Add mixed-upgrade HP adoption trajectory to rate design pipeline

[sherryzuo]

What

Wire NREL's MixedUpgradeScenario (buildstock-fetch v1.6.5) into the rate design pipeline so each CAIRO run can represent a specific year in a multi-year HP adoption trajectory rather than always using a single upgrade (all-baseline or all-HP).

Why

Today every run is either 100% baseline (upgrade=0) or 100% HP (upgrade=2). To model how rates perform as heat pump adoption grows over time — e.g. NYISO's NYCA electrification forecast from the Gold Book 2025 — we need runs where a fraction of buildings have adopted each upgrade in each calendar year.

How

1. Adoption config YAML (config/adoption/<name>.yaml): specifies per-upgrade adoption fractions for each year index and a year_labels list mapping indices to calendar years.
2. utils/pre/materialize_mixed_upgrade.py: assigns buildings to upgrades per year using MixedUpgradeScenario, writes year=<N>/metadata-sb.parquet (combined rows from each upgrade) and year=<N>/loads/ (symlinks to correct load parquets). No changes to run_scenario.py.
3. utils/pre/generate_adoption_scenario_yamls.py: generates scenarios_<utility>_adoption.yaml by copying selected runs from a base scenario YAML and rewriting path_resstock_metadata / path_resstock_loads per year.
4. Justfile recipes: materialize-adoption, generate-adoption-scenarios, run-adoption-scenario, run-adoption-all.
5. Tests in tests/pre/test_materialize_mixed_upgrade.py: building assignment fractions, metadata combination, symlink targets, scenario CSV, and validation error paths.

Prerequisites (separate tasks): download and process upgrades 01, 04, 05 through the data/resstock/ pipeline. Until available, the trajectory works with upgrade 02 only.

Deliverables

• Implemented materialize_mixed_upgrade.py and generate_adoption_scenario_yamls.py with passing tests
• NYCA adoption config with fractions derived from NYISO Gold Book 2025
• Justfile recipes wired end-to-end
• just check and just test pass

[linear]

RDP-174