New outcome space and encoding: BubbleSortSwaps and BubbleSortSwapsEncoding

[kahaaga]

As discussed in #384.

• Added BubbleSortSwaps outcome space, which when used to encode data, first embeds the data, then counts for each state vector the number of swaps necessary for the bubble sort algorithm to sort it.
• Added BubbleSortSwapsEncoding, which takes care of the actual encoding.
• Implemented codify, which uses BubbleSortSwapsEncoding and is used internally by BubbleSortSwaps.
• Added BubbleEntropy <: ComplexityEstimator, which implements both complexity and complexity_normalized, and computes a (normalized) difference between entropies for dimensions m + 1 and m. In the original paper, Renyi entropy of order q = 2 was used, but thanks to information_maximum, we can use any InformationMeasure.
• Tests.
• Documentation entries.
• Example showing how to compute the "bubble entropy".

[codecov]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (c191814) 88.80% compared to head (eea2c66) 89.00%.

Files	Patch %	Lines
...ding_implementations/bubble_sort_swaps_encoding.jl	96.15%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #390      +/-   ##
==========================================
+ Coverage   88.80%   89.00%   +0.20%     
==========================================
  Files          82       85       +3     
  Lines        2393     2447      +54     
==========================================
+ Hits         2125     2178      +53     
- Misses        268      269       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Datseris]

LGTM, only minor comments