Refactor Hausdorff distance to header-only API #538
Conversation
Co-authored-by: Jake Hemstad <[email protected]> Co-authored-by: Paul Taylor <[email protected]>
…ial into fea-header-only-haversine
harrism
requested review from
vyasr and
isVoid
and removed request for
thomcom and
zhangjianting
I discovered that this is because I have merged in individual commits from feature branches targetting branch-22.06 before the associated PRs were squash-merged into 22.06. Since this PR now targets 22.08, those commits still show up in the feature branch history. However, the changes are already squash-merged into 22.08, so reviewers only see the code changes from this PR. Even though the PR's history is not clean, when it is squash-merged into 22.08, it won't matter. Therefore I'm inclined to leave it alone. |
| * all threads have run to completion, all "maximums of the minumum distances" (aka, directed | ||
| * Hausdorff distances) reside in the output. | ||
| * | ||
| * @tparam T type of coordinate, either float or double. |
There was a problem hiding this comment.
Oops, I approved too soon. The docs for template parameter here should be added.
|
@gpucibot merge |
|
@gpucibot merge |
We can use `rapids_cpm_libcudacxx` directly without this extra function. The only value it adds is setting the `LIBCUDACXX_INCLUDE_DIR` variable, which we don't use. See also rapidsai/cuspatial#538 (comment) Authors: - Vyas Ramasubramani (https://github.com/vyasr) Approvers: - Mark Harris (https://github.com/harrism) - Robert Maynard (https://github.com/robertmaynard) URL: #11138
Adds header-only API
cuspatial::directed_hausdorff_distanceand refactors the existing API on top of the new API, and updates tests.Part of https://github.com/rapidsai/cuspatial/milestone/1