Add bulk rule loading from CSV and JSON files

[ColtonPayne]

Summary

Adds bulk rule loading from CSV and JSON files, following the same pattern established in PR #100 for bulk fact loading. Also updates the existing text file loader with a raise_errors parameter for consistency (set to False by default for the old text loader only to avoid breaking changes)

Resolves #117

Three-tier rule loading

Each file format supports a different level of expressiveness:

Feature	Text (.txt)	CSV (.csv)	JSON (.json)
rule_text	per rule	per rule	per rule
name	auto-generated	per rule	per rule
infer_edges	uniform (function param)	per rule	per rule
set_static	False (default)	per rule	per rule
custom_thresholds	not supported	not supported	per rule
weights	not supported	not supported	per rule
raise_errors default	False	True	True

• Text (add_rules_from_file): Simplest format. One rule per line. infer_edges is applied uniformly to all rules. raise_errors defaults to False for backwards compatibility.
• CSV (add_rule_from_csv): Per-rule control over infer_edges and set_static. Supports flexible boolean formats (True/true/1/yes). Auto-detects headers.
• JSON (add_rule_from_json): Full expressiveness. Supports all Rule parameters including custom_thresholds (as serialized Threshold objects) and weights per rule.

Other changes

• Adds duplicate rule name detection with configurable error/warning behavior
• Adds shared _parse_bool_param() and _parse_and_validate_rule_params() helpers
• Includes comprehensive test coverage (32 new tests across 3 test classes)

Test plan

• All 32 new tests pass (12 TestAddRulesFromFile + 9 TestAddRuleFromCSV + 11 TestAddRuleFromJSON)
• Existing tests remain unaffected
• Tests cover: valid loading, error handling, duplicate detection, edge cases (empty files, missing fields, invalid types), boolean format flexibility, threshold/weight parsing

🤖 Generated with Claude Code