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

homemaker-py-wkh: DP-exact hard pre-filter for driver.py's shape-feasibility prune

Browse source 
Adds shapecurve.is_feasible() (a non-mutating refactor of solve()'s check
phase) and a shapecurve_prune flag composing the DP's exact feasible/
infeasible verdict with operators.predicted_shape_fails' existing heuristic
prune: DP-feasible vetoes a heuristic prune outright; DP-infeasible only
hard-prunes when the incumbent already has zero total fails (exact, since
infeasible proves the shape-fail floor is >=1); otherwise defers unchanged
to today's heuristic threshold. Conservative by design since a wrong prune
is unrecoverable.

Validated 0/400 false negatives across two structurally distinct plots
(harbor-house-l0 + a newly-added programme-house sweep, the first genuinely
non-rectangular plot this DP has been checked against). The real
driver.search A/B on harbor-house-l0 measured NULL (byte-identical off/on)
for a root-caused, pre-existing reason: predicted_shape_fails rarely
triggers organically at this scale, so neither new branch had an opening to
fire -- not a defect in this change. Full writeup: DESIGN.md §37.5.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LSwQwpEaHFBkeVSDDWd75S
This commit is contained in:
Bruno Postle 2026-08-03 21:10:28 +01:00
parent 229c3c6933
commit 467c732260
9 changed files with 453 additions and 45 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

38
.beads/issues.jsonl
View file

@ -28,7 +28,7 @@
{"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-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-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-koo","title":"Multi-storey (below-link) support for the shape-curve DP","description":"homemaker-py-6xh item 3 (DESIGN.md §37.2/§37.4). src/homemaker_layout/shapecurve.py's solve()/realise() writes division on every divided node under a single level_root unconditionally -- it has no notion of upper-storey below-inherited (wall-stacked) fixed splits (see solver.free_branches: a branch is free only when b.below is None or not b.below.divided). shapecurve.eligible() currently guards this by requiring len(dom.levels(root)) == 1, so the DP warm-start (homemaker-py-6xh) never fires on multi-storey topologies -- which is most real programmes (e.g. examples/programme-house has storey_minimum=2, examples/harbor-house is multi-storey; only the purpose-built examples/harbor-house-l0 de-risk variant is single-storey). Needs: generalise the DP to run bottom-up per storey, treating below-inherited-and-divided branches as FIXED (their (w,h) contribution comes from the level below's already-realised geometry, not chosen by this level's DP) while still composing correctly through them to size the level's own free branches. Validate against the full (multi-storey) examples/harbor-house, the DP's original but not-yet-attempted target.","status":"open","priority":2,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-08-03T17:29:43Z","created_by":"Bruno Postle","updated_at":"2026-08-03T17:29:43Z","dependencies":[{"issue_id":"homemaker-py-koo","depends_on_id":"homemaker-py-2g7","type":"parent-child","created_at":"2026-08-03T18:31:48Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} {"id":"homemaker-py-koo","title":"Multi-storey (below-link) support for the shape-curve DP","description":"homemaker-py-6xh item 3 (DESIGN.md §37.2/§37.4). src/homemaker_layout/shapecurve.py's solve()/realise() writes division on every divided node under a single level_root unconditionally -- it has no notion of upper-storey below-inherited (wall-stacked) fixed splits (see solver.free_branches: a branch is free only when b.below is None or not b.below.divided). shapecurve.eligible() currently guards this by requiring len(dom.levels(root)) == 1, so the DP warm-start (homemaker-py-6xh) never fires on multi-storey topologies -- which is most real programmes (e.g. examples/programme-house has storey_minimum=2, examples/harbor-house is multi-storey; only the purpose-built examples/harbor-house-l0 de-risk variant is single-storey). Needs: generalise the DP to run bottom-up per storey, treating below-inherited-and-divided branches as FIXED (their (w,h) contribution comes from the level below's already-realised geometry, not chosen by this level's DP) while still composing correctly through them to size the level's own free branches. Validate against the full (multi-storey) examples/harbor-house, the DP's original but not-yet-attempted target.","status":"open","priority":2,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-08-03T17:29:43Z","created_by":"Bruno Postle","updated_at":"2026-08-03T17:29:43Z","dependencies":[{"issue_id":"homemaker-py-koo","depends_on_id":"homemaker-py-2g7","type":"parent-child","created_at":"2026-08-03T18:31:48Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-wkh","title":"DP-exact