Refactor and Add methods in Grid Partitioning

[Ali-Tehrani]

This pull request is regarding the grid partitioning methods. These partition space in bins and points are selected based on different bins at a time. There are two types of partitioning schemes, equisized which partitions such that bins have the same size, and equifrequent which partitions space so that each bin has approximately the same number of points. In each scheme, one can do independent and dependent, where in the latter, it partition in each dimension depends on the previous dimensions. There are examples explained in further detail in reference.

The following changes were made:

• The equifrequent independent and dependent were implemented. Before it raised a NonImplementError.
• The equisized methods were vectorized and refactored. One method was added partition_points_to_bins_equisized, which is used in both dependent and independent methods.
• PCA projection were removed as an option and should be implemented in a jupyter notebook example.

The following tests were made:

• Both independent and dependent should have the same bins and partitioning on a uniform grid. This the following test test_bins_from_both_methods_dependent_same_as_independent_on_uniform_grid.
• Cooked up examples in two dimensions were created.

In accordance to issue #156, I've added a simple fix until the pull-request #138 is merged.

[codecov[bot]]

Codecov Report

Merging #162 (16c26ee) into main (4eb670d) will increase coverage by 9.46%. The diff coverage is 98.80%.

:exclamation: Current head 16c26ee differs from pull request most recent head 17b8aa8. Consider uploading reports for the commit 17b8aa8 to get more accurate results

@@            Coverage Diff             @@
##             main     #162      +/-   ##
==========================================
+ Coverage   85.57%   95.03%   +9.46%     
==========================================
  Files          13       13              
  Lines        1026     1028       +2     
==========================================
+ Hits          878      977      +99     
+ Misses        148       51      -97     

Files	Coverage Δ
DiverseSelector/diversity.py	99.28% <100.00%> (ø)
DiverseSelector/methods/partition.py	97.85% <98.76%> (+42.00%)	:arrow_up:

[FanwangM]

Thanks for the improvement. @Ali-Tehrani

I have rebased the codes in order to fix the failing test. A patch/update for the calculation of diversity is also added. Sorry for force push. I will do the code review later.