Closed miles-dev closed 7 years ago
Current status:
south_sudan
).sourceCover: 'south_sudan',
to the tileReduce
object.south_sudan
).osm
and south_sudan
respectively).tiger2015
property to be osm
.sqlite3 south_sudan.mbtiles
in your terminal and then select * from metadata limit 15;
. Look for this line in the output json|{"vector_layers":[{"id":"osm","description":"","minzoom":12,"maxzoom":12,"fields":{}}]}
and use the value of the id
field.with a more relevant name (in my case
south_sudan`).npm install
node index.js <path_to_planet.mbtiles> <path_to_south_sudan.mbtiles>
node index.js > south_sudan_diff.geojson
@tcql - Inspecting the geojson I got back, I noticed that I'm getting results which doesn't look like roads to me. As an example, https://www.openstreetmap.org/edit?way=466049354#map=16/8.7761/27.4289. The large straight line marking the edge of the grasslands there is coming back in my diff. Given that isn't a road I wouldn't expect it to be part of my results. Does this look like an incorrectly mapped feature or something not being filtered out in this repo?
@miles-dev the diff.js
script only considers lines with highway
property and Tiger roads with FULLNAME
property. You may need to modify this line: https://github.com/mapbox/road-diff/blob/master/diff.js#L61
I'll close this since this is not an actual, actionable issue with the repository. It's a working example of using linematch
for road diffing in TileReduce, and tweaks when using other data sources are expected.
On a second thought, let's keep this open until we improve the readme and code comments to make it easier to customize for various use cases.
Created a new branch osm_to_osm_diff
which includes the updated README.md as well as changes to the index.js and diff.js to specifically target OSM vs OSM diffing. End goal here is for that to be 'walk up usable' with little to no code adjustment when diffing two OSM mbtiles files.
Question for @mourner and @aaronlidman - What're your thoughts on me merging my new osm_to_osm_diff into master? The differences as I'm tracking them:
@miles-dev I'm not sure about making OSM vs OSM the default implementation. I think OSM vs Tiger is a more common use case (comparing different datasets).
@mourner - Sounds good. I'll leave this as a separate branch for now then. I think I'm all set here with updating the readme and updating the code for osm-to-osm diffing. Closing out this ticket.
Working on a project to better understand how to use this repo. Using South Sudan as a sample use case for diff'ing.
Goals:
Resources:
/cc @tcql