motif composition metric notebook conversion

[ssenan]

• Resolves add motif composition/trajectory validation metric #212
• Adds the main functions from the motif composition notebook into main dnadiffusion library in motif_composition.py
• Adds function called motif_composition_helper that scans an input dataframe for motifs. This utilizes similar methodology to extract_motifs, so these functions can likely be combined in the future
• Introduces are a function version of the previously used SEQ_EXTRACT class called seq_extract, so we may look to remove the class in a future PR (this depends on integration of the rest of our validation metrics).
• Tests are provided for all the new functions except for motif_composition_matrix. We need to investigate further the best way to capture the performance of these metrics / if they are working correctly. (possible need to store a copy of hg38.fa and test if gimme motifs can work in CI)
• Reuses some fixtures across multiple files, which can probably be consolidated to one test_helper.py file in the future

[codecov]

Codecov Report

Merging #222 (3aa8282) into main (c2648c2) will increase coverage by 1.98%.
The diff coverage is 59.20%.

@@            Coverage Diff             @@
##             main     #222      +/-   ##
==========================================
+ Coverage   35.12%   37.11%   +1.98%     
==========================================
  Files          21       23       +2     
  Lines        1358     1482     +124     
  Branches      180      202      +22     
==========================================
+ Hits          477      550      +73     
- Misses        874      924      +50     
- Partials        7        8       +1     

Files	Coverage Δ
src/dnadiffusion/utils/data_util.py	49.66% <84.61%> (+3.34%)	⬆️
tests/test_data_util.py	92.94% <85.00%> (-7.06%)	⬇️
tests/test_motif_composition.py	41.37% <41.37%> (ø)
src/dnadiffusion/metrics/motif_composition.py	39.53% <39.53%> (ø)

[cameronraysmith]

perfect!