Various updates to filtering

[MattWellie]

Fixes

• Closes #304

Proposed Changes

• gnomAD genomes + exomes MAF test is applied to Cat. 1 Dominant Variants during MOI test
  • they still have an exception permitting low read depths
  • they also have an exception for recessive MOI (comp-Het or Hom) - this can be removed as well, but the main focus here is tidying up erroneous monoallelic variants
• F2 & F5 are on the 💩 list (requires a phenotype panel to be included in analysis)
• Lenient MOI flag application has changed
  • If gene MOI is Autosomal 'Mono or Biallelic'
  • If variant is Cat 1
  • If variant is not de novo for this sample
  • If this variant only satisfied an 'Autosomal Dominant' MOI

This takes the "does not qualify as any Recessive or X-Linked MOI" exclusions in #304 and reverses them - "we're only flagging Cat.1 non-de Novo variants with ambiguous MOIs, where we only satisfied an Autosomal Dominant MOI"

Misc

• Includes a few QOL fixes to the 'forbidden gene' checker required for the retrospective analysis. Now PanelApp activities endpoint queries are O(panels) instead of O(panels * timepoints)
• DocString formatting changes - non functional, but a more consistent look

Checklist

• [x] Related Issue created
• [wip] Tests covering new change
• [x] Linting checks pass