hard pre-filter: replace/augment predicted_shape_fails with shapecurve's boolean infeasibility","description":"homemaker-py-6xh item 1 (DESIGN.md §37.2/§37.4). The shapecurve DP (src/homemaker_layout/shapecurve.py, promoted from experiments/shapecurve_spike.py) gives an EXACT feasible/infeasible verdict for the size/width/proportion family, currently wired only as an NM warm-start (safe: never prunes). operators.predicted_shape_fails' threshold-based prune (driver._evaluate, feasibility_max_shape_fails/best_n_fails) still uses the older heuristic-count proxy. Using shapecurve.solve's infeasible verdict as an ADDITIONAL/replacement hard-prune signal would be stronger (exact, not a graduated heuristic) but riskier: unlike a bad warm-start, a wrong prune permanently discards a topology that could have beaten the incumbent. DESIGN.md §37.2 measured 0/200 false negatives (DP infeasible, NM reaches 0 anyway) on harbor-house-l0, but that is not a proven bound (the rectangle-vs-skew-quad approximation is a known ~7-12% error source, §37.2). Needs: (a) a design for how the DP's boolean signal composes with the existing pred\u003ethreshold\u0026\u0026pred\u003e=best_n_fails guard, (b) a false-negative-risk validation before enabling by default (a larger/less-rectangular topology sweep than the 200-topology harbor-house-l0 one), (c) a driver.search A/B (evals-to-N-hard-fails) against today's predicted_shape_fails-only filter.","status":"open","priority":2,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-08-03T17:29:13Z","created_by":"Bruno Postle","updated_at":"2026-08-03T17:29:13Z","dependencies":[{"issue_id":"homemaker-py-wkh","depends_on_id":"homemaker-py-2g7","type":"parent-child","created_at":"2026-08-03T18:31:46Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} {"id":"homemaker-py-wkh","title":"DP-exact hard pre-filter: replace/augment predicted_shape_fails with shapecurve's boolean infeasibility","description":"homemaker-py-6xh item 1 (DESIGN.md §37.2/§37.4). The shapecurve DP (src/homemaker_layout/shapecurve.py, promoted from experiments/shapecurve_spike.py) gives an EXACT feasible/infeasible verdict for the size/width/proportion family, currently wired only as an NM warm-start (safe: never prunes). operators.predicted_shape_fails' threshold-based prune (driver._evaluate, feasibility_max_shape_fails/best_n_fails) still uses the older heuristic-count proxy. Using shapecurve.solve's infeasible verdict as an ADDITIONAL/replacement hard-prune signal would be stronger (exact, not a graduated heuristic) but riskier: unlike a bad warm-start, a wrong prune permanently discards a topology that could have beaten the incumbent. DESIGN.md §37.2 measured 0/200 false negatives (DP infeasible, NM reaches 0 anyway) on harbor-house-l0, but that is not a proven bound (the rectangle-vs-skew-quad approximation is a known ~7-12% error source, §37.2). Needs: (a) a design for how the DP's boolean signal composes with the existing pred\u003ethreshold\u0026\u0026pred\u003e=best_n_fails guard, (b) a false-negative-risk validation before enabling by default (a larger/less-rectangular topology sweep than the 200-topology harbor-house-l0 one), (c) a driver.search A/B (evals-to-N-hard-fails) against today's predicted_shape_fails-only filter.","notes":"2026-08-03: Shipped the DP-exact hard pre-filter, DESIGN.md §37.5. Full\ndetails there; summary:\n\n- shapecurve.is_feasible() (new, non-mutating refactor of solve()'s check\n phase) + shapecurve_prune flag in driver._evaluate/search, threaded\n through to `homemaker-evolve --shapecurve-prune` (default off, mirrors\n --shapecurve-warmstart). Composition: DP-feasible vetoes a heuristic\n prune outright (skips predicted_shape_fails entirely); DP-infeasible only\n hard-prunes when the incumbent already has 0 total fails (exact, since\n infeasible proves the shape-fail floor \u003e=1); otherwise defers unchanged\n to the existing predicted_shape_fails threshold. Conservative by design\n per the bead's own risk framing (a wrong prune is unrecoverable, unlike a\n bad warm-start).\n- Validation (bead item b): pointed experiments/validate_shapecurve.py at\n the promoted product module (was still validating the frozen spike) and\n gave it a programme_dir CLI arg; ran the same 200-topology protocol\n against examples/programme-house (a genuinely skewed, non-axis-aligned\n plot, not just a rotated harbor-house-l0): 200/200 agreement, 0 false\n positives, 0 false negatives, 87.4x speedup. Combined with §37.2's\n original 200 on harbor-house-l0: 0/400 false negatives across two\n structurally distinct plots.\n- A/B (bead item c): experiments/ab_shapecurve_prune.py, same protocol as\n 6xh's warm-start A/B (harbor-house-l0, budget=2000, seeds 0-4). Result:\n byte-identical off/on across all 5 seeds -- NULL, not a regression.