0.5.0 — the building stock replaces the calibrated metamodel

Mapping the peak aggregation found three defects rather than a formula: it
divided by the peak efficiency twice, anchored 40 GW against the 2020 useful
heat rather than the 2020 peak load, and read the wrong efficiency row for
district heating. Corrected, the reference transition multiplies the
electric-heating peak by 1.27 instead of holding it flat: 39.45 -> 51.0 GW.