the key and value indexes. For the layer as a whole, including the string table:
(PS: Looks like I got my maths a little wrong before - the reductions are 35% and 7% for the minor roads count and layer size, rather than 38% and 8% - perhaps I summed up the counts and byte sizes together by mistake or something.) The GeoJSON for the |
queries.yaml
Outdated
properties: | ||
- bicycle_network | ||
- bicycle_shield_text | ||
- all_bicycle_networks |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Dropping all_bicycle_networks
and all_bicycle_shield_texts
all the time until the max zoom would be best (not limited to just lcn
, for any bicycle network).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Dropped in 2f7728a, and cleaned up some of the other logic for dropping all_bicycle_*
start_zoom: 0 | ||
end_zoom: 13 | ||
properties: | ||
- bicycle_network |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reviewing https://github.com/tangrams/walkabout-style/blob/gh-pages/layers/bike-interlay.yaml I see that the network IS referenced like bicycle_network: [icn,ncn,rcn]
at lower zooms, but rcn
is ignored at super low zooms like filter: { bicycle_network: rcn, $zoom: [8,9,10] }
:
https://github.com/tangrams/walkabout-style/blob/gh-pages/layers/bike-interlay.yaml#L1788-L1797
So it would be:
- okay to drop
bicycle_network
from 0 to 11 (or is that 10, I get confused) forrcn
- okay to drop
bicycle_shield_text
from 0 to 12 (one more zoom thanbicycle_network
is dropped) forrcn
- see other comment about always dropping the
all*
variants until the max zoom) even forrcn
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think I've captured this in 324e67f, but would be good to have a check over the tests to make sure the logic is correct.
start_zoom: 0 | ||
end_zoom: 16 | ||
properties: | ||
- bicycle_network |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It would be more conservative to do:
- okay to drop
bicycle_network
from 0 to 14 (2 less than max zoom) forlcn
- okay to drop
bicycle_shield_text
from 0 to 15 (one more zoom thanbicycle_network
is dropped) forlcn
– they should be available for labeling at zoom 15 like in https://tangrams.github.io/walkabout-style/#15/37.7844/-122.4416
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think I've captured this in 324e67f, but would be good to have a check over the tests to make sure the logic is correct.
params: | ||
source_layer: roads | ||
start_zoom: 0 | ||
end_zoom: 13 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
For reference this comes from:
https://github.com/tangrams/walkabout-style/blob/gh-pages/layers/bike-interlay.yaml#L569-L573
end_zoom: 13 | ||
properties: | ||
- bicycle_network | ||
- bicycle_shield_text |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should this include the all*
variants, too?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is answering a slightly different question, but: Since 2f7728a, we drop the all_bicycle_*
stuff at anything except max zoom, so we don't need to drop it anywhere else - it's already gone.
Look at the sample GeoJSON file we can probably also drop https://tangrams.github.io/walkabout-style/#13/37.8177/-122.5060 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great work! See a few comments and then I think we're good to go...
- all_walking_networks | ||
- all_walking_shield_texts | ||
- all_bus_networks | ||
- all_bus_shield_texts | ||
# drop all_bicycle_* until the max zoom: | ||
# https://github.com/tilezen/vector-datasource/pull/1707#discussion_r236522176 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
See my other comment in the other PR...
https://github.com/tilezen/vector-datasource/pull/1710/files#r236765704
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, though not sure if https://github.com/tilezen/vector-datasource/pull/1707/files#r236767971 applies to this PR or the other one since commits are in both PRs...
Please also make sure #1707 (comment) is tracked in an issue somewhere. |
Tracking the |
Drop RCN from 12 down to 0, LCN from 15 down to 0, and any network on track paths from 12 down to zero. These properties inhibit some degree of merging at mid zooms.
Context.
Connects to #1331. (Note that this is one part of many - submitting a PR for each to avoid it becoming unreadably large.)