\n Instrumented root cause: on this benchmark predicted_shape_fails itself\n (pre-existing 9gp.1, not this bead's code) rarely reaches best_n_fails\n organically -- tests/test_driver.py's own test_feasibility_filter_\n prunes_cheaply already had to force it to 999 to observe any real prune\n -- so neither the veto nor the exact-prune branch had an opening to fire\n (spied: 17/17 DP checks infeasible, incumbent total fails never reached\n 0). Not a wkh defect; 9gp.1 is documented as a \"scaling lever\", expected\n to matter at larger programmes/leaf counts than this benchmark, not here.\n\nTests: tests/test_shapecurve.py (+1), tests/test_driver.py (+4). Full\nsuite: 393 passed.\n\nFollow-up (not blocking this close, noted in DESIGN.md §37.5): re-run the\nA/B at a scale where predicted_shape_fails organically prunes to see wkh's\nmarginal value -- the more direct route there is homemaker-py-koo\n(multi-storey) and homemaker-py-tym (leaf-sharing), since today's DP\neligibility already excludes the real \u003e=2-storey, leaf-sharing-default\nprogrammes (programme-house, harbor-house) this would need to be measured\non.","status":"in_progress","priority":2,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-08-03T17:29:13Z","created_by":"Bruno Postle","updated_at":"2026-08-03T20:08:23Z","started_at":"2026-08-03T18:16:35Z","dependencies":[{"issue_id":"homemaker-py-wkh","depends_on_id":"homemaker-py-2g7","type":"parent-child","created_at":"2026-08-03T18:31:46Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-6xh","title":"Wire shapecurve DP prototype into driver.py as a real pre-filter + NM warm-start","description":"homemaker-py-2g7.4's prototype (experiments/shapecurve_spike.py, DESIGN.md\n§37.2) validated PASS on harbor-house-l0 (99% agreement vs shape-fail-only NM\nover 200 random topologies, 93.6x speedup at grid_n=150, 0 false negatives).\nIt is not yet wired into the product — it's a reference spike only, same\nstatus as experiments/autodiff_spike.py (§34).\n\nTo productionise per the original plan (DESIGN.md §37 point 2):\n- Replace/augment operators.predicted_shape_fails with the DP as driver.py's\n real per-child pre-filter (a single-sample heuristic today; the DP gives an\n exact yes/no plus a realizing ratio point).\n- Warm-start innerloop.optimise's NM from the DP's realised ratios instead of\n (or in addition to) the current proportion-aware target-geometry seed.\n- Multi-storey support: the DP only walks one level's leaves currently;\n below-linked nodes (wall-stacking across storeys) aren't modelled.\n- leaf_sharing/co_type target-adjustment: not modelled in leaf_constraints,\n needed for any programme that uses either (harbor-house-l0 doesn't).\n- Consider replacing the bounding-box leaf approximation with true skew-quad\n polygon algebra to remove the ~7-12% area approximation error identified\n as the root cause of both measured false positives (§37.2) -- or at least\n characterise it on a LESS rectangular plot than harbor-house-l0's\n near-rectangular trapezoid, where the error is likely worse.\n- A/B against the real driver.search: does DP-pre-filter + warm-start beat\n today's predicted_shape_fails + cold/proportion-aware start on wall-clock\n to N hard fails, on harbor-house (full) and/or a less-rectangular plot?","notes":"2026-08-03: Shipped NM warm-start (item 2) + a scoped A/B (item 5), left\nin_progress -- 3 of 5 description items deliberately deferred to new\ntracked beads (see below). Full details + measured numbers: DESIGN.md\n§37.4.\n\nWhat shipped: promoted experiments/shapecurve_spike.py into\nsrc/homemaker_layout/shapecurve.py (fixed a latent numpy.float64-in-division\nbug caught by round-tripping through dom.dumps in the new tests -- the spike\nnever round-tripped and so never caught it). Added shapecurve.eligible()\n(single storey, no leaf_sharing/superpose/max_share/multi_use). Wired into\ndriver._evaluate as an NM warm-start only (never a prune) behind\nshapecurve_warmstart=False default, threaded through driver.search and\nexposed as `homemaker-evolve --shapecurve-warmstart`. A/B\n(experiments/ab_shapecurve_warmstart.py) on harbor-house-l0, budget=2000,\n5 seeds: mean total-fails 16.6 (on) vs 19.6 (off), ~3.5x mean fitness\nimprovement; mean hard-fail count alone was a noise-level wash (4.6 vs 4.4\nat n=5). Tests: tests/test_shapecurve.py (4), tests/test_driver.py (+3).\nFull suite 388 passed.\n\nDeferred to new tracked beads (children of 2g7, per the epic's own\ndependency ordering):\n- homemaker-py-wkh: DP-exact hard pre-filter (item 1) -- replacing\n predicted_shape_fails' heuristic threshold with the DP's exact\n infeasibility verdict. Needed to actually chase \"evals to N hard fails\"\n rather than just improve soft-fail/fitness convergence.\n- homemaker-py-koo: multi-storey (below-link) DP support (item 3) --\n without this, the warm-start never fires on programme-house\n (storey_minimum=2) or full harbor-house, only the purpose-built\n single-storey harbor-house-l0.