bruno/homemaker-layout
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

h10: confirm §12.3 A/B was already at fixed worker count, no re-run needed

Browse source 
run_9gp_ab.sh never threaded a worker count through run_staged_search.py, so
every §12.3 arm ran at n_workers=1 (serial) — the one mode §12.4 already
proved byte-for-byte reproducible even before the completion-order
determinism fix (that bug is ProcessPoolExecutor as_completed-only).
Spot-checked empirically: same config run twice gave identical fail counts
at every checkpoint. Closes homemaker-py-h10 as confirmed-null without
re-spending the ~8 core-hours a full sweep re-run would cost.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01R8agJBT2ZpmF3ErW7wi2wY
This commit is contained in:
Bruno Postle 2026-07-30 09:51:27 +01:00
parent b78ba8c67d
commit eec075c4a7
2 changed files with 42 additions and 21 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

44
.beads/issues.jsonl
View file

@ -22,6 +22,8 @@
{"id":"homemaker-py-1p0","title":"Geometry inner loop: full-objective equal-offset ratio optimiser","description":"DESIGN.md §5.1, §7 Phase 1. Productionise experiments/optimize_fullfitness.py into homemaker: optimise(topology, x0=None) -\u003e (geometry, fitness). DOF = equal-offset division ratios of free branches (solver.free_branches, lowest-storey cut ownership), clipped to [eps, 1-eps]. Objective = full oracle fitness (never a proxy — §4.2 falsified). Must support warm-start x0 (§5.6) and a population/batch evaluation mode so each iteration scores via one batched oracle call (§4.6).","acceptance_criteria":"Reproduces or exceeds §4.5 gains (x1.24–x1.67, no new failures) on 2f45907, candidate-002, c964435; works as a library call on any corpus .dom","status":"closed","priority":1,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-06-11T23:36:58Z","created_by":"Bruno Postle","updated_at":"2026-06-12T08:46:31Z","started_at":"2026-06-12T00:14:19Z","closed_at":"2026-06-12T08:46:31Z","close_reason":"innerloop.optimise() lands: batched CMA-ES sigma ladder (0.05/0.15, IPOP popsize doubling, deterministic seeding) over equal-offset free-branch ratios vs full oracle fitness; warm-start x0 supported. Acceptance vs unprojected originals: x1.65/x1.66/x1.58 against bars x1.24/x1.67/x1.59, no new failures, 46 oracle calls vs NM's 200. Two near-bar results accepted as reproduced-within-noise (1% tol) — draw spread brackets the single-NM-draw bars; approved by Bruno 2026-06-12. Gotchas: equal-offset projection of legacy unequal cuts loses fitness/adds failures (midpoint projection used); pycma seed=0 means clock-seeded.","dependencies":[{"issue_id":"homemaker-py-1p0","depends_on_id":"homemaker-py-av5","type":"blocks","created_at":"2026-06-12T00:39:33Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":3,"comment_count":0}
{"id":"homemaker-py-8cs","title":"Experiment: warm-vs-cold start of inner loop (Lamarckian inheritance)","description":"DESIGN.md §5.6, §4.6. Warm-starting a child topology's inner loop from the parent's optimised ratios is the main lever for cutting per-topology cost (~3 min/topology cold). Apply single topology mutations to optimised corpus designs, re-optimise warm (surviving cuts keep values, new cuts get heuristic defaults) vs cold, compare oracle-call counts to convergence at equal final fitness.","acceptance_criteria":"Speedup factor measured across \u003e=10 mutated topologies; decision recorded (expect order-of-magnitude; if \u003c2x, revisit §4.6 Phase-2 scoping)","notes":"Experiment script committed (experiments/warm_vs_cold.py, 1cc86c8) and machinery validated oracle-free; one mutated child scored through the oracle OK. Waiting on homemaker-py-gp2 reference run to finish, then execute under URB_NO_OCCLUSION=1 (3 parents x 400 evals + 12 children x 2 x 200 evals, ~1.5-2 h oracle time). Default budgets: parent 400, child 200; target = evals to 95% of best final.","status":"closed","priority":1,"issue_type":"task","owner":"bruno@postle.net","created_at":"2026-06-11T23:36:58Z","created_by":"Bruno Postle","updated_at":"2026-06-12T11:44:45Z","closed_at":"2026-06-12T11:44:45Z","close_reason":"Measured (URB_NO_OCCLUSION=1, parent budget 400, child 200, 12 single mutations across 3 designs): cold start reached 95% of warm final in 0/12 cases within budget — speedup unbounded at practical budgets; warm finals beat cold finals x1.2-x4 in 12/12; 6/12 warm starts were within 95% at 1 eval (near-neutral mutations). Decision: Lamarckian warm-starting is MANDATORY in the memetic driver (homemaker-py-b39), not an optimisation; cold starts produce strictly worse geometry at equal budget. Note: 2 undivides were exactly fitness-neutral (same-type merge == Merge_Divided equivalence) — locality datum for homemaker-py-nyb.","dependencies":[{"issue_id":"homemaker-py-8cs","depends_on_id":"homemaker-py-1p0","type":"blocks","created_at":"2026-06-12T00:39:34Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-av5","title":"Batched oracle: score many .dom files per invocation","description":"oracle.py currently scores one .dom per urb-fitness.pl call (~1.65 s/dom). DESIGN.md §4.6: batching amortises Perl startup to ~0.99 s/dom and is required so population/batch optimisers can score a whole generation in one oracle call. Extend oracle.py with a batch API: write N .dom files, one perl invocation, parse N .score/.fails pairs. Keep the single-file path for compatibility.","acceptance_criteria":"Batch of 35 corpus files scores in one perl invocation; per-file results identical to single-file calls; measured s/dom reported","status":"closed","priority":1,"issue_type":"task","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-06-11T23:36:56Z","created_by":"Bruno Postle","updated_at":"2026-06-12T00:14:06Z","started_at":"2026-06-11T23:50:40Z","closed_at":"2026-06-12T00:14:06Z","close_reason":"score_batch() lands in oracle.py; 35-file corpus parity verified single-vs-batch (1e-12 rel fitness, exact fail sets); 0.98 s/dom batched vs 1.27 single, x1.30","dependency_count":0,"dependent_count":1,"comment_count":0}
{"id":"homemaker-py-1s3","title":"Multi-use leaves as permanent design goal (§26 path b, never attempted)","description":"§26 (homemaker-py-9o5/xi7/b3v) scoped two readings of multi-use leaves (a leaf legitimately serving several DIFFERENT compatible programme codes at once, e.g. study+guest-bedroom, kitchen+dining — Stewart Brand's 'loose-fit' rooms): (a) superposition as a SEARCH RELAXATION — carry uncommitted candidate types per leaf, collapse to one usage only at scoring time; (b) multi-use as the PERMANENT DESIGN GOAL, surviving into the output with no collapse. Only (a) was built and measured, and it was NULL/NEGATIVE — diagnosed as underperforming not from a relaxation gap (measured small, gap_ratio 1.01-1.23) but because the geometry floor dominates: type labels are not the binding constraint on these programmes, so easing them buys nothing while re-typing adds feasibility noise (fitness.py per-eval collapse perturbs counts/adjacency).\n\nPath (b) was never built. It's structurally different from (a): rather than a per-eval relabelling relaxation on top of the existing leaf count, it would permanently REDUCE leaf count by having one leaf serve two rooms' worth of programme requirement simultaneously — the same structural mechanism as leaf-sharing (§13.3, homemaker-py-x3b), which is the single biggest positive lever in the whole DESIGN.md log (harbor-house −21% to −32% at various stages, because it cuts leaf count in a way the search cannot mutate back — §13.4/13.5's key finding that levers the search 'cannot erode' compound, unlike shape-only levers that wash out over a 20k-eval budget).