Add parse_variables_sheet function with validation

[yulric]

Summary

This PR implements the parse_variables_sheet() function that validates variables sheets to ensure derived variables follow proper dependency rules.

Key Features

• Validation Logic: Ensures derived variables do not reference database columns directly (e.g., DerivedVar::[cchs::age] is invalid)
• Allowed Patterns:
  • Derived variables can use non-derived variables: DerivedVar::[age, sex] ✓
  • Derived variables can use other derived variables: DerivedVar::[var1, var2] ✓
  • Non-derived variables can use database columns: cchs::age or [age] ✓
• Return Values:
  • Success: Returns input with S3 class variables_sheet assigned
  • Failure: Returns structured error list with success = FALSE and detailed error information
• Error Handling: Collects all validation errors (not just the first one), with each error containing type, row number, and descriptive message

Implementation Details

New Files:

• R/parse-variables-sheet.R - Main function and helper functions
• tests/testthat/test-parse-variables-sheet.R - Comprehensive test suite

Helper Functions:

• .validate_variables_sheet_input() - Input validation (data frame, required columns)
• .validate_derived_variables() - Core validation logic for derived variables
• .extract_feeder_variables() - Extracts variables from DerivedVar::[...] syntax
• .is_database_column_reference() - Detects database::column pattern

Test Coverage

Comprehensive test suite with 12 test cases (6a through 6l):

• ✅ Non-derived variables validation
• ✅ Derived variables using valid dependencies
• ✅ Detection of invalid database column references
• ✅ Multiple error collection
• ✅ Empty data frames and DerivedVar syntax
• ✅ Input validation (missing columns, invalid types)
• ✅ Real-world data validation (pbc_variables.csv)

All 30 tests passing ✓

Test plan

• Run all tests with rig run -r 4.3.2 -e "devtools::test(filter = 'parse-variables-sheet')"
• Verify all 30 tests pass
• Update NAMESPACE via devtools::document()
• Verify function works with real-world data (pbc_variables.csv)

🤖 Generated with Claude Code