Crosswalk naming

[binh-dam-ibigroup]

Summary

This PR adds a CrosswalkNamer class to assign names to crosswalks, such as "crossing crosswalk over 10th Street" (see screenshot), so that they have better names than just "path" for turn-by-turn directions. As of writing, only crosswalks that have 3 or more nodes in OSM and a shared node with a street are covered. The PR also apply the crosswalk name to short sidewalk segments if they are the only adjacent way at either end of the crosswalk.

The CrosswalkNamer is copied over from from SidewalkNamer (both classes refactored for common logic), and a few supporting helper methods are added to OsmWay for determining footways and adjacent ways.

Screenshot:

Issue

Closes #6864

Things to discuss/resolve

• Should the PR cover crosswalks that are split into multiple ways (e.g. sidewalk-to-street, street-to-sidewalk)
  • Example: https://www.openstreetmap.org/way/931095546 and https://www.openstreetmap.org/way/1186278549
• i18n
• What is a turn lane? Below are some examples:
  • Atlanta, GA, USA area:
    • https://www.openstreetmap.org/way/631281805: unnamed one-lane, one-way tertiary link (I tested with lanes similar to this one)
    • https://www.openstreetmap.org/way/527649398: unnamed one-way motorway link with turn:lanes attribute
    • https://www.openstreetmap.org/way/96019593: unnamed one-way secondary link
    • Not a turn lane: https://www.openstreetmap.org/way/96020432: unnamed one-way motorway link with "negative" turn lane attribute.
    • Not a turn lane: https://www.openstreetmap.org/way/1167192380: unnamed one-way motorway link (not tested yet)
  • Portland, OR, USA area:
    • https://www.openstreetmap.org/way/204306312: unnamed, one-way primary link (no lane info)
    • Not a turn lane: https://www.openstreetmap.org/way/44762684: unnamed, one-way secondary link but with description
    • Not a turn lane: https://www.openstreetmap.org/way/344579996: unnamed, one-way motorway link, 3 lanes
  • Seattle, WA, USA:
    • https://www.openstreetmap.org/way/10738253: unnamed one-way primary link
    • Not a turn lane: https://www.openstreetmap.org/way/61014437: named one-way residential
    • Not considered a turn lane https://www.openstreetmap.org/way/130345931: unnamed, one-way motorway link

Unit tests

Unit tests are added for the process for finding and applying names to various types of crosswalks (over a named street, over a service road, over a turn lane).

Documentation

Added.

Bumping the serialization version id

Not necessary, I think

[codecov]

Codecov Report

❌ Patch coverage is 83.58209% with 33 lines in your changes missing coverage. Please review.
✅ Project coverage is 72.20%. Comparing base (9966fd4) to head (7d52252).

Files with missing lines	Patch %	Lines
...raph_builder/module/osm/naming/CrosswalkNamer.java	85.07%	4 Missing and 6 partials ⚠️
...lder/module/osm/naming/SidewalkCrosswalkNamer.java	0.00%	10 Missing ⚠️
...uilder/module/osm/naming/PreciseBufferFactory.java	73.33%	4 Missing ⚠️
...graph_builder/module/osm/naming/SidewalkNamer.java	87.50%	0 Missing and 3 partials ⚠️
...ain/java/org/opentripplanner/osm/model/OsmWay.java	86.66%	0 Missing and 2 partials ⚠️
...ilder/module/osm/naming/BufferedEdgeProcessor.java	96.87%	0 Missing and 1 partial ⚠️
...aph_builder/module/osm/naming/StreetEdgeIndex.java	90.00%	0 Missing and 1 partial ⚠️
...pplanner/graph_builder/services/osm/EdgeNamer.java	50.00%	1 Missing ⚠️
...ing/algorithm/mapping/StatesToWalkStepsMapper.java	93.33%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@              Coverage Diff              @@
##             dev-2.x    #6918      +/-   ##
=============================================
+ Coverage      72.18%   72.20%   +0.01%     
- Complexity     19838    19895      +57     
=============================================
  Files           2155     2161       +6     
  Lines          80048    80185     +137     
  Branches        8082     8104      +22     
=============================================
+ Hits           57786    57896     +110     
- Misses         19415    19430      +15     
- Partials        2847     2859      +12     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[leonardehrenfried]

We are getting there. I took the liberty to not use to many static methods: ibi-group@ee9d9bb

If you agree with this, can you cherry-pick it into this branch.

Also, there are conflicts now.

[binh-dam-ibigroup]

We are getting there. I took the liberty to not use to many static methods: ibi-group@ee9d9bb

Good idea!

[VillePihlava]

Good job refactoring and extracting stuff out of classes. Recently I looked at similar OsmEntity way logic related to levels and it was really confusing. Can there be a good reason that an EdgeNamer can have OsmEntity way as parameter or should they always be OsmWay way? I think this should be changed if possible.

[VillePihlava]

If this is really a way, we should use the OsmWay type. If it has to be an OsmEntity there might be a logic error somewhere because relations and nodes can also be given as parameter in theory.

[binh-dam-ibigroup]

I also personally prefer OsmWay and I also got confused because a lot of way/edge code is in OsmEntity. Does a StreetEdge always come from an OsmWay? Happy to discuss how to proceed in an upcoming call.

[VillePihlava]

I think another option is to maybe do it in a separate PR. Lets discuss in Thursday's meeting to decide how to proceed

[VillePihlava]

Can you check if all of the namer-related OsmEntity way uses can be changed to OsmWay way?