\n\nMultiple independent diagnostics (§12.3 calibration, §12.4's conclusion, §13.1's per-leaf saturation analysis) converge on: the residual fail floor at harbor/maple scale is driven by having as many leaves as there are distinct rooms (52 rooms -\u003e 73 leaves at 44% utilisation gives every leaf a high perimeter/area ratio). Leaf-sharing already exploits this for SAME-code multiplicities; path (b) would extend the same leverage to DIFFERENT-but-compatible codes, which is a materially larger addressable set on programmes with many small single-instance rooms (offices, WCs, meeting rooms — see health-centre, examples/health-centre, 19 distinct codes).\n\nTask: design + build path (b) — likely: SpaceReq gains a compatibility/co-location relation (reuse or extend programme.derive_interchange_classes' S1-S4 guards, or a new explicit 'co_locate' declaration since 'interchangeable' semantics don't fit two DIFFERENT codes coexisting), a leaf can be permanently typed as serving code-pair (or code-set) X, and check_space_counts/quality functions treat the areas as shared per the k-instances-per-leaf model §13.3 already uses for same-code sharing. Gate behind a new conf flag (default OFF, bit-identical when off, per this project's established pattern for every §13.x/§20+ lever). A/B against the current default stack on harbor-house and health-centre (the diverse-room-type programme built for xyu/9yx, §31/§32) before considering a default flip — same discipline as every other lever in this log.","status":"open","priority":2,"issue_type":"task","owner":"bruno@postle.net","created_at":"2026-07-30T07:55:56Z","created_by":"Bruno Postle","updated_at":"2026-07-30T07:55:56Z","dependency_count":0,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-h10","title":"Re-run §12.3 reassociate/shape-feasibility A/B at fixed worker count","description":"§12.3 (homemaker-py-9gp) measured mutate_reassociate (M3 Wong-Liu move) and the shape-feasibility pre-filter as negative: +3.3/+4.0 fails on maple/harbor. But that A/B ran BEFORE §12.4 (homemaker-py-c3g) found and fixed a real nondeterminism bug — driver._run_batch admitted parallel futures in completion order rather than submission order, producing ±3-6 fail noise between otherwise-identical runs. §12.4's own writeup flags this explicitly: 'sub-±3 effects (the §12.3 +3-4 negatives, the §12.4 ±1.7) should be re-run at a single fixed worker count before being trusted as magnitudes.' That re-run was never done for §12.3.\n\nReassociate is the only search-machinery move in the whole DESIGN.md log verified to reach genuinely new tree topologies (confirmed on synthetic cases in §12.3's own tests) — every other outer-search-machinery lever tried (niching, graded objective, island model, grain annealing, graded connectivity, circulation repair, beam search, ruin-recreate, bubble-diagram signal) is independently null-to-negative for other reasons, so this is the one candidate whose 'negative' verdict might be pure measurement artifact rather than a real finding.\n\nTask: re-run experiments/run_9gp_ab.sh (maple-court + harbor, seeds 0/1/2, 20000 evals, staged) with the post-c3g determinism fix in place, at a single fixed worker count (matching whatever the original run used — check the script/log for workers=N). If the negative holds at fixed worker count, close as confirmed-null (upgrade §12.3's confidence). If it flips positive or neutral, this reopens the reachability question closed in §12.3/§12.4's 'residual is geometry floor, not search-reachability' conclusion — would need a larger-N confirmation before any default flip, per this project's own evidentiary bar (cf. f1d/1ph/e01 pattern).","status":"closed","priority":2,"issue_type":"task","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-07-30T07:55:15Z","created_by":"Bruno Postle","updated_at":"2026-07-30T08:14:43Z","started_at":"2026-07-30T08:06:25Z","closed_at":"2026-07-30T08:14:43Z","close_reason":"Confirmed without a full re-run: run_9gp_ab.sh/run_staged_search.py never threaded a worker count, so every §12.3 arm already ran at n_workers=1 (serial) — the mode §12.4 already proved byte-for-byte reproducible even before the completion-order fix (that bug is ProcessPoolExecutor-as_completed-only). Spot-checked empirically too: same config run twice (harbor-house s0, budget 300) gave identical fail counts at every checkpoint. §12.3's negative verdict is CONFIRMED at fixed worker count; DESIGN.md §12.4 updated with the finding.","dependency_count":0,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-9wi","title":"Adjacency-aware discrete assignment for finish-time collapse (QAP/CP-SAT)","description":"fitness.collapse_superposition (homemaker-py-9o5/94g family) already does exact optimal relabelling of superposed leaves via brute-force permutation (CLASS_CAP\u003c=4) or Hungarian (linear_sum_assignment) beyond that -- but _best_assignment's docstring is explicit that the objective is deliberately SEPARABLE per leaf (quality_size * quality_width * quality_proportion only); perpendicular/crinkliness/access/adjacency are assumed usage-invariant within a class and left out, because adjacency quality depends on PAIRS of leaf-label assignments, not one leaf at a time, which breaks the exact separable solve.\n\nProposal: extend the collapse step to account for adjacency between candidate labels -- either a quadratic-assignment-style local search (2-opt swaps over the current Hungarian solution, accepting swaps that improve total adjacency satisfaction) or a CP-SAT (OR-Tools) encoding of the labelling problem with pairwise adjacency terms. This directly extends the one search-adjacent technique (exact/near-exact discrete assignment) that has actually paid off in this project, into territory the current separable solve cannot reach.\n\nMeasure against the current Hungarian-only collapse on harbor-house (heavy interchange-class usage: neighborhoods, meeting rooms, individual rooms) where adjacency-blind relabelling is most likely to leave adjacency fails on the table.","status":"closed","priority":2,"issue_type":"task","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-07-25T20:18:24Z","created_by":"Bruno Postle","updated_at":"2026-07-26T19:57:55Z","started_at":"2026-07-26T14:50:36Z","closed_at":"2026-07-26T19:57:55Z","close_reason":"2-opt local search DELIVERED beyond collapse_global's Jacobi adjacency relaxation:\nFitness._collapse_value (refactored per-leaf-per-code value, shared by the\nassignment matrix and the polish) + Fitness._two_opt_adjacency_polish (same-level\npairwise label-swap search, kept only on strict improvement -- monotone by\nconstruction) + collapse_global(local_search=..., local_search_passes=...) +\nhomemaker-collapse --local-search CLI flag + regression test\n(test_two_opt_polish_escapes_jacobi_plateau) that proves the Jacobi loop can\n2-cycle between two labellings satisfying ZERO of a satisfiable adjacency set,\nand that 2-opt escapes it.\n\nWent with 2-opt over CP-SAT/OR-Tools: no new dependency (project has no\nortools), directly extends the existing Jacobi machinery, and the issue listed\nit as the first alternative. QAP is NP-hard in general so this is a local\nsearch, not an exact solve, but it strictly dominates the Jacobi-only result\n(never worse, by construction).\n\nMeasured against harbor-house per the issue's own instruction: swept all 11\nevolved-*/3m/materialised .dom files, comparing collapse_global(local_search=False)\nvs (local_search=True). 