\n- homemaker-py-tym: leaf_sharing/co_type modelling (item 4) -- without this,\n the warm-start never fires when leaf_sharing=True, which is\n driver.search's own default.\n- homemaker-py-ekc: true skew-quad polygon algebra (the §37.2-quantified\n ~7-12% rectangle-approximation error) -- not a new bead-description item,\n but the explicit \"consider replacing the bounding-box leaf approximation\"\n bullet.\n\nNet: 6xh's own acceptance (a real evals-to-N-hard-fails win over\npredicted_shape_fails + cold/proportion-aware start) is NOT yet met --\ntoday's result is a safe, positive-but-partial step (soft-fail/fitness\nconvergence, not hard-fail convergence, and only on the single-storey\nno-sharing envelope). Keeping 6xh in_progress rather than closing it.","status":"in_progress","priority":2,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-08-02T22:39:46Z","created_by":"Bruno Postle","updated_at":"2026-08-03T17:40:52Z","started_at":"2026-08-03T15:51:29Z","dependency_count":0,"dependent_count":0,"comment_count":0} {"id":"homemaker-py-6xh","title":"Wire shapecurve DP prototype into driver.py as a real pre-filter + NM warm-start","description":"homemaker-py-2g7.4's prototype (experiments/shapecurve_spike.py, DESIGN.md\n§37.2) validated PASS on harbor-house-l0 (99% agreement vs shape-fail-only NM\nover 200 random topologies, 93.6x speedup at grid_n=150, 0 false negatives).\nIt is not yet wired into the product — it's a reference spike only, same\nstatus as experiments/autodiff_spike.py (§34).\n\nTo productionise per the original plan (DESIGN.md §37 point 2):\n- Replace/augment operators.predicted_shape_fails with the DP as driver.py's\n real per-child pre-filter (a single-sample heuristic today; the DP gives an\n exact yes/no plus a realizing ratio point).\n- Warm-start innerloop.optimise's NM from the DP's realised ratios instead of\n (or in addition to) the current proportion-aware target-geometry seed.\n- Multi-storey support: the DP only walks one level's leaves currently;\n below-linked nodes (wall-stacking across storeys) aren't modelled.\n- leaf_sharing/co_type target-adjustment: not modelled in leaf_constraints,\n needed for any programme that uses either (harbor-house-l0 doesn't).\n- Consider replacing the bounding-box leaf approximation with true skew-quad\n polygon algebra to remove the ~7-12% area approximation error identified\n as the root cause of both measured false positives (§37.2) -- or at least\n characterise it on a LESS rectangular plot than harbor-house-l0's\n near-rectangular trapezoid, where the error is likely worse.\n- A/B against the real driver.search: does DP-pre-filter + warm-start beat\n today's predicted_shape_fails + cold/proportion-aware start on wall-clock\n to N hard fails, on harbor-house (full) and/or a less-rectangular plot?","notes":"2026-08-03: Shipped NM warm-start (item 2) + a scoped A/B (item 5), left\nin_progress -- 3 of 5 description items deliberately deferred to new\ntracked beads (see below). Full details + measured numbers: DESIGN.md\n§37.4.\n\nWhat shipped: promoted experiments/shapecurve_spike.py into\nsrc/homemaker_layout/shapecurve.py (fixed a latent numpy.float64-in-division\nbug caught by round-tripping through dom.dumps in the new tests -- the spike\nnever round-tripped and so never caught it). Added shapecurve.eligible()\n(single storey, no leaf_sharing/superpose/max_share/multi_use). Wired into\ndriver._evaluate as an NM warm-start only (never a prune) behind\nshapecurve_warmstart=False default, threaded through driver.search and\nexposed as `homemaker-evolve --shapecurve-warmstart`. A/B\n(experiments/ab_shapecurve_warmstart.py) on harbor-house-l0, budget=2000,\n5 seeds: mean total-fails 16.6 (on) vs 19.6 (off), ~3.5x mean fitness\nimprovement; mean hard-fail count alone was a noise-level wash (4.6 vs 4.4\nat n=5). Tests: tests/test_shapecurve.py (4), tests/test_driver.py (+3).\nFull suite 388 passed.\n\nDeferred to new tracked beads (children of 2g7, per the epic's own\ndependency ordering):\n- homemaker-py-wkh: DP-exact hard pre-filter (item 1) -- replacing\n predicted_shape_fails' heuristic threshold with the DP's exact\n infeasibility verdict. Needed to actually chase \"evals to N hard fails\"\n rather than just improve soft-fail/fitness convergence.\n- homemaker-py-koo: multi-storey (below-link) DP support (item 3) --\n without this, the warm-start never fires on programme-house\n (storey_minimum=2) or full harbor-house, only the purpose-built\n single-storey harbor-house-l0.\n- homemaker-py-tym: leaf_sharing/co_type modelling (item 4) -- without this,\n the warm-start never fires when leaf_sharing=True, which is\n driver.search's own default.