On a second though, I think to_mvt_unscaled is more accurate. raw here would be confusing.
There is no need to store bounds and do additional math ops on it. Can precompute a few things earlier on.
Default can be auto-derived
P.S. @pka I think its better to push PRs to individual's forks, and only keep project-maintained versions in the primary repo. This way there would be no experimental branches in the main repo, reducing confusion.
to_mvt_unscaled
is more accurate.raw
here would be confusing.P.S. @pka I think its better to push PRs to individual's forks, and only keep project-maintained versions in the primary repo. This way there would be no experimental branches in the main repo, reducing confusion.