Configuration¶
Canonical configs live under configs/.
| Area | Path |
|---|---|
| Genomes Analyzer | configs/genomes_analyzer/ |
| Genotype predictor | configs/predictors/genotype_based/ |
| Variant transformer | configs/predictors/variant_transformer/ |
| SNP ancestry | configs/predictors/snp_ancestry/ |
| Non-longevous dataset builder | configs/workflows/non_longevous_dataset/ |
| AlphaGenome workflow | configs/workflows/alphagenome/ |
| VCF to 23andMe converter | configs/converters/vcf_to_23andme/ |
| Legacy neural ancestry code | legacy/neural_ancestry_predictor_deprecated/configs/ |
Auditing Configs¶
Use the config audit before long runs:
--fail-on-active-legacy is the recommended gate for new experiments. It permits deprecated configs to exist but fails if an active config still depends on legacy data paths.
Overrides¶
Several CLI commands accept path overrides such as:
--dataset-dir--processed-cache-dir--results-dir--processed-dir--consensus-dataset-dir
Overrides are applied only when explicitly passed. This keeps config files reproducible while allowing smoke tests and temporary runs.