Suspect Data Workflow Flags

[crashfrog]

Parent

#1 - MLST Multi-Scheme Typing Implementation

What to build

Workflow parameters for suspect data handling. If quality.json present, workflow can include/exclude suspect alleles/loci/profiles. Flag naming: positive semantics (--include-suspect-alleles) not double-negatives.

Default: include suspect data (no filtering). Users opt-in to filtering.

Acceptance criteria

• Workflow reads quality.json if present
• CLI flags: --include-suspect-alleles (default), --exclude-suspect-alleles
• Additional flags: --exclude-suspect-loci, --exclude-suspect-profiles
• Workflow filters allele database based on flags before MinHash/alignment
• Results note which alleles/loci were excluded (if any)
• Works when quality.json absent (no filtering)
• Tests verify filtering behavior at all three levels
• Documentation: flag semantics and defaults

Blocked by

#6 - HITL - Quality Analysis Review
#18 - Complete MLST Workflow

[crashfrog]

Implementation PR #48 closed due to backward compatibility issues with existing workflow. Filtering approach needs rework - recommend creating separate workflow or making filter task truly conditional in WDL.

[crashfrog]

Completed in PR #73 - implemented quality.json filtering with three hierarchical levels (alleles/loci/profiles), added CLI flags, and integrated filtering into all typing workflows.