\n- homemaker-py-ekc: true skew-quad polygon algebra (the §37.2-quantified\n ~7-12% rectangle-approximation error) -- not a new bead-description item,\n but the explicit \"consider replacing the bounding-box leaf approximation\"\n bullet.\n\nNet: 6xh's own acceptance (a real evals-to-N-hard-fails win over\npredicted_shape_fails + cold/proportion-aware start) is NOT yet met --\ntoday's result is a safe, positive-but-partial step (soft-fail/fitness\nconvergence, not hard-fail convergence, and only on the single-storey\nno-sharing envelope). Keeping 6xh in_progress rather than closing it.","status":"in_progress","priority":2,"issue_type":"feature","assignee":"Bruno Postle","owner":"bruno@postle.net","created_at":"2026-08-02T22:39:46Z","created_by":"Bruno Postle","updated_at":"2026-08-03T17:40:52Z","started_at":"2026-08-03T15:51:29Z","dependency_count":0,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-2g7.9","title":"Parallel best-of-N + racing harness (use all cores, kill stragglers early)","description":"§14 measured islands \u003c= best-of-N, and the 3M runs used workers=1-2 on a 4-core box — independent seeds are the proven shape and we are not even using the local machine. Build a harness: launch N independent search_staged seeds across all cores (processes, not threads — mind the cvw id()-keyed cache bug), checkpoint fail-counts periodically, successively halve (hyperband-style: kill runs above median hard-fail count at each rung, reallocate budget to survivors). Fix/respect homemaker-py-b8g (parallel non-determinism) and homemaker-py-cvw first or work around with process isolation. This multiplies whatever eval cost the shape-curve DP issue achieves; on its own it is a free 4x locally and scales to any box. Report best + variance across seeds (the seed-variance in §12-§13 tables is huge — 78 vs 97 same config — so best-of-N is worth several levers combined).","acceptance_criteria":"harness runs N=16 seeds on 4 cores with racing; at equal total native-eval budget beats the single-seed mean on harbor by at least the observed seed spread; deterministic per-seed replay","status":"open","priority":2,"issue_type":"task","owner":"bruno@postle.net","created_at":"2026-08-02T09:15:58Z","created_by":"Bruno Postle","updated_at":"2026-08-02T09:15:58Z","dependencies":[{"issue_id":"homemaker-py-2g7.9","depends_on_id":"homemaker-py-2g7","type":"parent-child","created_at":"2026-08-02T10:15:58Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-2g7.9","depends_on_id":"homemaker-py-b8g","type":"blocks","created_at":"2026-08-02T10:16:16Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-2g7.9","depends_on_id":"homemaker-py-cvw","type":"blocks","created_at":"2026-08-02T10:16:15Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":2,"dependent_count":0,"comment_count":0} {"id":"homemaker-py-2g7.9","title":"Parallel best-of-N + racing harness (use all cores, kill stragglers early)","description":"§14 measured islands \u003c= best-of-N, and the 3M runs used workers=1-2 on a 4-core box — independent seeds are the proven shape and we are not even using the local machine. Build a harness: launch N independent search_staged seeds across all cores (processes, not threads — mind the cvw id()-keyed cache bug), checkpoint fail-counts periodically, successively halve (hyperband-style: kill runs above median hard-fail count at each rung, reallocate budget to survivors). Fix/respect homemaker-py-b8g (parallel non-determinism) and homemaker-py-cvw first or work around with process isolation. This multiplies whatever eval cost the shape-curve DP issue achieves; on its own it is a free 4x locally and scales to any box. Report best + variance across seeds (the seed-variance in §12-§13 tables is huge — 78 vs 97 same config — so best-of-N is worth several levers combined).","acceptance_criteria":"harness runs N=16 seeds on 4 cores with racing; at equal total native-eval budget beats the single-seed mean on harbor by at least the observed seed spread; deterministic per-seed replay","status":"open","priority":2,"issue_type":"task","owner":"bruno@postle.net","created_at":"2026-08-02T09:15:58Z","created_by":"Bruno Postle","updated_at":"2026-08-02T09:15:58Z","dependencies":[{"issue_id":"homemaker-py-2g7.9","depends_on_id":"homemaker-py-2g7","type":"parent-child","created_at":"2026-08-02T10:15:58Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-2g7.9","depends_on_id":"homemaker-py-b8g","type":"blocks","created_at":"2026-08-02T10:16:16Z","created_by":"Bruno Postle","metadata":"{}"},{"issue_id":"homemaker-py-2g7.9","depends_on_id":"homemaker-py-cvw","type":"blocks","created_at":"2026-08-02T10:16:15Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":2,"dependent_count":0,"comment_count":0}
