harold
harold
Thanks, this project hasn't been updated in a while and seems to no longer work with newer versions of python You may want to try SciPy's solution- https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.linear_sum_assignment.html
(I'll try to fix this at some point, but not sure when. Meanwhile I'll add a link to SciPy in the README)
Although as discussed in #4 theirs is maybe slower, so keep this alive until theirs has been improved? (And help with that?)
#9 also claims that this is faster than numpy, so I'm changing this issue to be about trying to figure out why and merge the faster version into numpy 🎉
- [ ] benchmark this versus numpy to see if it actually is faster - [ ] speed up numpy
Hi @charnley! Sadly there is zero progress on this. Any interest in taking a look?! Either way, just bringing it back into my consciousness is helpful too, thanks! I'll add...
Cool if you get a chance that would be great! And yes good point, making sure the pip version is working should be highest priority. Issue #12 is about that.
@berhane that's awesome, thanks! Looks like it would be even better to bake a LAPJV implementation into SciPy.
@vmarkovtsev Awesome, sounds perfect! I think the very first step is that someone should figure out how people contribute to SciPy, and then maybe open a placeholder issue (if that's...
@pnagasanthosh did you get this working? (Sorry for the slow response, I've been traveling.)