Closed aneksteind closed 4 years ago
Yes looks like some kahypar
build related issue. I've just been using the pypi
(pip install) version of kahypar on linux. Can you see if kahypar's own tests run? If not might be a issue to raise over there.
I also installed using pypi
, I will try and separate the two and update with any progress to rule out building kahypar from source
@jcmgray I was able to get the example working on an Ubuntu VM, perhaps this is a macOS thing. I tried running the example from a fresh virtual environment on macOS and got the same error, this time without the kahypar source involved and only cotengra and quimb
Hmm, yes sorry I don't have access to any mac to test. Does this minimal python kahypar example in their readme run?
import os
import kahypar as kahypar
num_nodes = 7
num_nets = 4
hyperedge_indices = [0,2,6,9,12]
hyperedges = [0,2,0,1,3,4,3,4,6,2,5,6]
node_weights = [1,2,3,4,5,6,7]
edge_weights = [11,22,33,44]
k=2
hypergraph = kahypar.Hypergraph(num_nodes, num_nets, hyperedge_indices, hyperedges, k, edge_weights, node_weights)
context = kahypar.Context()
context.loadINIconfiguration("<path/to/config>/km1_kKaHyPar_dissertation.ini")
context.setK(k)
context.setEpsilon(0.03)
kahypar.partition(hypergraph, context)
(you would have to download one of the .ini
files to run this - cotengra keeps a copy of them)
@jcmgray even just the import causes the same difficulty, unfortunately
Seems to be related to the kahypar pybind dependency, not a cotengra issue: https://github.com/kahypar/kahypar/issues/58
Glad you were able to get it fixed- thanks for keeping this issue updated!
In an attempt to run the Sycamore depth 12 example, I've run into the following exception:
Helpful info: OS: macOS Catalina $PYTHONPATH: /Users/david/Library/Python/3.7/lib/python
I've tried changing my environment variables and editing the CMake of kahypar such that
CC=clang
,CXX=clang++
and, as a sanity check, toCC=gcc
,CXX=g++
-- taking care to uninstall and reinstall the kahypar python bindings with pip each time.Any insight would be much appreciated.