{"id":"homemaker-py-2g7.7","title":"LLM repair operator at stagnation (dom+fails -\u003e targeted compound edits)","description":"Generalize the §4.10 lesson: deceptive valleys are crossed by COMPOUND edits (move room + re-home displaced room + fix ratios atomically), which we currently hand-code one per valley (mutate_level_compound_fix). Our fail messages are semantically rich and localized ('me1 on wrong level', 'level 1 not connected', '0/rlrlr proportion') and the .dom is readable — ideal LLM input. Loop: on stagnation (no fail-tier improvement for N evals), serialize best individual + .fails + programme summary -\u003e LLM proposes 3-5 multi-step repairs as structured edit scripts (a small DSL over existing operator primitives: swap/divide/retype/rotate with explicit paths — NOT freeform dom text, so proposals are always well-formed) -\u003e apply, inner-loop, lex-accept as usual. Native fitness disposes; a bad proposal costs one child budget. Cost discipline: one LLM call ~ thousands of native evals, so plateau-only, cache by (signature, fails) key. Benchmark: the 3M-run best sat on 'level 0 not connected' + 'me1 on wrong level' for \u003e1M evals — moves a plan-reader fixes in one edit. Use claude via API (see claude-api skill); temperature\u003e0 for diverse proposals. Later extension (separate issue): AlphaEvolve-style operator-code synthesis using our existing A/B harness as the evaluator.","acceptance_criteria":"on the evolved-3M-nols-3 15-fail plateau seed: repair loop reduces hard-fail count where 1M+ blind evals did not, within \u003c=20 LLM calls; edit-DSL rejects malformed proposals; A/B at equal native-eval budget shows strictly better final fails on \u003e=2/3 seeds","status":"open","priority":2,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-08-02T09:15:54Z","created_by":"Bruno Postle","updated_at":"2026-08-02T09:15:54Z","dependencies":[{"issue_id":"homemaker-py-2g7.7","depends_on_id":"homemaker-py-2g7","type":"parent-child","created_at":"2026-08-02T10:15:53Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":0,"dependent_count":1,"comment_count":0} {"id":"homemaker-py-2g7.7","title":"LLM repair operator at stagnation (dom+fails -\u003e targeted compound edits)","description":"Generalize the §4.10 lesson: deceptive valleys are crossed by COMPOUND edits (move room + re-home displaced room + fix ratios atomically), which we currently hand-code one per valley (mutate_level_compound_fix). Our fail messages are semantically rich and localized ('me1 on wrong level', 'level 1 not connected', '0/rlrlr proportion') and the .dom is readable — ideal LLM input. Loop: on stagnation (no fail-tier improvement for N evals), serialize best individual + .fails + programme summary -\u003e LLM proposes 3-5 multi-step repairs as structured edit scripts (a small DSL over existing operator primitives: swap/divide/retype/rotate with explicit paths — NOT freeform dom text, so proposals are always well-formed) -\u003e apply, inner-loop, lex-accept as usual. Native fitness disposes; a bad proposal costs one child budget. Cost discipline: one LLM call ~ thousands of native evals, so plateau-only, cache by (signature, fails) key. Benchmark: the 3M-run best sat on 'level 0 not connected' + 'me1 on wrong level' for \u003e1M evals — moves a plan-reader fixes in one edit. Use claude via API (see claude-api skill); temperature\u003e0 for diverse proposals. Later extension (separate issue): AlphaEvolve-style operator-code synthesis using our existing A/B harness as the evaluator.","acceptance_criteria":"on the evolved-3M-nols-3 15-fail plateau seed: repair loop reduces hard-fail count where 1M+ blind evals did not, within \u003c=20 LLM calls; edit-DSL rejects malformed proposals; A/B at equal native-eval budget shows strictly better final fails on \u003e=2/3 seeds","status":"open","priority":2,"issue_type":"feature","owner":"bruno@postle.net","created_at":"2026-08-02T09:15:54Z","created_by":"Bruno Postle","updated_at":"2026-08-02T09:15:54Z","dependencies":[{"issue_id":"homemaker-py-2g7.7","depends_on_id":"homemaker-py-2g7","type":"parent-child","created_at":"2026-08-02T10:15:53Z","created_by":"Bruno Postle","metadata":"{}"}],"dependency_count":0,"dependent_count":1,"comment_count":0}
@ -127,27 +127,27 @@
{"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.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-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} {"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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":"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."}

138
DESIGN.md
View file

@ -4338,3 +4338,141 @@ test: `homemaker-evolve init.dom --programme-dir . --no-leaf-sharing
--shapecurve-warmstart` in `examples/harbor-house-l0` runs to completion and --shapecurve-warmstart` in `examples/harbor-house-l0` runs to completion and
the emitted `.dom` scores cleanly with `homemaker-fitness` (score matches the emitted `.dom` scores cleanly with `homemaker-fitness` (score matches
the run's own reported best). the run's own reported best).
