Normalize taxon_by handling to match area_by/period_by in cohort grouping

[shauryam2807]

Fixes #808

Problem

_prep_samples_for_cohort_grouping() normalizes area_by → "area" and period_by → "period" by copying user-specified columns into standard internal columns. However, taxon_by is treated inconsistently — it keeps the original column name instead of being normalized to "taxon".

This forces _build_cohorts_from_sample_grouping() to accept taxon_by as a parameter and maintain two separate label-generation code paths (one for the default "taxon" column, one for custom columns).

Re: discussion in #694 (comment): #694 (comment)

Solution (Option A, per @jonbrenas)

Normalize taxon_by to a standard "taxon" column in _prep_samples_for_cohort_grouping(), consistent with area_by and period_by. This:

• Removes the need for taxon_by in _build_cohorts_from_sample_grouping()
• Simplifies label generation to a single code path
• Makes all three *_by parameters consistent in behavior

Changes

• malariagen_data/anoph/frq_base.py — Add taxon column normalization in _prep_samples_for_cohort_grouping(); remove taxon_by param from _build_cohorts_from_sample_grouping(); simplify label logic
• malariagen_data/anoph/snp_frq.py — Update groupby and remove taxon_by from _build_cohorts_from_sample_grouping() calls
• malariagen_data/anoph/cnv_frq.py — Same
• malariagen_data/anoph/hap_frq.py — Same
• tests/anoph/test_frq_base.py — Add tests for taxon normalization

Testing

• Default taxon_by="taxon" works unchanged (backward compat) ✅
• Custom taxon_by creates standard "taxon" column ✅
• Original custom column is preserved ✅
• Label generation works correctly ✅
• ruff check passes ✅

[shauryam2807]

Hello
Hi @jonbrenas,
When you get a chance, could you take a look at this PR?
Let me know if anything needs improvement—I’ll update it right away.

[jonbrenas]

The issue is that "taxon" is already a column. It is why it is treated differently.

[shauryam2807]

Ah @jonbrenas , I see! Because df_samples already has a taxon column with real data, overwriting it destroys that data. To fix #808, should we normalize it into a new column name like "cohort_taxon", or should we ditch the normalization approach altogether since taxon can't be treated exactly like area and period?

[jonbrenas]

Thanks @shauryam2807, I think using a new column name is the best solution.

[shauryam2807]

Hello
Hi @jonbrenas,
Thanks for the suggestion! I've updated the implementation to use new column names (cohort_taxon, cohort_area, cohort_period) inside _prep_samples_for_cohort_grouping, ensuring the original user metadata columns are never overwritten.

Changes in this update:

• _prep_samples_for_cohort_grouping: Now writes to cohort_taxon, cohort_area, and cohort_period instead of overwriting taxon, area, and period.
• _build_cohorts_from_sample_grouping: Renames these back to taxon, area, and period after aggregation, keeping the downstream API (datasets, plots, etc.) fully unchanged.
• Downstream Updates: Updated groupby() calls in snp_frq.py, hap_frq.py, and cnv_frq.py to use the new internal column names.
• Tests: Updated test_frq_base.py to match the new column names and behavior.
• Linting & Formatting:
  • Fixed duplicate constant definitions in af1.py that were causing the pre-commit linting step to fail.
  • Applied ruff format across all modified files to ensure CI passes.

On a separate note — as a GSoC 2026 applicant who has applied for the MalariaGEN project, I just wanted to reiterate how much I've been enjoying contributing to the codebase! While waiting for the final results, I plan to continue addressing issues and helping out wherever I can.

Please let me know if this looks good to merge or if there's anything else you'd like me to address here!