With the upcoming v1.0 release, we should start publishing hypertools to conda-forge in addition to PyPI. It shouldn’t take much extra work (the build/publish workflow is pretty straightforward) and would make installing/managing dependencies a lot easier for users in a lot of environments. Most popular data science-related packages are available via conda-forge, and conda & pip don't always interact cleanly when resolving dependencies, so pip-installing hypertools into a primarily-conda environment can sometimes lead to issues (especially with the stricter dependency constraints in the 1.0 update).
With the upcoming v1.0 release, we should start publishing hypertools to conda-forge in addition to PyPI. It shouldn’t take much extra work (the build/publish workflow is pretty straightforward) and would make installing/managing dependencies a lot easier for users in a lot of environments. Most popular data science-related packages are available via conda-forge, and conda & pip don't always interact cleanly when resolving dependencies, so pip-installing hypertools into a primarily-conda environment can sometimes lead to issues (especially with the stricter dependency constraints in the 1.0 update).