Age | Commit message (Collapse) | Author | |
---|---|---|---|
2020-12-10 | Merge pull request #409 from mglisse/fardist | Marc Glisse | |
Pass a distance function instead of a kernel to choose_n_farthest_points | |||
2020-12-10 | Merge pull request #408 from mglisse/subsample | Marc Glisse | |
Clean-ups in subsampling | |||
2020-12-10 | Fix memory leak | ROUVREAU Vincent | |
2020-11-30 | Fix pykeops 1.4.2 where version is very verbose on 2 lines | ROUVREAU Vincent | |
2020-11-26 | handle old boost | Marc Glisse | |
2020-11-22 | Handle squared radius | Marc Glisse | |
Make it work without a breaking change, we can always make a change later in a separate PR. | |||
2020-11-18 | merge master and resolve conflicts | ROUVREAU Vincent | |
2020-11-13 | Merge pull request #405 from mglisse/botneg | Marc Glisse | |
Bottleneck distance for one point or negative coordinates | |||
2020-11-13 | Test with negative coordinates | Marc Glisse | |
2020-11-10 | Revert change to the example | Marc Glisse | |
It would require updating the figure. | |||
2020-11-09 | Don't test the equality case in sparsify_point_set. sqrt. | Marc Glisse | |
2020-11-09 | Merge remote-tracking branch 'origin/master' into fardist | Marc Glisse | |
and remove example installation | |||
2020-11-09 | static_assert message | Marc Glisse | |
2020-11-08 | Test BettiCurve with infinite value | Marc Glisse | |
+ black reformatting | |||
2020-11-07 | Swap clip with int conversion | Marc Glisse | |
This allows correct handling of infinite values. | |||
2020-11-05 | More numpy in BettiCurve | Marc Glisse | |
suggested by @raphaeltinarrage | |||
2020-11-04 | Merge branch 'master' into weighted_alpha_complex_dD | ROUVREAU Vincent | |
2020-11-04 | Merge pull request #421 from mglisse/pythonfirst | Vincent Rouvreau | |
Build python first | |||
2020-11-04 | import matplotlib only when necessary in examples, in function of args. ↵ | ROUVREAU Vincent | |
Tests are without matplotlib, so no need to test if present | |||
2020-11-04 | Merge pull request #418 from VincentRouvreau/do_not_install_compiled_examples | Vincent Rouvreau | |
Do not install compiled examples | |||
2020-11-04 | Revert "Fix #416 by adding full path to introduction.rst - maybe strange ↵ | ROUVREAU Vincent | |
with conda or pip installation" This reverts commit 2d4c69d698df7e40709ec0e38463b4cf30ee388d. | |||
2020-11-04 | Build python first | Marc Glisse | |
This subdirectory is built sequentially and takes a very long time. Starting it first, in a parallel build, it is still the last to end, but since it started earlier it also ends earlier. Usually I build the plugin directly with python3 setup.py build_ext --inplace -j20 But I don't think there is a natural way to do that from the makefile (except for the unlimited -j), since I doubt setuptools knows how to cooperate with make, ninja, etc to share a pool of workers. | |||
2020-11-03 | Fix #415 by removing all install cmake targets from examples | ROUVREAU Vincent | |
2020-11-03 | Fix #416 by adding full path to introduction.rst - maybe strange with conda ↵ | ROUVREAU Vincent | |
or pip installation | |||
2020-11-03 | Merge branch 'master' into weighted_alpha_complex_dD | ROUVREAU Vincent | |
2020-11-03 | code review: use Marc's trick to test relative error | ROUVREAU Vincent | |
2020-11-03 | code review: use relative error | ROUVREAU Vincent | |
2020-11-03 | doc review: explain briefly test aims | ROUVREAU Vincent | |
2020-11-03 | Add comment in code for get_boundaries if not interested in filtration values | ROUVREAU Vincent | |
2020-11-02 | Merge master and resolve conflicts | ROUVREAU Vincent | |
2020-11-01 | The example does not need CGAL/eigen anymore. | Marc Glisse | |
I had done a search and replace without checking. I don't understand why there are "install" directives for examples... | |||
2020-11-01 | Doc tweaks. | Marc Glisse | |
2020-11-01 | Document that only double is supported. | Marc Glisse | |
2020-10-31 | long line | Marc Glisse | |
2020-10-31 | s/kernel/distance/ for choose_n_farthest_points argument | Marc Glisse | |
2020-10-31 | Clean-ups in subsampling | Marc Glisse | |
2020-10-28 | emplace_back | Marc Glisse | |
2020-10-27 | Add a test about the filtration | ROUVREAU Vincent | |
2020-10-27 | Add a test for non visible points (hidden by weights) | ROUVREAU Vincent | |
2020-10-27 | zero weighted test is guaranteed only for exact version | ROUVREAU Vincent | |
2020-10-26 | code review: bad copy/paste that needs rephrase | ROUVREAU Vincent | |
2020-10-26 | for CGAL 5.2 pre-release to work | ROUVREAU Vincent | |
2020-10-22 | test with duplicated point | Marc Glisse | |
2020-10-22 | Simplify strange iterator use | Marc Glisse | |
the syntax with [] is already used a few lines above | |||
2020-10-22 | Document the change | Marc Glisse | |
It would be possible to emit the duplicate points instead of stopping, but the current implementation makes that inconvenient. | |||
2020-10-19 | Handle duplicated points | Marc Glisse | |
2020-10-17 | Handle diagrams with a single point | Marc Glisse | |
2020-10-17 | Don't put a (useless) point on the diagonal in the example | Marc Glisse | |
2020-10-17 | Replace diameter (actually max y) with max distance to diagonal | Marc Glisse | |
2020-10-16 | Merge pull request #396 from VincentRouvreau/tensorflow_wasserstein_test | Vincent Rouvreau | |
Added tests for wasserstein distance with tensorflow |