### 37.5 `homemaker-py-wkh` DP-exact hard pre-filter — measured 2026-08-03, ACCEPTANCE: PARTIAL
**What was built.** `6xh`'s own deferred item 1: use `shapecurve`'s exact
feasible/infeasible verdict alongside `operators.predicted_shape_fails`'
heuristic-count pre-filter (§12.3/9gp.1) in `driver._evaluate`, instead of
only as an NM warm-start. Added `shapecurve.is_feasible(level_root, fit,
grid_n)` — a read-only refactor of `solve`'s own check phase (`_check`, now
shared by both) that never calls `realise()`/writes `division`, so the new
`shapecurve_prune` flag composes cleanly with `shapecurve_warmstart` and the
two can be A/B'd independently without one experiment's tree mutation
contaminating the other's measurement (a real risk: `solve` always writes a
realised point in place when feasible).
**Composition (the design item the bead's own description flagged as
needed).** Conservative by construction, chosen to extend today's prune
guard (`pred > threshold && pred >= best_n_fails`) rather than replace it,
because — per the bead's own risk framing — a wrong prune permanently
discards a topology that could have beaten the incumbent, unlike a bad
warm-start:
- **DP feasible → veto.** A real ratio point exists clearing every leaf's
size/width/proportion threshold, so a heuristic-triggered prune must have
come from `predicted_shape_fails`' own single (proportion-aware) layout
being an unlucky, non-representative sample — not the topology's true
floor. Never prunes in this case, and skips the `predicted_shape_fails`
eval entirely (redundant once the DP has already answered the question it
approximates).
- **DP infeasible + incumbent already at 0 total fails → exact prune.**
Infeasible proves the shape-fail floor is ≥1 (0/400 measured false
negatives across both validation sweeps below), which alone beats a
zero-fail incumbent — no heuristic count needed, and again the
`predicted_shape_fails` eval is skipped.
- **DP infeasible + incumbent >0 total fails → defer to the heuristic,
unchanged.** Infeasible only proves the floor is ≥1, not that it reaches
an arbitrary `best_n_fails>0`; asserting that would need a hard
*count*, which the DP (a boolean feasibility oracle) does not give.
`predicted_shape_fails` still runs and its threshold decides, exactly as
before `shapecurve_prune` existed.
Threaded as `search(…, shapecurve_prune=False)` (mirrors `shapecurve_warmstart`'s
threading exactly — `_evaluate`, the parallel-batch tuple, the explicit
single-seed call) and `homemaker-evolve --shapecurve-prune` (default off).
Note: like the pre-existing `feasibility_filter`/`feasibility_max_shape_fails`
it augments, `shapecurve_prune` is a no-op unless `feasibility_filter=True`
is also set — that pair has never been exposed as its own CLI flag (a
pre-existing gap in `evolve.py`, not introduced here), so
`--shapecurve-prune` alone only reaches the Python `driver.search` API today.
**False-negative-risk validation (bead item (b)): a second, genuinely
non-rectangular plot, not just a rotated copy of harbor-house-l0.**
`experiments/validate_shapecurve.py` was pointed at the *promoted product
module* (`homemaker_layout.shapecurve`, not the frozen `experiments/
shapecurve_spike.py` it validated in §37.2) — the actual code path
`wkh`'s hard-prune now trusts — and given a `programme_dir` CLI arg (was
silently hardcoded to harbor-house-l0 before) to run against
`examples/programme-house`: an authentically skewed parallelogram plot
(`node:` corners not axis-aligned, unlike harbor-house-l0's near-rectangle),
its own 6-space single-storey programme, 200 random topologies, seed 12345
(same protocol as §37.2):
| metric | harbor-house-l0 (re-run, product module) | programme-house (skewed) |
|---|---|---|
| agreement | 20/20 = 100.0% (n=20 smoke) | 200/200 = **100.0%** |
| false positives | 0 | **0** |
| false negatives | 0 | **0** |
| DP feasible / NM 0-shape-fail | — | 20/200 both |
| speedup | 97.6x | **87.4x** |
Zero false negatives on a structurally distinct, genuinely non-rectangular
plot — the DP-infeasible verdict the hard-prune branch relies on has now
been checked on 400 combined topologies (200 harbor-house-l0 from §37.2 +
200 here) across two plots with no measured false negative either time.
This clears the bead's own bar ("a larger/less-rectangular topology sweep
… before enabling by default" — still shipped **off** by default, matching
every other experimental flag in this codebase, but the safety case for a
future default-on is now measured, not just argued).
