Implement LineString simplification #55
Conversation
This trait currently uses the Ramer–Douglas–Peucker algorithm to simplify a LineString, while preserving its overall shape
| assert_eq!(dist, 0.7071067811865475); | ||
| } | ||
| #[test] | ||
| fn rdp_test() { |
There was a problem hiding this comment.
Can you also add a test doing simplify on an empty LineString?
There was a problem hiding this comment.
Or should we expect all LineStrings to have at least two points? If so, we should make it explicit in the doc comment and add an assert.
There was a problem hiding this comment.
Hmm. A 2-point LineString can't be simplified anyway, and both it and an empty LineString are valid, so it makes sense to me to just return them both unmodified.
Return it unmodified
These can't be simplified, so the output should be identical to the input
|
Everything here looks good to me, is this still a WIP? |
|
No, I'm pretty happy with it. How do you feel about the thoroughness of the tests, though? I feel like some tests for pathological input (tiny epsilon, gigantic LineString) would probably be a good idea at some point, although in the latter case that should probably happen further up the chain. |
|
Sounds great, thanks for this! 🎊 |
|
\o/ |
|
:D One of these I want to get some sort of basic Rust GUI for geo data (maybe written in conrod) so that we can visualize all these rust-geo algorithms and operations. |
This trait currently uses the Ramer–Douglas–Peucker algorithm to
simplify a LineString (see #54), while preserving its overall shape.
This is WIP, for two reasons:
rdp()currently usesunwrap(), because thefirst()andlast()methods on slices returnOption. It's not immediately clear to me how these could fail in this context, but all input is welcome.point_line_distance()andrdp()are implemented as helper functions, and called by theSimplifytrait. This is mainly becauserdp()is recursive and uses sub-slices of its initial input, which AFAICS can't be easily obtained from aLineString.