benchmark_vs_urbevolve.py results (2026-06-13, budget=2000, URB_NO_OCCLUSION=1):
- Seeded designs: memetic beats urb-evolve 1.91× (c964435) and 1.63× (2f45907)
- Blank slate init.dom: memetic at 18 fails vs urb-evolve at 6 fails (topology
diversity gap from single-seed mutation chain vs random-population init)
Bug fixed: run_search.py was calling oracle.score on out.parent without
patterns.config present — causing the re-score to return near-zero instead of
the correct tracked fitness. Added shutil.copy to propagate patterns.config
alongside the output .dom before the standalone re-score.
Gate recorded in DESIGN.md §7. Closes homemaker-py-way.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
4-way comparison (NM / CMA-ES / compass / compass-ms) over 3 corpus files ×
3 seeds at budget 200, cold-start, URB_NO_OCCLUSION=1. CMA-ES wins on
batch-efficiency (18 oracle calls vs 200 for NM, 12x speedup on Perl startup
amortisation per §4.6) with acceptable quality (x1.41 @200 vs NM's x1.56).
Compass stalls on narrow-valley landscapes and introduces fail regressions.
NM flagged as Phase 3+ candidate once native fitness removes oracle call overhead.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Port Urb's programme-driven fitness leaf quality factors (perpendicular,
proportion, size, width, crinkliness, daylight, access), value rates,
and cost model (per-leaf area costs, interior/exterior wall edge costs,
boundary costs) to Python. Passes 0-mismatch parity against the Urb
oracle across all 35 corpus files (407 leaves, 2849 factors), using
URB_NO_OCCLUSION=1 simple crinkliness (illumination factor pinned to 1).
Key fixes: _dist must use math.sqrt not math.hypot (1-ULP difference
flips boundary overlap predicates); leaf-scope fail regex requires ^\d+/
prefix to exclude building-level failure messages.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
run_urbevolve took (seed, budget, pop, cell) but cells call
fn(seed, budget, cell, **kw) — every urb-evolve cell died on TypeError,
deferred silently by pool.map. mutate_swap lacked the empty-candidates
noop guard the other operators have, crashing on init.dom-style bare
plots. Regression test: every mutation survives an undivided tree.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
experiments/benchmark_vs_urbevolve.py (homemaker-py-way): 3 seed designs
(init from scratch, c964435 weak, 2f45907 strong) x 2000-eval runs, the
1000-eval tier read from each run's best-so-far log; urb-evolve gets two
population sizes (default 128 = ~16 generations at this budget, and 16 =
~130) and credit for its better one. Counts via the MAX_EVALS counter
patch in urb-evolve.pl (committed in the urb repo); both systems under
URB_NO_OCCLUSION=1; all finals re-scored through urb-fitness.pl as the
common deterministic yardstick.
run_search.py generalised to (budget, rng-seed, seed.dom, out.dom);
innerloop.optimise now handles 0-DOF topologies (an undivided plot like
init.dom scores once instead of crashing CMA).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two bugs fixed in boundary_id / leaf_graph:
1. 'bid not in "abcd"' used Python substring check, silently dropping the
root-division boundary (empty-string id). Fixed to frozenset membership.
2. Upper-storey nodes store their own rotation in the YAML but Urb::Quad::Rotation
delegates to Below->Rotation. boundary_id now walks the below-chain to the
ground-floor rotation, matching Perl exactly.
After fixes all 35 corpus files produce edge counts matching Perl oracle.
Added:
- src/homemaker/graph.py: build_graphs (two-phase pattern), has_adjacency,
has_vertical_connection (faithful no-overlap stub per DESIGN §8.1),
find_missing_spaces, check_adjacency, check_level_constraints,
check_vertical_connectivity
- src/homemaker/dom.py: @dataclass(eq=False) on Node for NetworkX hashability;
is_outside, is_supported, is_unsupported, merge_divided
- tests/test_graph.py: 7 tests, edge counts vs Perl oracle on all 35 files,
exact widths for 2f45907, merge_divided smoke, two-phase independence
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Bruno's correction: 'C' was never a 'covered' type — Is_Covered is a
geometric predicate. Urb generics are canonically uppercase (get_space_types
qw/C O S/; corpus 100% uppercase). The driver/operator type pool emitted
lowercase 'c'/'o', creating mixed-case designs that fragmented Dom->Ratios
class buckets and fired the latent ratio_type first-match nondeterminism
(which the search promptly reward-hacked). Operators now emit uppercase
generics only and class checks match case-insensitively (t[0].lower() in
'cos', cf. Is_Circulation/Is_Outside). The Urb-side class-sum patch remains
as defensive hardening, zero-impact on canonical designs (35/35 parity).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
driver.py (homemaker-py-b39): tournament selection, operators.mutate (storey
ops down-weighted) + area-matched crossover, every child's geometry
delegated to the warm-started inner loop (Lamarckian write-back; children
use a single local CMA phase - the exploratory ladder phase exists for cold
projections children never face). Budget stated and accounted in oracle
evaluations; near-duplicate fitness guard against population collapse
(neutral mutations are common, per 8cs).