10/11 matched exactly (Jacobi was already at the local\n2-opt optimum), 0 regressed, 1 improved (evolved-anneal-3M.dom 21-\u003e19 fails --\nresolved a genuine mutual da1\u003c-\u003ek1 adjacency miss). Runtime \u003c1s even on the\nlargest file (90-fail evolved-3M.dom). Findings + the plateau proof saved via\nbd remember. Default left OFF (opt-in via local_search=True /\nhomemaker-collapse --local-search) pending a broader sweep and evolve.py CLI\nwiring -- spun off as homemaker-py-cdl.\n\n298/298 tests pass (7 in test_collapse_global.py, up from 6).","dependency_count":0,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-f1d","title":"Ruin-and-recreate LNS: rebuild wings with the adjacency-aware constructor mid-search","description":"DESIGN.md's own experiment log shows every 'search machinery' change (niching+restarts 11.5, graded objective 11.4, Wong-Liu reassociation+shape-feasibility 12.3, granularity 12.4, island model 14, grain annealing 16, circulation-repair ops 21/22) has come back null-to-negative, while construction/seeding quality (adjacency-aware seeding 11.6/11.7, proportion-aware seeding 12.2) is the only lever that has ever moved the fail count. operators._assign_adjacency_aware currently only runs once, at seeding.\n\nProposal: a large-neighbourhood-search move that periodically un-divides a whole wing/subtree of the CURRENT BEST individual and reconstructs just that region using the same proven adjacency-aware constructive heuristic (seeded from the surviving circulation spine as fixed_circ, same mechanism lift_base_to_storeys already uses for upper floors), instead of relying only on small local mutation operators to find improvements. Reuses the one technique with a real track record, applied repeatedly during search rather than once at initialisation.\n\nA/B against the current baseline on programme-house and harbor-house at a fixed worker count (see 12.4's determinism-fix note about serial vs parallel admission order before trusting sub-±3 effects).","notes":"Implementation landed (uncommitted, pending A/B): operators.mutate_ruin_recreate — picks a divided, live-cut wing of one storey (2..half its leaves), un-divides it, regrows+retypes it via a scope-generalised operators._assign_adjacency_aware (new 'scope' param restricts retyping to a leaf subset while fixed_circ seeds can be border leaves outside that subset), seeded from already-typed circulation leaves bordering the wing. Room-code budget inside the wing is preserved exactly; circ/outside counts rebuilt at circ_divisor=3/outside_divisor=3. Gated like reassociate/bridge_circulation: mutation_weights['ruin_recreate']=0.0 unless driver.search(enable_ruin_recreate=True); CLI flag --ruin-recreate/--no-ruin-recreate added to evolve.py (default off). 297 existing tests pass unchanged; added smoke coverage (200 applications on harbor-house constructive seeds: zero missing-space regressions, all canonical). A/B now running in background: experiments/run_f1d_ab.sh, qpk protocol (harbor-house budget=2500 seeds 1-3, programme-house budget=3000 seeds 1-5, workers=4, both arms finished with standard --collapse), results -\u003e scratch/f1d_ab_results.tsv.","status":"closed","priority":2,"issue_type":"task","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-07-25T20:18:02Z","created_by":"Bruno Postle","updated_at":"2026-07-26T08:30:13Z","started_at":"2026-07-25T20:20:56Z","closed_at":"2026-07-26T08:30:13Z","close_reason":"DONE (positive, size-dependent). operators.mutate_ruin_recreate landed: un-divides one wing of a storey (2..half its leaves), rebuilds it via a scope-generalised _assign_adjacency_aware seeded from bordering circulation, mirroring lift_base_to_storeys' core-inheritance mechanism. Gated like reassociate/bridge_circulation (enable_ruin_recreate, default off; --ruin-recreate CLI flag). Initial uniform-weight A/B was null but underpowered (op fired ~1/32 children); a weight=3.0 follow-up (_MUTATION_WEIGHTS['ruin_recreate']=3.0, kept in source) showed a statistically significant win on programme-house across 15 seeds (8W/1L/6T, mean fails 7.07-\u003e6.00, Wilcoxon p=0.041, sign-test p=0.020) but no consistent effect on harbor-house across 8 seeds (3W/2L/3T, mean fails 73.0-\u003e74.5, slight negative lean). Kept default OFF pending a size-threshold follow-up (not filed) -- same conservative bar qpk/1ph applied before its own larger-N confirmation. Full writeup: DESIGN.md §23.","dependency_count":0,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-mi7","title":"Prototype: 3D bubble-diagram relaxation of programme adjacency as a fitness signal","description":"Explore building a spring/force relaxation over the programme's required-space adjacency graph (multiple random-restart solutions), then score how well an actual Dom layout's real adjacency-graph distances correlate with a relaxed target's distances. Goal: an additional fitness term / search-guidance signal beyond the existing binary adjacency checks in graph.py. Prototype module: bubble.py. Validate by correlating similarity score against existing fitness .score on examples/programme-house's 36 scored .dom files.","notes":"Harbor-house real trajectory (driver.search, budget=6000, n=75 recorded individuals, fitness 3e-28 -\u003e 3.9e-17, fails 83-\u003e51):\n- embedding similarity(): spearman=0.164 p=0.16 (n.s.)\n- topological_similarity(): spearman=-0.160 p=0.17 (n.s.)\n\nFINAL PICTURE across 4 tests (2 programmes x 2 metric formulations, plus 2 canned-batch tests earlier): no statistically significant correlation anywhere between either the spring/embedding bubble-diagram similarity or the pure topological/abstract-graph-fitting similarity, and existing programme-driven fitness score. programme-house's real-trajectory result (rho=0.05 / rho=-0.06, n=100) is the cleanest data point — that programme has zero multi-count anonymous codes, so matched_leaves' known centroid-order matching heuristic cannot be confounding it, and it's still flat. Harbor-house is noisier (heavy anonymous counts: n x5, m x3, t x6, r x10, of x2 — the fixed centroid-order matching there is a real, uncontrolled confound) but tells the same story.\n\nRecommendation: do not pursue graph-relaxation-derived or pure-topological adjacency-matching as a fitness signal for this project without a fundamentally different formulation — two independent formulations, tested on two programmes with real evolved trajectories (not just static examples), both came back null. If revisited later, the harbor-house confound (anonymous-code instance matching) would need a real assignment solver (Hungarian/brute-force per homemaker-py-9o5's CLASS_CAP pattern) before drawing any conclusion there specifically, but programme-house's clean null already argues against the core idea. bubble.py is left in the repo (uncommitted) as a documented, working prototype/reference — not wired into fitness.py.","status":"closed","priority":2,"issue_type":"task","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-07-25T12:42:02Z","created_by":"Bruno Postle","updated_at":"2026-07-25T20:12:20Z","started_at":"2026-07-25T12:42:26Z","closed_at":"2026-07-25T20:12:20Z","close_reason":"Two independent formulations (spring/embedding bubble-diagram, pure topological hop-distance) both tested null against real evolve trajectories on programme-house (n=100, clean — no anonymous-code confound) and harbor-house (n=75). No positive correlation with existing fitness score found. bubble.py left in repo, uncommitted, as documented reference; not wired into fitness.py. Consistent with the project's broader pattern (see DESIGN.md §11.4/11.5/12.3/12.4/14/16/21/22): 'search machinery' / fitness-shaping changes have been null-to-negative essentially every time they've been tried; only construction/seeding quality and representation-relaxation changes (leaf-sharing, type superposition, global collapse) have ever moved the needle. This session's result is another data point for that pattern, not an exception.","dependency_count":0,"dependent_count":0,"comment_count":0}
