update criterion settings

[oojo12]

This PR updates the criterion settings for our benched algorithms. Over the course of a day or a few days I will update this description with the timing stats from before and after changes for each algorithm. Unless, it is okay to just do it for a few then I'll do that instead

linfa-ica

Old - 1min 39secs
New - 3min 38secs

linfa-pls

Old - 4mins 16secs
New - 9mins 42secs

linfa-linear

Old - 1min 29secs
New - 3min 37secs

linfa-nn

Old - 3mins 43secs
New - 8mins 57secs

linfa-clustering

k-means:

Old - 4min 6secs
New - 7mins 34secs

gaussian_mixture

Old - 51secs
New - 2mins 5secs

dbscan

Old - 1min 14secs
New - 2mins 50secs

appx_db_scan

Old - 47secs
New - 1min 27secs

linfa-ftrl

Old - 1min 43secs
New - 3min 57secs

linfa-trees

Old - 55secs
New - 2min 38secs

[YuhanLiin]

Should we refactor the benchmark configs to a common location? Doing it will be a bit involved, since we'll have to add a new benchmark feature to linfa and put the benchmark configs under that feature.

[oojo12]

Yeah an importable default would be nice for others and might reduce barriers to contributing benchmarks.

[YuhanLiin]

Also can you remove the (1000, 5) param for linfa-pls? It takes super long currently.

[codecov-commenter]

Codecov Report

Base: 38.59% // Head: 38.44% // Decreases project coverage by -0.15% ⚠️

Coverage data is based on head (cc0b235) compared to base (11c4d5b).
Patch coverage: 0.00% of modified lines in pull request are covered.

❗ Current head cc0b235 differs from pull request most recent head a4de64e. Consider uploading reports for the commit a4de64e to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #276      +/-   ##
==========================================
- Coverage   38.59%   38.44%   -0.16%     
==========================================
  Files          93       95       +2     
  Lines        6223     6227       +4     
==========================================
- Hits         2402     2394       -8     
- Misses       3821     3833      +12     

Impacted Files	Coverage Δ
algorithms/linfa-nn/tests/nn.rs	75.60% <0.00%> (ø)
src/benchmarks/mod.rs	0.00% <0.00%> (ø)
src/lib.rs	0.00% <0.00%> (ø)
...rithms/linfa-trees/src/decision_trees/algorithm.rs	37.94% <0.00%> (-1.79%)	⬇️
src/composing/platt_scaling.rs	28.69% <0.00%> (-1.74%)	⬇️
src/dataset/impl_dataset.rs	42.61% <0.00%> (-1.04%)	⬇️
src/metrics_classification.rs	37.97% <0.00%> (+0.63%)	⬆️

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.

[oojo12]

When I find more time to work on fixing this I will do the above and add a configuration for profiling with pprof. In the spirit of helping out with the investigation related issues.

[oojo12]

@YuhanLiin how do you add a feature? I tried unsuccessfully and couldn't find good docs on the subject.

[YuhanLiin]

Docs are here. You need to add the feature to the [features] section of Cargo.toml and introduce Criterion as an optional dependency, then have the new feature include Critierion.

[oojo12]

Got the feature in but now there is a peculiar failure for the ubuntu builds regard the tests for linfa-nn. Not sure what's prompting this error as I haven't adjusted the source code. This failure appears on this branch but is not evident when running the same check from wsl on the master branch

cargo check --workspace --all-targets

[oojo12]

assert_eq!(out, Vec::<ArrayBase<ViewRepr<&f64>, Dim<[usize; 1]>>>::new()) this line fixes the error and seems to work from my local testing. How do you feel about me updating the test?

Actually, pushing if it isn't desirable, we can always revert.

[YuhanLiin]

assert_eq!(out, Vec::<ArrayBase<ViewRepr<&f64>, Dim<[usize; 1]>>>::new()) this line fixes the error and seems to work from my local testing. How do you feel about me updating the test?

Actually, pushing if it isn't desirable, we can always revert.

Try ArrayView1<&f64> instead of ArrayBase<_>

[oojo12]

ArrayView1<&f64>

no dice with Vec::<ArrayView1<&f64>, Dim<[usize; 1]>>::new() instead of Vec::<ArrayBase<ViewRepr<&f64>, Dim<[usize; 1]>>>::new()

[YuhanLiin]

Try Vec::<ArrayView1<&f64>>

[oojo12]

Try Vec::<ArrayView1<&f64>>

no dice