free_with_keys/ratio_map/warm_x0 promoted from the 8cs experiment into
innerloop.py as the Lamarckian inheritance API; alignment with
solver.free_branches asserted across the corpus.
tests/test_driver.py fakes the inner loop: budget accounting, monotone
improvement history, warm-start + sigma plumbing, valid .dom output.
31 tests pass. experiments/run_search.py is the end-to-end acceptance run.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
operators.py (homemaker-py-nyb): divide/undivide/retype/swap/rotate/
level_add/level_delete + Urb-style area-matched base-storey crossover.
Operators edit the decoded Node tree; genome.encode absorbs all repair
(dangling deltas, storey misalignment) so every child is a valid genome
by construction. Geometry moves deliberately absent — the inner loop owns
continuous DOF, and 8cs made Lamarckian re-optimisation mandatory.
Fixes dom._link to CLEAR stale below-links when a path vanishes from the
storey below (undividing a base branch left upper nodes pointing at
orphaned quads; oracle scoring unaffected but in-process geometry crashed).
Acceptance (experiments/operator_locality.py, flag-on): 115/115 children
scored without error; geometry perturbation small for core ops (retype
0.07, divide/undivide 0.14, swap/crossover 0.16-0.17), fitness
perturbation large for all (0.68-0.99 rel) — the 0.5^n cliff flags most
raw moves, confirming warm-started re-optimisation + penalty reshaping
as the load-bearing design choices. 27 tests pass.
Closes homemaker-py-nyb.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Corpus: all 35 scores shift (x1.0-1.24, daylight pinned), one expected
failure-set change (458aa8b8 +2 crinkliness), oracle ~8% faster batched.
New deterministic-seed reference gains become the accept_innerloop bars:
x1.63 / x1.70 / x1.68 at budget 400, ~35 oracle calls per topology.
urb-evolve respects the flag by construction (in-process fitness reads
ENV at call time). Old flag-off numbers kept in DESIGN as historical.
Closes homemaker-py-gp2 (Urb-side patch lives in /home/bruno/src/urb,
uncommitted there pending review).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
experiments/warm_vs_cold.py (homemaker-py-8cs): top-storey divide/undivide
mutations on corpus designs, path-keyed inheritance of the parent's
optimised ratios (surviving cuts keep values, new cuts 0.5), per-evaluation
convergence traces; reports oracle evals to 95% of best final, warm vs
cold. Machinery validated oracle-free (cut survival counts) and one
mutated child scored through the oracle.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The sigma-ladder default splits the budget into restart phases, so a
400-eval run reaches ~0.996 on the smooth test objective rather than
0.999+. Test now matches the component's contract. (Previous commit
landed with this failing because piping pytest to tail masked its exit
code.)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
innerloop.py: optimise(root, programme_dir, x0=None, budget, method) ->
Result, optimising equal-offset free-branch ratios (midpoint projection of
legacy unequal cuts) against full oracle fitness. OracleEvaluator scores
each population in one batched perl call. Methods: cma (default) — multi-
start sigma ladder (0.05 local, 0.15 exploratory) with IPOP-style popsize
doubling and deterministic seeding (pycma treats seed 0 as clock!) — and
compass with Hooke-Jeeves pattern moves, kept for the d0s bake-off.
Acceptance (experiments/accept_innerloop.py, §4.5 bars vs unprojected
originals, within-noise tolerance 1%): x1.65 / x1.66 / x1.58 against bars
x1.24 / x1.67 / x1.59, no new failures, 46 oracle calls vs Nelder-Mead's
200. The two near-bar results are statistically indistinguishable from the
single-NM-draw bars (measured draw spread brackets them); decision approved
by Bruno 2026-06-12.
