Optimize distance calculations

[JCGoran]

Fixes #202

Describe your changes

• use numpy.linalg.norm instead of scipy.spatial.distance (see this SO comment and this SO answer for performance benchmarks)
• cache all Circle instances when we create an instance of Rings, and compute all separations via numpy

I've benchmarked the above 2 on the bullseye and rings shapes, and the results are nice:

# bullseye on main
218292773 function calls (215844061 primitive calls) in 85.438 seconds
# bullseye w/ this PR
211510859 function calls (209069591 primitive calls) in 82.753 seconds
# rings on main
224466386 function calls (222017681 primitive calls) in 86.428 seconds
# rings w/ this PR
210318700 function calls (207877434 primitive calls) in 82.098 seconds

While the relative performance improvement appears marginal, it only appears that way due to Pandas-related operations being the bottleneck, and for bullseye we get 6.7M less function calls, while for rings we get 14.1M less function calls in total.

Note that no changes were made to the public API.

Checklist

• [x] Test cases have been modified/added to cover any code changes.
• [x] Docstrings have been modified/created for any code changes.
• [x] All linting and formatting checks pass (see the contributing guidelines for more information).

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 98.41%. Comparing base (51b8d26) to head (244db0a). Report is 9 commits behind head on main.

Additional details and impacted files

[](https://app.codecov.io/gh/stefmolin/data-morph/pull/203?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Stefanie+Molin) ```diff @@ Coverage Diff @@ ## main #203 +/- ## ========================================== - Coverage 98.42% 98.41% -0.01% ========================================== Files 43 43 Lines 1775 1772 -3 Branches 358 357 -1 ========================================== - Hits 1747 1744 -3 Misses 25 25 Partials 3 3 ``` | [Files](https://app.codecov.io/gh/stefmolin/data-morph/pull/203?dropdown=coverage&src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Stefanie+Molin) | Coverage Δ | | |---|---|---| | [src/data\_morph/shapes/bases/line\_collection.py](https://app.codecov.io/gh/stefmolin/data-morph/pull/203?src=pr&el=tree&filepath=src%2Fdata_morph%2Fshapes%2Fbases%2Fline_collection.py&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Stefanie+Molin#diff-c3JjL2RhdGFfbW9ycGgvc2hhcGVzL2Jhc2VzL2xpbmVfY29sbGVjdGlvbi5weQ==) | `92.68% <100.00%> (+0.37%)` | :arrow_up: | | [src/data\_morph/shapes/bases/shape.py](https://app.codecov.io/gh/stefmolin/data-morph/pull/203?src=pr&el=tree&filepath=src%2Fdata_morph%2Fshapes%2Fbases%2Fshape.py&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Stefanie+Molin#diff-c3JjL2RhdGFfbW9ycGgvc2hhcGVzL2Jhc2VzL3NoYXBlLnB5) | `100.00% <100.00%> (ø)` | | | [src/data\_morph/shapes/circles.py](https://app.codecov.io/gh/stefmolin/data-morph/pull/203?src=pr&el=tree&filepath=src%2Fdata_morph%2Fshapes%2Fcircles.py&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Stefanie+Molin#diff-c3JjL2RhdGFfbW9ycGgvc2hhcGVzL2NpcmNsZXMucHk=) | `94.33% <100.00%> (+0.46%)` | :arrow_up: | ... and [4 files with indirect coverage changes](https://app.codecov.io/gh/stefmolin/data-morph/pull/203/indirect-changes?src=pr&el=tree-more&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=Stefanie+Molin)