Input Validation for Rules

[ColtonPayne]

Summary

Closes #104

Adds comprehensive input validation to the rule parser (rule_parser.py) so that malformed rules, invalid bounds, and bad parameters produce clear, actionable error messages instead of cryptic runtime failures.

Changes

• Entry-point validation — rule text must be a non-empty string with exactly one <- separator; head and body must both be non-empty after splitting
• Predicate & variable name validation — names must match [a-zA-Z_][a-zA-Z0-9_]*; digit-leading names and special characters are rejected with specific messages
• Bound validation — intervals must contain exactly two finite numeric values in [0, 1] with lower <= upper; NaN, Inf, and out-of-range values are caught
• Negation with explicit bounds — ~pred(X):[l,u] now correctly computes [1-u, 1-l] for both head and body clauses
• :True / :False shorthand — body and head bounds accept boolean string literals as aliases for [1,1] and [0,0]
• Double negation detection — ~~ in head or body raises immediately
• forall() syntax validation — the inner expression must contain a predicate with variables
• Threshold validation — empty dicts, negative keys, and length mismatches raise ValueError
• Weight validation — must be numeric, finite, non-negative, and match clause count; plain lists are auto-converted to np.float64 arrays
• Error type cleanup — generic Exception raises replaced with TypeError / ValueError
• Refactored internals — extracted _split_body_into_clauses, _parse_head, _validate_predicate_name, and _validate_variable_name helpers
• Makefile typo fix — corrected don_disable_jit → dont_disable_jit in test suite target

Test coverage

Added 722-line test file (tests/unit/dont_disable_jit/test_rule_parser.py) with three test classes:

• TestValidRuleParsing — node/edge rules, temporal deltas, annotation functions, bounds, negation, :True/:False, forall, head functions, thresholds, weights, and a complex combined-parameter case
• TestInvalidRuleParsing — type errors, empty inputs, missing arrows, bad bounds, wrong-length thresholds/weights, NaN/Inf weights, malformed forall
• TestEdgeCasesAndBoundary — boundary bounds, large delta_t, all six comparison operators, single-char predicates, negation inversion arithmetic, NaN/Inf bounds, digit-leading names, double negation

🤖 Generated with Claude Code