and do not output '"' for repeated values in sheet, this allows for
sorting/filtering afterwards.

mainly use data frames now.

- Remove get_rate_model_to_files.
- Replace get_AIC, get_gains_losses with parse_badirate_result.
- Partially refactor get_badirate_output_frame using dicts as
  main data structure.

Turns out v1.5 was v1.70b all along, so now we use that one again which
does support -bmodel GR and has the output expected by Lennart's code.

and include species even when they have no parent node.

and:

- Mention count_groups.tsv in doc.
- Use snake case TSV headers instead.

- Split run_badirate and add to main.nf
- Various refactors

- Rename phylo_tree -> tree
- Remove unused fam_to_genes
- Derive fam_to_species_to_count where it's needed, locally

Refactors:

- Do use a dummy size file to be able to use -print_ids, but use a tmp file
  instead.
- ...

Changes:

- Do not write tree_ids.newick. If needed downstream, we'll write out a more
  convenient format.

Various refactors.

Changes:

- Create count_groups.tsv instead a gene_families.txt. Basically the
  same file as before minus the genes column as that can already be
  derived from the Orthogroups.txt file.
- Name groups deterministically. This will avoid cache misses in Nextflow.
- Do not write 0.txt. This was a fake input file used downstream to get
  list of species. Instead, one can get the species from a species file
  or as CLI args.