openvax / topiary / 24228314598

89%

master: 90%

Build Type

Pull #103

github

Committed by iskandr

Commit Message

Expression DSL v2: scoped contexts, len, count, repr round-trips

Replace WT() wrapper with scope prefix system (wt., shuffled., self.)
for alternate peptide contexts. The scope is a reserved keyword that
modifies which DataFrame column a Field reads from.

Core changes:
- Add scope parameter to Field and KindAccessor
- Add Scope class with wt, shuffled module-level instances
- Add Len(Expr) for peptide length from precomputed column
- Add Count(Expr) for dynamic amino acid counting from peptide string
- Add __repr__ to all Expr subclasses enabling parse→repr→parse round-trips
- Remove WT class entirely (replaced by wt.Affinity.score syntax)

Parser:
- wt, shuffled, self are reserved context keywords
- Support wt.affinity.score, wt.len, wt.count('C'), count('KR'), len
- Fix _build_ranking_strategy to route dotted exprs through parse_expr
- Fix broken test_parse_expr_abs, add parse_expr to __all__
- Remove unused mhc_binding_predictor_from_args import
- Deduplicate transform name set in CLI args

Tests:
- 15 CLI integration tests (test_cli_ranking.py)
- 28 AST↔string equivalence tests
- 25 round-trip tests (parse→repr→parse→evaluate)
- 20 WT→scope, bracket+chain, multi-dot, underscore-qualified tests
- 30 len/count/scope Python API/parsing/round-trip tests
- Update all existing WT() tests to wt. prefix syntax

Pull Request Pull Request #103: Expression DSL v2: scoped contexts, len, count, repr round-trips

Coverage Stats

390 of 423 new or added lines in 3 files covered. (92.2%)

26 existing lines in 1 file now uncovered.

1786 of 2014 relevant lines covered (88.68%)

2.66 hits per line