pysal / segregation

Segregation Measurement, Inferential Statistics, and Decomposition Analysis

https://pysal.org/segregation/

BSD 3-Clause "New" or "Revised" License

112 stars 26 forks source link

remove mamba reference #166

Closed jGaboardi closed 3 years ago

jGaboardi commented 3 years ago

This PR removes the mamba reference in unittests.yml, which appears to be screwing up the channel priority settings resulting in a bad rtree install on Windows.

See pysal/spaghetti#610 and conda-forge/rtree-feedstock#31

codecov-commenter commented 3 years ago

Codecov Report

Merging #166 (974624a) into master (2fbc439) will not change coverage. The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #166   +/-   ##
=======================================
  Coverage   85.79%   85.79%           
=======================================
  Files          66       66           
  Lines        2788     2788           
=======================================
  Hits         2392     2392           
  Misses        396      396

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 2fbc439...974624a. Read the comment docs.

martinfleis commented 3 years ago

Does this essentially disable mamba? It would be great to figure out if there's an actual bug in mamba or some other way how to make it work (with mamba).

jGaboardi commented 3 years ago

Does this essentially disable mamba? It would be great to figure out if there's an actual bug in mamba or some other way how to make it work (with mamba).

Yep, the only thing it does is remove the mamba-version: '*' keyword from conda-incubator/setup-miniconda. I think the most likely reason for the weird behavior is that we were using it incorrectly. Maybe we can get @goanpeca or @bollwyvl's opinion on this?

jGaboardi commented 3 years ago

Maybe I am wrong. See here.

jGaboardi commented 3 years ago

I now see that the values acceptable in conda for channel-priority have changed since we first started using it and we had not updated. Now channel-priority: true defaults to channel-priority: flexible. We should be using channel-priority: strict.

knaaptime commented 3 years ago

thanks for this @jGaboardi. I'll keep an eye on what you're doing in spaghetti and try and incorporate those changes in the refactor branch also

jGaboardi commented 3 years ago

thanks for this @jGaboardi. I'll keep an eye on what you're doing in spaghetti and try and incorporate those changes in the refactor branch also

@martinfleis and I had a major success with micromamba over the weekend that reduced total testing duration by more than half. Have a look and let me know if you'd like a PR for segregation.

knaaptime commented 3 years ago

awesome. if you're up for a PR against my 2.0 branch, that would be killer

knaaptime commented 3 years ago

(then we could close this one, actually)