@ -96,27 +98,27 @@
{"id":"homemaker-py-erc.6","title":"Experiment: inner-loop slack-expansion objective term","description":"Inner-loop counterpart to plot-fill construction. If Diagnostic B shows the inner loop has room to expand leaves into slack but no objective gradient to do so (the scalar rewards hitting target area but not exceeding it where slack exists), add a term/incentive so the ratio optimiser pushes leaf boundaries out to consume neighbouring slack and satisfy size, rather than parking at target.\n\nCONDITIONAL on Diagnostic B: build this only if B localizes the gap to the inner loop (room to expand, no gradient); if B shows construction targets too-small dims, prefer the plot-fill construction sibling. Must preserve the §5.4 inner-loop cliff / §4.9 lexicographic protection — the term sits where it cannot displace the fail-count ordering. A/B vs §12.2 baseline, seeds 0/1/2, 20000 evals, staged, default-OFF. Record DESIGN.md §13.6.","notes":"DEPRIORITISED by Diagnostic B (§13.2). B shows the inner loop CANNOT repair undersize: the slack is depth-driven maldistribution baked into the frozen topology, and the equal-offset ratio DOF cannot shrink a 14x leaf to feed a starved one without trading into shape fails (0.5^n cliff). Wrong DOF and wrong direction — the blocker is slicing POSITION, not a missing expansion reward. Fix belongs upstream in construction/topology (erc.4 re-scoped, erc.3). Keep as a low-priority follow-up only if a depth-balanced construction still leaves a residual size gradient the inner loop could pick up.","status":"closed","priority":4,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-06-22T23:16:24Z","created_by":"Bruno Postle","updated_at":"2026-06-28T13:22:22Z","closed_at":"2026-06-28T13:22:22Z","close_reason":"wont-fix (DESIGN §13.7): Diag B (§13.2) showed the inner loop cannot repair undersize (wrong DOF — slicing position, frozen-topology ratios). Superseded by depth-balanced construction (erc.4). Condition unmet.","dependencies":[{"issue_id":"homemaker-py-erc.6","depends_on_id":"homemaker-py-erc","type":"parent-child","created_at":"2026-06-23T00:16:23Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-erc.6","depends_on_id":"homemaker-py-erc.2","type":"blocks","created_at":"2026-06-23T00:16:47Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-erc.5","title":"Experiment: compactness-aware cuts (minimize leaf perimeter/area)","description":"Attacks the #1 factor, crinkliness (346) — a per-leaf perimeter/area property DISTINCT from proportion (aspect ratio). Proportion-aware seeding (leu.2) sizes splits but does not bias toward balanced, square-ish subdivision. Add a KD-tree-style 'keep both children compact' cut rule (prefer the cut orientation/position that minimises summed child perimeter/area) in construction.\n\nCONDITIONAL on Diagnostic A: if A shows per-leaf shape-fail is FLAT across densities (floor intrinsic to slicing density), better cuts at the same leaf count will not pay → this should be closed wont-fix in favour of leaf-sharing. Only build if A shows shape-fail RISES with density. A/B vs §12.2 baseline, seeds 0/1/2, 20000 evals, staged, default-OFF. Record DESIGN.md §13.5.","notes":"DEPRIORITISED by erc.1 verdict (§13.1): per-leaf shape-fail flat vs slicing density and cuts already squarest (_size_divisions_from_targets picks squarest rotation) yet still ~1.8 fails/leaf =\u003e little compactness headroom at fixed leaf count. Floor is intrinsic to leaf COUNT, not cut quality. Revisit only if leaf-sharing (erc.3) underdelivers.","status":"closed","priority":4,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-06-22T23:16:21Z","created_by":"Bruno Postle","updated_at":"2026-06-28T13:22:17Z","closed_at":"2026-06-28T13:22:17Z","close_reason":"wont-fix (DESIGN §13.7): Diag A (§13.1) showed the floor is intrinsic to leaf COUNT not cut quality; revisit condition was 'only if leaf-sharing underdelivers' but leaf-sharing OVER-delivered (−32…−39%, §13.3). Condition unmet.","dependencies":[{"issue_id":"homemaker-py-erc.5","depends_on_id":"homemaker-py-erc","type":"parent-child","created_at":"2026-06-23T00:16:21Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-erc.5","depends_on_id":"homemaker-py-erc.1","type":"blocks","created_at":"2026-06-23T00:16:43Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":1,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-2g5","title":"Rebuild occlusion/daylight/sun subsystem in Python (post-Phase-5, after optimisation fully native)","description":"DESIGN.md §6 port scope — a whole subsystem, not a term. quality_daylight (Leaf.pm:281-296) needs Urb::Misc::Sun + Urb::Field::Occlusion (+CIESky); quality_uncrinkliness also takes the occlusion object. Indoor spaces return 1 for daylight; cost is outdoor spaces + crinkliness. Port Sun_horizontal (262980-minute normalisation) and the occlusion wall set from Dom-\u003eWalls.","acceptance_criteria":"Daylight and crinkliness factors match Perl (float tolerance) across the corpus, including multi-storey cases","notes":"Re-scoped 2026-06-12: occlusion disabled in the Urb oracle instead of ported (see homemaker-py-gp2). Native fitness ships with simple crinkliness (illumination factor = 1, in homemaker-py-gnw). This issue is now the eventual Python occlusion rebuild, only after optimisation works entirely in Python. Restores outdoor-daylight and shaded-wall selection pressure.\nReframed 2026-06-17: orthogonal to epic homemaker-py-c4c. This is fitness FIDELITY (restoring daylight + shaded-wall selection pressure to match Perl), not search CAPABILITY — it changes what 'good' means, not the search's ability to find good. It will NOT improve final designs in the sense currently sought. Stays P4, deferred until the topology-search-quality epic lands and optimisation is fully native.","status":"open","priority":4,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-06-11T23:38:25Z","created_by":"Bruno Postle","updated_at":"2026-06-17T19:14:48Z","dependency_count":0,"dependent_count":0,"comment_count":0}
{"_type":"memory","key":"collapse-global-s-jacobi-adjacency-relaxation-homemaker-py","value":"collapse_global's Jacobi adjacency relaxation (homemaker-py-94g) is a synchronous per-round linear-assignment re-solve, which can 2-cycle indefinitely between two labellings that each satisfy ZERO adjacency requirements even though a permutation satisfying ALL of them exists -- proven on a minimal 4-cell chain (p1-q1-p2-q2, two disjoint adjacency pairs p1\u003c-\u003ep2/q1\u003c-\u003eq2) in test_two_opt_polish_escapes_jacobi_plateau. homemaker-py-9wi added Fitness._two_opt_adjacency_polish: a same-level pairwise-swap local search run after the Jacobi fixpoint, gated behind collapse_global(local_search=True) (default off, exposed as homemaker-collapse --local-search). Monotone by construction (a swap is kept only if it strictly increases total reward). Empirically on the 11 harbor-house evolved-*.dom/3m.dom/materialised-3M.dom layouts: 10 matched Jacobi-only exactly, 0 regressed, and evolved-anneal-3M.dom improved 21-\u003e19 fails (fixed a genuine mutual da1\u003c-\u003ek1 adjacency miss the Jacobi loop couldn't reach)."}