Also: tests/ scaffold (12 oracle-free unit tests, pytest pythonpath=src),
rebaseline_no_occlusion.py for homemaker-py-gp2, cma>=3.0 dependency
(installed via dnf), dead-variable cleanup in solver.py.
Closes homemaker-py-1p0.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Strategy decision (Bruno): occlusion is orthogonal to building a scalable
optimiser. Instead of porting Sun/Occlusion/CIESky, disable it in Urb
behind an env flag (daylight -> 1, illumination factor -> 1 so crinkliness
is unweighted external wall area / floor area). Python occlusion rebuild
deferred until optimisation is fully native.
Tracker: new homemaker-py-gp2 (flag + re-baseline) gates gnw/way/uxz;
homemaker-py-2g5 re-scoped to the post-Phase-5 Python rebuild (P4).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
oracle.score_batch() writes/cleans N outputs and runs urb-fitness.pl once
with all file names; oracle.score() is now a thin wrapper. Adds
Score.fail_lines (sorted) because Perl hash-order randomisation shuffles
.fails line order between runs, and documents Urb's ~1-ULP score
nondeterminism (compare with rel tolerance, never ==).
experiments/bench_batch_oracle.py validates batch-vs-single parity on the
35-file corpus and benchmarks: 0.98 s/dom batched vs 1.27 s/dom single
(x1.30), all files identical (fitness to 1e-12 rel, exact failure sets).
Closes homemaker-py-av5.
Self-contained design + plan structured for breaking into bd (beads) tasks:
domain constraints that fix the slicing representation; what was built;
full empirical record (geometry port validated 35/35; area-proxy solver
falsified; perpendicular artifact resolved via equal-offset cuts; full-fitness
frozen-topology optimisation validated with 24-67% headroom; 0.5^n cliff);
validated memetic architecture; component plan; phased roadmap; risks/open
questions; repro steps; gotchas.
Oracle throughput measured: ~0.99s/dom batched vs 1.65s single (assessment-
dominated). urb-fitness.pl batches many doms per call, so native fitness is a
later speed/scale optimisation, not a gate; favour population/batch optimisers
and prototype the search on the batched oracle first.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Driving equal-offset cut ratios with Nelder-Mead against the REAL oracle
fitness (full objective, no proxy) improves all three test candidates with
zero new failures:
2f45907 (best evolved) 0.012617 -> 0.015684 x1.24 (2->2 fails)
candidate-002 0.007375 -> 0.012319 x1.67 (2->2 fails)
c964435 (baseline) 0.003667 -> 0.005836 x1.59 (3->3 fails)
Headroom widens on weaker designs. The EA under-optimises geometry by
24-67% even on its best result. This validates a full-fitness geometry
inner loop (NOT the earlier area-proxy solver) and motivates a memetic
architecture: topology search outside, full-objective geometry optimise
inside, gated on a native Python fitness (oracle at ~3s/call is too slow).
Adds the bottom-up ratio solver, programme parser, Perl-oracle bridge,
and two experiments. Headline finding: the "isolated size solver on a
frozen topology" hypothesis is NOT validated.
- resolve_ratios.py: re-solving candidate-002 from programme targets
recovers areas accurately but scores below the original (introduces
width/perpendicular/crinkliness failures the area objective ignores).
- refine_sweep.py: warm-start refine of all 34 evolved candidates
regresses 34/34 (fails 124->297 perpendicular-tied; 124->626 area-only
with free skew). Moving cuts to fix room area breaks the coupled
adjacency/access/shape constraints those designs balanced.
Conclusion: sizing is not separable from the rest of Urb's fitness;
a geometry inner loop must optimise the full objective, not an area proxy.
Geometry port remains validated byte-identical to Urb.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Clean-room Python successor to Urb for programme-driven layout search.
This initial commit establishes the .dom bridge format and a faithful
port of Urb's top-down quad geometry, validated byte-identical against
Urb across all 35 programme-house example files (including the wall
inset and multi-storey wall-stacking inheritance).
- dom.py: .dom YAML <-> Node tree, parent/below/position linkage,
wall_outer inset on load, raw-corner stash for round-tripping
- geometry.py: Coordinate/Coordinate_a/_b/Area/Length + Coordinate_Offset
- experiments/dump_areas.{py,pl}: geometry regression harness
