Uh oh!
There was an error while loading. Please reload this page.
Fix point bounding box query bug (occurring for non-trivial transformations within non-matching 3D axes) - #1187
Conversation
f8c57d5 to
1600425CompareCodecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@## main #1187 +/- ##
==========================================
- Coverage 92.40% 91.85% -0.56%
==========================================
Files 52 53 +1 Lines 7872 7928 +56 ==========================================
+ Hits 7274 7282 +8 - Misses 598 646 +48
🚀 New features to boost your workflow:
|
For reviewing:
|
7021386 to
98c7d85CompareUh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
98c7d85 to
c203796Comparec203796 to
a8c1d7aCompareLucaMarconato
commented
Aug 25, 2026
(@ajkswamy did a rebase to meet the requirements of the stacked PR machinery since a new PR got merged into main while this was open). |
ported from transfo: supporting z and c; improved order of returned transformations also: split into simple/full; changed return type to tuples; supporting permutation of input/output axes for the transformation
a8c1d7a to
f4cfb43Compareajkswamy
commented
Aug 26, 2026
@LucaMarconato I've added some commits to the PR, some before and some after your commits (1d775ad and 89ca4cd). The commits before were needed for fixing the benchmarking setup and adding a new benchmark. This helps us compare the benchmark results before and after your commits. The commits after were minor refactors/cleanups, feel free to revert them if you see fit. Your changes for fixing the issue look good to me, in the context of the current issue. We could do other improvements like better typing and code refactoring for readability and maintainability, but we can do that later as needed. I ran benchmarks on my Windows Laptop (i7-11800H, 8 cores, 2.3 Ghz, 16GB RAM) to compare the execution times of the newly added benchmark before and after your changes. The results showed a significant increase of about 35%. It would be good to add another benchmark with larger points data to help get a sense of how the execution time might scale with larger data. Here are the benchmarking commands and results: Comparing benchmarks of `bbf9c0` with `89ca4c` (i.e., just your commits)Comparing benchmarks of `bbf9c0` with PR `HEAD` (i.e. including my commits after yours) |
ajkswamy
left a comment
There was a problem hiding this comment.
LGTM, I've added some commits, please see my last comment on the PR
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
LucaMarconato
commented
Aug 27, 2026
Thanks for the review for extending the I have ran new benchmarks with larger data sizes and the regression in performance is only occurring for small-medium sample sizes. With large data (most interesting use case) the performance is not significantly different than before this PR. If in the future we observe the need to improve performance also for small sample sizes, we can open a follow up PR centered on performance. Ready to merge! 🚀 |
Uh oh!
There was an error while loading. Please reload this page.
Closes#1175