feat: run variant index step

[ireneisdoomed]

This PR includes minor fixes to generate a variant index that is based on the variant annotation.

• Change to the path of the GWAS Catalog Study locus with the aim of having all study locus records under the same parent directory.
• Refactor of unique_variants_in_locus: simpler, more efficient way of extracting all variants of interest in a StudyLocus dataset.
  • This set will contain all variantIds of the lead variant in the locus and the other relevant variants in the locus.
• Refactor of the logic in the step: now filtering variant annotation happens inside VariantIndex.from_variant_annotation, so that the step is lighter.
• Downgrade of Hail to 0.2.122 to fix the bug reported in #3088

Important notes

• This variant index is based on the most up to date and latest StudyLocus. Right now these are based on:
  • gs://genetics_etl_python_playground/output/python_etl/parquet/XX.XX/study_locus/catalog_study_locus_updated_locus/: the latest PICSed version of GWASCatalog top hits
  • gs://genetics_etl_python_playground/output/python_etl/parquet/XX.XX/study_locus/finngen_study_locus: the latest PICSed version of Finngen's summary statistics generated by @DSuveges
• Given these inputs, the current variant index contains: 754_306 variants
• Job runs in ~3min

[codecov-commenter]

Codecov Report

Merging #188 (2e812f5) into main (6389ff1) will increase coverage by 0.20%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##             main     #188      +/-   ##
==========================================
+ Coverage   90.74%   90.95%   +0.20%     
==========================================
  Files          67       67              
  Lines        1459     1459              
==========================================
+ Hits         1324     1327       +3     
+ Misses        135      132       -3     

Files	Coverage Δ
src/otg/dataset/study_locus.py	95.94% <100.00%> (-0.06%)	⬇️
src/otg/dataset/variant_index.py	100.00% <100.00%> (ø)
src/otg/common/session.py	75.00% <0.00%> (ø)
src/otg/variant_index.py	68.75% <0.00%> (+4.04%)	⬆️

... and 1 file with indirect coverage changes

[DSuveges]

All looks good. Things got much nicer.