{"_type":"memory","key":"experiment-seeding-pitfall-run-search-scaled-py-s","value":"Experiment seeding pitfall: run_search_scaled.py's default PH_SEED (c964…dom) is a FINISHED programme-house design — passing it warm-starts and floors at ~3 fails, NOT a blank-slate topology search. For blank-slate runs comparable to §11.5/§11.6 baselines, seed from examples/programme-house/init.dom (a bare undivided plot; driver bootstrap auto-triggers only on bare plots). Bit the 6zy sweep — first pass used c964 and falsely showed 3-fail floor across the whole grid."}
{"_type":"memory","key":"never-use-corpus-filenames-candidate-001-dom-candidate","value":"Never use corpus filenames (candidate-001.dom, candidate-002.dom, generated.dom, init.dom, etc.) as --output targets when running experiments. These are test fixtures. Always write experimental outputs to scratch/ or a timestamped path. Lesson from 2026-06-14: warm-start runs overwrote candidate-001/002.dom and broke graph tests."}
{"_type":"memory","key":"programme-house-optimisation-result-2026-06-14-15","value":"Programme-house optimisation result (2026-06-14/15): best achievable is 1 fail (l1 wrong level, score ~0.005). 0 fails is geometrically impossible: l1 (min 27m²) must occupy ll (~23m²) at level 0, which eliminates the t3-adj-C provider; dividing ll into lll(l1)+llr(C) gives llr proportion ~6:1 (fails). Python memetic optimizer achieves 1 fail in 50k evals vs Perl optimiser's 2-3 fails. Winning topology: TWO C nodes at level 0 — ll(C) for t3-adj-C via geometric contact, rl(C) for staircase via tree-sibling adjacency to rrr(O). Best .dom: scratch/from-warmstart-fixed.dom and scratch/from-compound3-fixed.dom."}
{"_type":"memory","key":"proportion-aware-constructive-seeding-leu-2-12-2","value":"Proportion-aware constructive seeding (leu.2/§12.2): sizing seed cuts from target AREAS only regresses (thin slivers wreck aspect); you must ALSO pick each cut's rotation for child squareness. It is a convergence ACCELERATOR via a deeper local optimum around the constructed topology: wins where that topology is roughly right and budget is scarce (harbor -13%, maple -10% at 20k evals) but DELAYS small programmes where the seed must be restructured by undivide (programme-house regresses at fixed budget, yet reaches the floor given budget - speed, not asymptote). Default-on. Also: n_storeys must honour storey_minimum, not just level: keys (programme-house storey_minimum:2, all rooms level:0 - was seeded 1 storey short; cq1)."}
{"_type":"memory","key":"homemaker-py-pythonpath-set-pythonpath-home-bruno-src","value":"homemaker-layout PYTHONPATH: package installed as 'homemaker-layout' via pip install -e . so 'import homemaker_layout' works from anywhere without PYTHONPATH. For running tests use 'python -m pytest' from project root /home/bruno/src/homemaker-layout (pyproject.toml adds src/ automatically). Never try pip show homemaker — that's the old homemaker-addon conflict."}
{"_type":"memory","key":"strategy-decision-2026-06-12-bruno-occlusion-daylight","value":"Strategy decision 2026-06-12 (Bruno): occlusion/daylight is ORTHOGONAL to building a scalable optimiser. Disable it in Urb (env flag, homemaker-py-gp2) rather than port it; native fitness uses simple crinkliness (illumination factor = 1); rebuild occlusion in Python only after optimisation is fully native (homemaker-py-2g5, now P4). Consequence: all scores change when the flag flips — re-baseline corpus/.score, DESIGN \\$4.5 gains, gate bars at one clean boundary AFTER homemaker-py-1p0 closes; Phase-2 urb-evolve benchmark must run with the same flag."}
{"_type":"memory","key":"urb-fitness-bug-found-fixed-2026-06-12","value":"Urb fitness bug found+fixed 2026-06-12 (patch in /home/bruno/src/urb, uncommitted): ProgrammeDriven.pm ratio_o/ratio_type grepped case-insensitively over the ratios hash and took the FIRST key — nondeterministic (x4.5 score swings) for designs with mixed-case type classes (both 'c' circulation and 'C' covered). Fixed to SUM the class (matches Is_Circulation//Is_Outside semantics); 35/35 corpus scores unchanged. CRITICAL for homemaker-py-3y7/gnw: the native port must implement class-SUM ratios. Building.pm has the same unpatched pattern (site-driven path, not used by our oracle). Also: the memetic search reward-hacked this bug before the fix — search results predating it are noise artifacts."}
{"_type":"memory","key":"warm-x0-initialization-bug-pattern-when-a-topology","value":"warm_x0 initialization bug pattern: when a topology operator explicitly sets division ratios on a newly-created node (e.g. compound_fix sets node.division=[0.25,0.25] for t3), parent.ratios has no entry for that node (it was a leaf). warm_x0 defaults it to 0.5, corrupting the inner loop's starting point and making the operator invisible to lex comparison. Fix: only propagate child ratios for nodes where the parent node was NOT already divided; stale hidden nodes revealed by structural mutations (swap flipping b.below) must NOT contribute their pre-writeback values. See driver.py lines 259-267 (fixed 2026-06-14)."}
{"_type":"memory","key":"cli-tool-style-prefer-python-m-homemaker-module","value":"CLI tool style: prefer python -m homemaker.module --parameters pattern, installable via pip install -e . with pyproject.toml entry_points. Not standalone bin/ scripts."}
{"_type":"memory","key":"urb-oracle-nondeterminism-urb-fitness-pl-output-varies","value":"Urb oracle nondeterminism: urb-fitness.pl output varies run-to-run from Perl hash-order randomisation — .fails line ORDER shuffles (compare sorted, use oracle.Score.fail_lines) and the score float can flip by ~1 ULP (compare with math.isclose rel_tol=1e-12, never ==). Not a batching artifact; affects single runs too. Matters for the Phase 3 native-fitness parity gate (homemaker-py-uxz)."}
{"_type":"memory","key":"correction-to-urb-fitness-bug-memory-bruno-2026","value":"CORRECTION to urb-fitness-bug memory (Bruno, 2026-06-12): 'C' is NOT a 'covered' type — Is_Covered is a geometric predicate (indoor space above). Urb's generic types are canonically UPPERCASE: C=circulation, O=outside, S=sahn (get_space_types qw/C O S/; corpus is 100% uppercase, never 'c'/'o' leaves). The mixed-case designs that fired the latent ratio_type first-match bug were created by homemaker's own operator type pool emitting lowercase 'c'/'o' — fixed: driver/operators now emit uppercase generics only, and class checks use t[0].lower() in 'cos'. The Urb class-sum patch stays as defensive hardening (zero impact on canonical designs). Native port (3y7/gnw): treat type classes case-insensitively, generics canonically uppercase."}
{"_type":"memory","key":"island-model-psk-14-is-a-null-priming","value":"Island model (psk, §14) is a NULL: priming a population from N converged independent elites + crossover-heavy migration does not beat best-of-N at equal total budget (maple island 124 vs control 116). The child_probe instrument shows WHY: area-matched crossover across independently-converged elites almost never synthesizes (1-3 of ~64 children beat the better parent, max drop 2-5) because the slicing encoding is non-canonical (9gp), so splices are disruptive not combinatorial. Search-machinery null #3 after graded-objective and niching/restarts; residual stays geometry/shape-bound."}
{"_type":"memory","key":"run-to-run-reproducibility-in-homemaker-layout-serial","value":"Run-to-run reproducibility in homemaker-layout: serial search (workers=1) is byte-for-byte deterministic; parallel (workers\u003e1) is now deterministic too AFTER fixing driver._run_batch to admit futures in submission order (was as_completed/completion order, bug xcy). Reproducibility holds only for a FIXED worker count — serial vs parallel differ because children-per-iteration is 1 vs n_workers (different batch granularity), which is expected, not a bug. The constructive seeder was NEVER nondeterministic: _assign_adjacency_aware has unique idx tiebreaks; comparing topologies with Python builtin hash() of the signature STRING is invalid (PYTHONHASHSEED salts str hashing per process) — use a stable hash (sha1) or genome.signature equality."}
