This recipe for smetric is the fastest

[eriknw]

There's a big difference in performance of the three implementations in the code. Benchmarking made clear which is the fastest. Lesson reinforced: try not use use outer followed by reduction.

Also, we can probably remove the extra with_authority=False option in HITS (using the authority matrix is slower and uses more memory), but I'll run a few more benchmarks on this to make sure.

[codecov-commenter]

Codecov Report

Base: 72.66% // Head: 72.67% // Increases project coverage by +0.01% 🎉

Coverage data is based on head (cf16b85) compared to base (725a0b9).
Patch coverage: 100.00% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #38      +/-   ##
==========================================
+ Coverage   72.66%   72.67%   +0.01%     
==========================================
  Files          69       69              
  Lines        2568     2569       +1     
  Branches      475      475              
==========================================
+ Hits         1866     1867       +1     
  Misses        531      531              
  Partials      171      171              

Impacted Files	Coverage Δ
graphblas_algorithms/algorithms/smetric.py	71.42% <100.00%> (+4.76%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.