Enhance entity mapping with flexible aggregation methods and custom values

[nwoodruff-co]

Summary

This PR enhances the entity mapping system to support custom value arrays and multiple aggregation methods, making it more flexible for complex analysis workflows.

Changes

Core enhancements:

• Added values parameter to map custom value arrays instead of requiring existing columns
• Extended how parameter with new aggregation methods:
  • Person → Group: sum (default), first
  • Group → Person: project (default), divide
  • Group → Group: sum, first, project, divide

Refactoring:

• Created base YearData class in src/policyengine/core/dataset.py
• UKYearData and USYearData now inherit from base class
• Eliminated duplicate map_to_entity implementations

Documentation:

• Added comprehensive entity mapping section to docs/core-concepts.md
• Added examples to UK and US model documentation
• Documented all aggregation methods with practical use cases

Testing

All existing tests pass, confirming backward compatibility:

• tests/test_entity_mapping.py: 9 tests passing
• tests/test_us_entity_mapping.py: 5 tests passing

Use cases

1. Custom value mapping: Map imputed or calculated values that don't exist as columns
2. Equal splitting: Divide household-level values equally among persons
3. Representative values: Take first person's value as group representative
4. Flexible aggregation: Control how values combine across entity levels

Example usage

# Map custom per-person values to household level
household_totals = dataset.data.map_to_entity(
    source_entity="person",
    target_entity="household",
    values=custom_values_array,
    how="sum"
)

# Split household rent equally among persons
person_rent_share = dataset.data.map_to_entity(
    source_entity="household",
    target_entity="person",
    columns=["rent"],
    how="divide"
)

Closes #183