{"_type":"memory","key":"adjacency-in-binary-slicing-tree-is-structural-not","value":"Adjacency in binary slicing tree is structural, not geometric: the inner-loop NM cannot fix topological adjacency failures. Two paths exist: (1) tree-sibling adjacency — a node is adjacent to its sibling in the tree; (2) cross-zone geometric adjacency — leaves from different subtrees that happen to share a boundary. Staircase/adjacency fails require a topology mutation that changes which nodes are siblings or which zones touch. This was proved empirically on programme-house: staircase fail from rot=0 layout could not be fixed by NM but was fixed by level_retype creating a two-C topology (2026-06-14/15)."}
{"_type":"memory","key":"experiment-harness-gotcha-the-leaf-sharing-relaxed-objective","value":"Experiment harness gotcha: the leaf-sharing RELAXED objective (§13.3) is injected ONLY by monkeypatching fitness.load_config in the parent process (run_staged_search.py / probe scripts). This is parent-process-only and does NOT propagate into ProcessPoolExecutor workers (n_workers\u003e1), which re-import fitness fresh and score under the STRICT on-disk patterns.config -\u003e r.n_fails MISMATCH (worker strict vs parent relaxed re-score). ALL §13.x floor runs were therefore SERIAL. Any future PARALLEL leaf-sharing experiment will silently mis-score until leaf_sharing lives on disk/CLI (tracked: homemaker-py-x3b). The parallel driver itself is correct; both paths score via load_config(programme_dir)."}
{"_type":"memory","key":"homemaker-py-3l6-fix-leaf-sharing-evolve-runs","value":"homemaker-py-3l6 fix: leaf-sharing evolve runs now auto-finish before write via driver.polish_finish — unfold_shared_leaves() then a warm-started leaf_sharing=False polish search (--polish-budget, default budget//2). Makes the written .dom honest under canonical homemaker-fitness (internal==canonical when leaf_sharing off). Interrupt path forces polish_budget=0 (unfold+rescore only). This is yaa's unfold-then-polish, made automatic; Schedule B annealing is still kpu."}
{"_type":"memory","key":"user-preference-bruno-this-is-a-fedora-system","value":"User preference (Bruno): this is a Fedora system — NEVER install Python packages via pip without asking first; always ask whether to install the rpm via dnf (e.g. python3-cma) before considering pip. Applies to any dependency additions."}
{"_type":"memory","key":"9o5-multi-use-leaves-is-path-a-superposition","value":"9o5 multi-use leaves is path (a) — superposition as SEARCH RELAXATION that COLLAPSES to specific usage at the end, NOT path (b) loose-fit/no-collapse. Bruno's intent: codes with SIMILAR leaf requirements form an interchangeable equivalence class; during evolution the solver doesn't commit which leaf serves which specific usage (smoother landscape, no fighting over exact leaf usage); at the end the layout is CONDENSED to specific usages by brute-forcing the in-class assignment (3 interchangeable usages over 3 leaves = 3! = 6 combinations to check, pick best). 'Derive automatically' compatibility = requirement-similarity grouping. This reverses the issue's stated 'path b preferred' note."}
{"_type":"memory","key":"deceptive-valleys-in-topology-search-when-every-single","value":"Deceptive valleys in topology search: when every single-step mutation from a target state passes through a high-fail intermediary (e.g. level_fix displaces a room into 5+ new fails), a compound operator that atomically applies two coordinated changes can escape. Design compound operators to land on the low-fail state directly, bypassing the deceptive gradient. Programme-house example: level_compound_fix atomically moves the level-constrained room AND re-inserts the displaced room adjacent to C in one step (operators.py, 2026-06-14)."}
{"_type":"memory","key":"multi-storey-staircase-consistency-when-dividing-or-retyping","value":"Multi-storey staircase consistency: when dividing or retyping a circulation (C) leaf at one level, the same structural change should be propagated to the matching leaf on ALL other storeys so the stair core path is maintained. The optimizer cannot fix staircase disruptions through trial-and-error geometry alone — it requires a synchronized multi-level operator that applies the same topology change to every storey simultaneously."}
{"_type":"memory","key":"collapse-global-94g-and-any-label-usage-optimisation","value":"collapse_global (94g) and any label/usage optimisation CANNOT fix geometry-intrinsic fails. The harbor-house 15-fail best layout contains long-thin cells that are useless whatever room usage is assigned — their width/proportion/crinkliness fails are shape-bound, not label slack. Two consequences: (1) do not over-claim collapse gains — only ~2-3 of that layout's fails are reclaimable relabel slack, the rest are geometry- or building-level bound; (2) the threshold objective must not be tuned to 'pass' a degenerate cell via a permissive room type — a metric-pass on a physically useless space is gaming, not a fix. Real remedies for these are geometry/topology search (cell shape) and circulation placement, filed separately, not the collapse."}
{"_type":"memory","key":"ld2-13-6-interior-o-seed-diagnostic-all","value":"ld2/§13.6 interior-O seed diagnostic: ALL crinkliness fails in the constructed bal+share seed are UNDER-exposed (crink\u003c0.62, landlocked rooms with no facade + no uncovered-O neighbour) — zero over-exposed sliver fails. So the erc crinkliness residual is genuine under-daylighting, validating the interior light-well premise. Default outside_divisor=6 was too sparse (null: harbor 147-\u003e142, crinkliness even rose). odiv=3 is the seed-optimal joint setting: harbor seed fails 147-\u003e129 (-18), maple 219-\u003e206 (-14), landlocked fails drop, at cost of more leaves (harbor +4, maple +8). Because it ADDS leaves it carries the §13.4 wash-out risk; A/B to convergence pending."}
{"_type":"memory","key":"programme-house-optimisation-result-2026-06-14-15","value":"Programme-house optimisation result (2026-06-14/15): best achievable is 1 fail (l1 wrong level, score ~0.005). 0 fails is geometrically impossible: l1 (min 27m²) must occupy ll (~23m²) at level 0, which eliminates the t3-adj-C provider; dividing ll into lll(l1)+llr(C) gives llr proportion ~6:1 (fails). Python memetic optimizer achieves 1 fail in 50k evals vs Perl optimiser's 2-3 fails. Winning topology: TWO C nodes at level 0 — ll(C) for t3-adj-C via geometric contact, rl(C) for staircase via tree-sibling adjacency to rrr(O). Best .dom: scratch/from-warmstart-fixed.dom and scratch/from-compound3-fixed.dom."}
{"_type":"memory","key":"unfold-strategy-for-shared-leaves-homemaker-py-8iv","value":"Unfold strategy for shared leaves (homemaker-py-8iv, resolved 2026-07-16): use the BALANCED GRID (operators._grow_balanced/_size_subtree_equal), NOT circulation-aware slicing. Slicing a shared leaf perpendicular to its access edge so every child touches the corridor was implemented + A/B-tested and LOST decisively (150k-eval warm-start polish from evolved-3M: slice 41 fails/3.5e-14 vs grid 25 fails/2.4e-09, grid ahead at every milestone). Reason: k rooms all touching one wall are intrinsically thin slices; that geometric debt (proportion/long/width) is unfixable without topology change, while the grid's squarer children let local search re-route access cheaply via level_retype/place_missing/level_fix. Lesson: at the sharing-\u003eno-sharing transition, prioritise squarer children and leave access to local search; do not reintroduce slicing in Schedule B (kpu)."}
