Python rewrite of the Urb/Homemaker stack
Bruno Postle
d52cce6863
Global relabel of inside-room leaves via one optimal assignment over the full leaf set — the 9o5 per-class collapse generalised to N leaves ↔ M required rooms — as a one-shot finish-time polish on a committed layout. Assignable room_codes exclude any starting c/o/s to match the scorer's own partition (check_space_counts skips those; cr1/st1/st2 collide with the circulation/structure convention). Hard level constraint via a -1e12 forbid penalty. Adjacency handled as an iterated relaxation: geometry is fixed at finish time so each leaf's graph neighbours are fixed; warm-start from evolved labels, each pass a linear assignment over quality + an adjacency bonus (has_adjacency vs current labels), Jacobi to a fixpoint. Measured (level+adjacency): best evolved layout 15→14 fails, rougher ones 32→28 and 90→83; adjacency-on beats adjacency-off everywhere (off regresses the best layout +1). Substrate only — not wired into search or a CLI yet. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01M8566xAxTnwtJTkpXjYNZm |
||
|---|---|---|
| .beads | ||
| .claude | ||
| examples | ||
| experiments | ||
| src/homemaker_layout | ||
| tests | ||
| .gitignore | ||
| AGENTS.md | ||
| CLAUDE.md | ||
| DESIGN.md | ||
| pyproject.toml | ||
| README.md | ||
homemaker-layout
Programme-driven building-layout search over slicing trees. A clean-room Python successor to the Perl Urb project, intended to eventually be 100% Python.
Why a rewrite
Urb represents a building as a binary slicing tree where room sizes are derived top-down from division ratios. That makes room area an emergent property of every cut above it, which:
- gives the genome low locality (a cut near the root rescales every descendant),
- makes target room sizes nearly impossible to hit, so the gaussian size penalty dominates fitness, and
- defeats crossover (transplanted subtrees lose their proportions).
homemaker inverts this: leaves carry target dimensions from the programme and division ratios are solved bottom-up for a fixed topology. The evolutionary search then only explores topology + types + adjacency.
Phase plan
Solver experiment: port Urb's geometry, re-solve ratios from programme targets, score the result against the original via the Perl oracle.✓Native Python fitness (retire the Perl oracle).✓- Memetic search (current): canonical slicing genome + high-locality operators + Nelder-Mead inner loop.
Layout
src/homemaker_layout/dom.py— read/write Urb.domYAML into aNodetree.src/homemaker_layout/geometry.py— faithful port of Urb's top-down geometry.src/homemaker_layout/programme.py— parsepatterns.configspace requirements.src/homemaker_layout/solver.py— bottom-up ratio solve (scipy).src/homemaker_layout/fitness.py— native Python fitness evaluator.src/homemaker_layout/fitness_cmd.py—homemaker-fitnessCLI (drop-in forurb-fitness.pl).src/homemaker_layout/graph.py— leaf-adjacency graph for programme-driven checks.src/homemaker_layout/genome.py— topology genome: base-floor tree + per-storey deltas.src/homemaker_layout/operators.py— high-locality mutation and subtree crossover.src/homemaker_layout/innerloop.py— ratio optimisation inner loop (Nelder-Mead / CMA-ES).src/homemaker_layout/driver.py— memetic search outer loop.src/homemaker_layout/evolve.py—homemaker-evolveCLI entry point.src/homemaker_layout/oracle.py— legacy Perl shim, kept for cross-validation only.