Validation for stream and river direction

[Geimas5]

Because of an ongoing import of N50 data, we have in Norway a huge amount of lakes and ponds with the elevation specified. Because of this, it would be very useful to have a validation in JOSM for the direction of streams and rivers based on this elevation information.

This can, of course be made very complicated to cover every case, but even a simple validation of the cases:

• (ele:200)<----(ele:100)
• (ele:200)<----(node)<---(ele:100)
• <----(node)---->(node)<----

Would help a lot.

From experience, it would avoid a lot of false positives if 1-2 meters of "wrong direction" was allowed.

This means that while:

• (ele:200)<----(ele:100)

would result in an error:

• (ele:101)<----(ele:100)

would not.

It would also be natural to assume natural=coastline was at ele:0.

[simon04]

It might already be possible to do using Help/Validator/MapCSSTagChecker. Maybe we'd need a ​sibling selector `~`?

[Klumbumbus]

The nodes do have the ele tagged? Please link an example from the osm database or attach a sample osm file.

[Geimas5]

Example of case with two streams

[Geimas5]

Replying to Klumbumbus:

The nodes do have the ele tagged? Please link an example from the osm database or attach a sample osm file.

No, the nodes do not have an ele tag. Its just very common for the streams to be split into multiple connected ways.

[Klumbumbus]

I can't find an error in that example file.

[Geimas5]

Two connected ways, with error

[Geimas5]

Replying to Klumbumbus:

I can't find an error in that example file.

Sorry, I created it as an example of a stream with two parts. I have attached one where the streams are in the wrong direction.

[Klumbumbus]

Replying to Geimas5:

• (ele:200)<----(ele:100)
• (ele:200)<----(node)<---(ele:100)

JOSM has no internal elevation database, so it can't know if the direction is right.

• <----(node)---->(node)<----

For this case a warning should appear after #17533 is fixed.