{"_type":"memory","key":"warm-x0-initialization-bug-pattern-when-a-topology","value":"warm_x0 initialization bug pattern: when a topology operator explicitly sets division ratios on a newly-created node (e.g. compound_fix sets node.division=[0.25,0.25] for t3), parent.ratios has no entry for that node (it was a leaf). warm_x0 defaults it to 0.5, corrupting the inner loop's starting point and making the operator invisible to lex comparison. Fix: only propagate child ratios for nodes where the parent node was NOT already divided; stale hidden nodes revealed by structural mutations (swap flipping b.below) must NOT contribute their pre-writeback values. See driver.py lines 259-267 (fixed 2026-06-14)."}
{"_type":"memory","key":"correction-to-urb-fitness-bug-memory-bruno-2026","value":"CORRECTION to urb-fitness-bug memory (Bruno, 2026-06-12): 'C' is NOT a 'covered' type — Is_Covered is a geometric predicate (indoor space above). Urb's generic types are canonically UPPERCASE: C=circulation, O=outside, S=sahn (get_space_types qw/C O S/; corpus is 100% uppercase, never 'c'/'o' leaves). The mixed-case designs that fired the latent ratio_type first-match bug were created by homemaker's own operator type pool emitting lowercase 'c'/'o' — fixed: driver/operators now emit uppercase generics only, and class checks use t[0].lower() in 'cos'. The Urb class-sum patch stays as defensive hardening (zero impact on canonical designs). Native port (3y7/gnw): treat type classes case-insensitively, generics canonically uppercase."}
{"_type":"memory","key":"deceptive-valleys-in-topology-search-when-every-single","value":"Deceptive valleys in topology search: when every single-step mutation from a target state passes through a high-fail intermediary (e.g. level_fix displaces a room into 5+ new fails), a compound operator that atomically applies two coordinated changes can escape. Design compound operators to land on the low-fail state directly, bypassing the deceptive gradient. Programme-house example: level_compound_fix atomically moves the level-constrained room AND re-inserts the displaced room adjacent to C in one step (operators.py, 2026-06-14)."}
{"_type":"memory","key":"user-preference-bruno-this-is-a-fedora-system","value":"User preference (Bruno): this is a Fedora system — NEVER install Python packages via pip without asking first; always ask whether to install the rpm via dnf (e.g. python3-cma) before considering pip. Applies to any dependency additions."}
{"_type":"memory","key":"experiment-harness-gotcha-the-leaf-sharing-relaxed-objective","value":"Experiment harness gotcha: the leaf-sharing RELAXED objective (§13.3) is injected ONLY by monkeypatching fitness.load_config in the parent process (run_staged_search.py / probe scripts). This is parent-process-only and does NOT propagate into ProcessPoolExecutor workers (n_workers\u003e1), which re-import fitness fresh and score under the STRICT on-disk patterns.config -\u003e r.n_fails MISMATCH (worker strict vs parent relaxed re-score). ALL §13.x floor runs were therefore SERIAL. Any future PARALLEL leaf-sharing experiment will silently mis-score until leaf_sharing lives on disk/CLI (tracked: homemaker-py-x3b). The parallel driver itself is correct; both paths score via load_config(programme_dir)."}
{"_type":"memory","key":"never-use-corpus-filenames-candidate-001-dom-candidate","value":"Never use corpus filenames (candidate-001.dom, candidate-002.dom, generated.dom, init.dom, etc.) as --output targets when running experiments. These are test fixtures. Always write experimental outputs to scratch/ or a timestamped path. Lesson from 2026-06-14: warm-start runs overwrote candidate-001/002.dom and broke graph tests."}
{"_type":"memory","key":"run-to-run-reproducibility-in-homemaker-layout-serial","value":"Run-to-run reproducibility in homemaker-layout: serial search (workers=1) is byte-for-byte deterministic; parallel (workers\u003e1) is now deterministic too AFTER fixing driver._run_batch to admit futures in submission order (was as_completed/completion order, bug xcy). Reproducibility holds only for a FIXED worker count — serial vs parallel differ because children-per-iteration is 1 vs n_workers (different batch granularity), which is expected, not a bug. The constructive seeder was NEVER nondeterministic: _assign_adjacency_aware has unique idx tiebreaks; comparing topologies with Python builtin hash() of the signature STRING is invalid (PYTHONHASHSEED salts str hashing per process) — use a stable hash (sha1) or genome.signature equality."}
{"_type":"memory","key":"urb-oracle-nondeterminism-urb-fitness-pl-output-varies","value":"Urb oracle nondeterminism: urb-fitness.pl output varies run-to-run from Perl hash-order randomisation — .fails line ORDER shuffles (compare sorted, use oracle.Score.fail_lines) and the score float can flip by ~1 ULP (compare with math.isclose rel_tol=1e-12, never ==). Not a batching artifact; affects single runs too. Matters for the Phase 3 native-fitness parity gate (homemaker-py-uxz)."}
{"_type":"memory","key":"collapse-global-94g-and-any-label-usage-optimisation","value":"collapse_global (94g) and any label/usage optimisation CANNOT fix geometry-intrinsic fails. The harbor-house 15-fail best layout contains long-thin cells that are useless whatever room usage is assigned — their width/proportion/crinkliness fails are shape-bound, not label slack. Two consequences: (1) do not over-claim collapse gains — only ~2-3 of that layout's fails are reclaimable relabel slack, the rest are geometry- or building-level bound; (2) the threshold objective must not be tuned to 'pass' a degenerate cell via a permissive room type — a metric-pass on a physically useless space is gaming, not a fix. Real remedies for these are geometry/topology search (cell shape) and circulation placement, filed separately, not the collapse."}
{"_type":"memory","key":"collapse-global-s-jacobi-adjacency-relaxation-homemaker-py","value":"collapse_global's Jacobi adjacency relaxation (homemaker-py-94g) is a synchronous per-round linear-assignment re-solve, which can 2-cycle indefinitely between two labellings that each satisfy ZERO adjacency requirements even though a permutation satisfying ALL of them exists -- proven on a minimal 4-cell chain (p1-q1-p2-q2, two disjoint adjacency pairs p1\u003c-\u003ep2/q1\u003c-\u003eq2) in test_two_opt_polish_escapes_jacobi_plateau. homemaker-py-9wi added Fitness._two_opt_adjacency_polish: a same-level pairwise-swap local search run after the Jacobi fixpoint, gated behind collapse_global(local_search=True) (default off, exposed as homemaker-collapse --local-search). Monotone by construction (a swap is kept only if it strictly increases total reward). Empirically on the 11 harbor-house evolved-*.dom/3m.dom/materialised-3M.dom layouts: 10 matched Jacobi-only exactly, 0 regressed, and evolved-anneal-3M.dom improved 21-\u003e19 fails (fixed a genuine mutual da1\u003c-\u003ek1 adjacency miss the Jacobi loop couldn't reach)."}
{"_type":"memory","key":"island-model-psk-14-is-a-null-priming","value":"Island model (psk, §14) is a NULL: priming a population from N converged independent elites + crossover-heavy migration does not beat best-of-N at equal total budget (maple island 124 vs control 116). The child_probe instrument shows WHY: area-matched crossover across independently-converged elites almost never synthesizes (1-3 of ~64 children beat the better parent, max drop 2-5) because the slicing encoding is non-canonical (9gp), so splices are disruptive not combinatorial. Search-machinery null #3 after graded-objective and niching/restarts; residual stays geometry/shape-bound."}
{"_type":"memory","key":"proportion-aware-constructive-seeding-leu-2-12-2","value":"Proportion-aware constructive seeding (leu.2/§12.2): sizing seed cuts from target AREAS only regresses (thin slivers wreck aspect); you must ALSO pick each cut's rotation for child squareness. It is a convergence ACCELERATOR via a deeper local optimum around the constructed topology: wins where that topology is roughly right and budget is scarce (harbor -13%, maple -10% at 20k evals) but DELAYS small programmes where the seed must be restructured by undivide (programme-house regresses at fixed budget, yet reaches the floor given budget - speed, not asymptote). Default-on. Also: n_storeys must honour storey_minimum, not just level: keys (programme-house storey_minimum:2, all rooms level:0 - was seeded 1 storey short; cq1)."}
{"_type":"memory","key":"urb-fitness-bug-found-fixed-2026-06-12","value":"Urb fitness bug found+fixed 2026-06-12 (patch in /home/bruno/src/urb, uncommitted): ProgrammeDriven.pm ratio_o/ratio_type grepped case-insensitively over the ratios hash and took the FIRST key — nondeterministic (x4.5 score swings) for designs with mixed-case type classes (both 'c' circulation and 'C' covered). Fixed to SUM the class (matches Is_Circulation//Is_Outside semantics); 35/35 corpus scores unchanged. CRITICAL for homemaker-py-3y7/gnw: the native port must implement class-SUM ratios. Building.pm has the same unpatched pattern (site-driven path, not used by our oracle). Also: the memetic search reward-hacked this bug before the fix — search results predating it are noise artifacts."}
{"_type":"memory","key":"adjacency-in-binary-slicing-tree-is-structural-not","value":"Adjacency in binary slicing tree is structural, not geometric: the inner-loop NM cannot fix topological adjacency failures. Two paths exist: (1) tree-sibling adjacency — a node is adjacent to its sibling in the tree; (2) cross-zone geometric adjacency — leaves from different subtrees that happen to share a boundary. Staircase/adjacency fails require a topology mutation that changes which nodes are siblings or which zones touch. This was proved empirically on programme-house: staircase fail from rot=0 layout could not be fixed by NM but was fixed by level_retype creating a two-C topology (2026-06-14/15)."}
{"_type":"memory","key":"cli-tool-style-prefer-python-m-homemaker-module","value":"CLI tool style: prefer python -m homemaker.module --parameters pattern, installable via pip install -e . with pyproject.toml entry_points. Not standalone bin/ scripts."}
{"_type":"memory","key":"strategy-decision-2026-06-12-bruno-occlusion-daylight","value":"Strategy decision 2026-06-12 (Bruno): occlusion/daylight is ORTHOGONAL to building a scalable optimiser. Disable it in Urb (env flag, homemaker-py-gp2) rather than port it; native fitness uses simple crinkliness (illumination factor = 1); rebuild occlusion in Python only after optimisation is fully native (homemaker-py-2g5, now P4). Consequence: all scores change when the flag flips — re-baseline corpus/.score, DESIGN \\$4.5 gains, gate bars at one clean boundary AFTER homemaker-py-1p0 closes; Phase-2 urb-evolve benchmark must run with the same flag."}
{"_type":"memory","key":"experiment-seeding-pitfall-run-search-scaled-py-s","value":"Experiment seeding pitfall: run_search_scaled.py's default PH_SEED (c964…dom) is a FINISHED programme-house design — passing it warm-starts and floors at ~3 fails, NOT a blank-slate topology search. For blank-slate runs comparable to §11.5/§11.6 baselines, seed from examples/programme-house/init.dom (a bare undivided plot; driver bootstrap auto-triggers only on bare plots). Bit the 6zy sweep — first pass used c964 and falsely showed 3-fail floor across the whole grid."}
{"_type":"memory","key":"homemaker-py-pythonpath-set-pythonpath-home-bruno-src","value":"homemaker-layout PYTHONPATH: package installed as 'homemaker-layout' via pip install -e . so 'import homemaker_layout' works from anywhere without PYTHONPATH. For running tests use 'python -m pytest' from project root /home/bruno/src/homemaker-layout (pyproject.toml adds src/ automatically). Never try pip show homemaker — that's the old homemaker-addon conflict."}
{"_type":"memory","key":"ld2-13-6-interior-o-seed-diagnostic-all","value":"ld2/§13.6 interior-O seed diagnostic: ALL crinkliness fails in the constructed bal+share seed are UNDER-exposed (crink\u003c0.62, landlocked rooms with no facade + no uncovered-O neighbour) — zero over-exposed sliver fails. So the erc crinkliness residual is genuine under-daylighting, validating the interior light-well premise. Default outside_divisor=6 was too sparse (null: harbor 147-\u003e142, crinkliness even rose). odiv=3 is the seed-optimal joint setting: harbor seed fails 147-\u003e129 (-18), maple 219-\u003e206 (-14), landlocked fails drop, at cost of more leaves (harbor +4, maple +8). Because it ADDS leaves it carries the §13.4 wash-out risk; A/B to convergence pending."}
{"_type":"memory","key":"9o5-multi-use-leaves-is-path-a-superposition","value":"9o5 multi-use leaves is path (a) — superposition as SEARCH RELAXATION that COLLAPSES to specific usage at the end, NOT path (b) loose-fit/no-collapse. Bruno's intent: codes with SIMILAR leaf requirements form an interchangeable equivalence class; during evolution the solver doesn't commit which leaf serves which specific usage (smoother landscape, no fighting over exact leaf usage); at the end the layout is CONDENSED to specific usages by brute-forcing the in-class assignment (3 interchangeable usages over 3 leaves = 3! = 6 combinations to check, pick best). 'Derive automatically' compatibility = requirement-similarity grouping. This reverses the issue's stated 'path b preferred' note."}

19
DESIGN.md
View file

@ -1533,6 +1533,25 @@ pays **end-to-end**.
determinism fix (to even see sub-±3 effects) or a representational change beyond
the slicing tree — not another seed/search tweak at this scale.
- *§12.3 re-run at fixed worker count — CONFIRMED, no new run needed
(`homemaker-py-h10`, 2026-07-30).* §12.4's own writeup flagged the §12.3
reassoc/feas negatives (+3.3/+4.0) as sub-±3-adjacent and asked for a re-run
"at a single fixed worker count" post-fix, since they predate the
completion-order determinism fix above. Checked before re-running the full
8.3-hour sweep: `experiments/run_9gp_ab.sh` invokes `run_staged_search.py`,
which never threads a worker count through to `driver.search_staged` —
every §12.3 arm therefore already ran at `n_workers=1` (serial), the one
mode §12.4 itself already proved "was already byte-for-byte reproducible"
even *before* the fix (the bug was in `ProcessPoolExecutor`
`as_completed` ordering, parallel-only; serial has no futures to reorder).
Confirmed empirically too: re-running one arm (harbor-house seed 0,
baseline config, budget 300) twice back-to-back reproduced identical fail
counts at every logged checkpoint. So the §12.3 table was already measured
at a fixed (and the most reproducible available) worker count — the
determinism fix changes nothing for it. **Verdict stands as CONFIRMED-NULL**
without re-spending the ~8 core-hours a full re-run would cost; upgrades
§12.3's negative from "should be re-run" to "already valid as measured."
## 13. Phase 8 — lowering the geometry/shape floor (`homemaker-py-erc`)
Phase 8 runs DIAGNOSTICS FIRST to decide *which* floor-lowering lever to invest