**`driver.search` A/B (bead item (c)): NULL on harbor-house-l0 at the
6xh-matching protocol.** `experiments/ab_shapecurve_prune.py`, same
benchmark/budget/seed protocol as §37.4's warm-start A/B
(`feasibility_filter=True, feasibility_max_shape_fails=0`, budget=2000,
seeds 0-4, `leaf_sharing=False`):
| seed | off hard/soft/fit/topo | on hard/soft/fit/topo |
|---|---|---|
| 0 | 3/13/1.222e-08/25 | 3/13/1.222e-08/25 |
| 1 | 4/13/1.234e-08/25 | 4/13/1.234e-08/25 |
| 2 | 6/15/6.95e-10/25 | 6/15/6.95e-10/25 |
| 3 | 3/18/3.954e-10/25 | 3/18/3.954e-10/25 |
| 4 | 6/17/5.6e-11/26 | 6/17/5.6e-11/26 |
Byte-identical off/on across all 5 seeds. Instrumented to find out why
(`shapecurve.is_feasible` call-count/verdict spy, seed 0 alone): 17 calls,
**0 feasible, 17 infeasible** — the veto branch never fired (needs at least
one DP-feasible verdict on a would-be-pruned candidate; got none) and the
incumbent's total fails never reached 0 in this run (best hard=3, soft=13,
so the exact-prune branch's own precondition, `best_n_fails<=0`, was never
true either) — every one of the 17 eligible checks fell through to "defer
to heuristic, unchanged" by construction, so nothing *could* have differed.
Root cause is upstream of `wkh`: at `feasibility_max_shape_fails=0`,
`predicted_shape_fails` itself rarely reaches `best_n_fails` (≈16-18 here)
on harbor-house-l0's modest leaf counts — `test_feasibility_filter_
prunes_cheaply` (tests/test_driver.py) already had to monkeypatch it to a
forced 999 to observe *any* real prune, a pre-existing characteristic of
9gp.1 (documented there as a "scaling lever", i.e. expected to bite on
larger programmes/leaf counts, not this benchmark) — not something `wkh`'s
composition introduced or could route around, since it only ever refines a
decision the base heuristic was already about to make.
**ACCEPTANCE: PARTIAL.** Composition designed and landed conservatively
(never prunes anything the pre-`wkh` filter wouldn't have, per the veto/
defer rules above); DP-exactness (0 false negatives) independently
re-validated on a second, structurally distinct plot at the same 200-
topology scale as §37.2's original result — items (a) and (b) from the
bead's own description are done. Item (c), the `driver.search` A/B, is
measured but **NULL** on harbor-house-l0 at this budget/threshold, for the
structurally-understood reason above (the base 9gp.1 filter barely engages
organically at this scale, so there is nothing for `wkh`'s refinement to
change) rather than a defect in the new logic. A benchmark/threshold where
`predicted_shape_fails` organically prunes — a larger programme or leaf
count, where 9gp.1 is itself expected to start mattering — is the natural
next measurement, tracked as a follow-up rather than blocking this landing;
the multi-storey (`homemaker-py-koo`) and leaf-sharing (`homemaker-py-tym`)
follow-ups remain the more direct route to that (today's DP eligibility
excludes `programme-house`/`harbor-house`'s real ≥2-storey, leaf-sharing-
default programmes, the same gap §37.4 already flagged).
**Verification.** `tests/test_shapecurve.py` (+1 test): `is_feasible` agrees
with `solve`'s own verdict on both the feasible and infeasible fixtures
already used there, and never writes `division` in either case.
`tests/test_driver.py` (+4 tests): off/on parity when the flag is off; the
veto branch (DP feasible skips `predicted_shape_fails` and never prunes,
even when the heuristic would have via a forced 999 return); the exact-prune
branch (DP infeasible + `best_n_fails<=0` prunes for 1 eval, skipping
`predicted_shape_fails`); the defer branch (DP infeasible + `best_n_fails>0`
still consults and obeys `predicted_shape_fails`, unchanged). Full suite:
393 passed.

77
experiments/ab_shapecurve_prune.py Normal file
View file

@ -0,0 +1,77 @@
"""A/B: does the shape-curve DP's exact feasible/infeasible verdict, composed
with the existing heuristic-count shape-feasibility pre-filter, beat the
heuristic-only filter on wall-clock/evals-to-fail-count, on the real
``driver.search`` loop (homemaker-py-wkh, DESIGN.md §37.5)?
Both arms run with ``feasibility_filter=True, feasibility_max_shape_fails=0``
(the existing §12.3 pre-filter switched on) so the only variable is whether
``shapecurve_prune`` additionally consults the DP (veto a heuristic prune when
DP-feasible; hard-prune immediately when DP-infeasible and the incumbent
already has zero total fails). Scoped to the DP's validated envelope
(DESIGN.md §37.2): single storey, no leaf_sharing/superpose/max_share/
multi_use -- ``examples/harbor-house-l0``, same benchmark and protocol as
``ab_shapecurve_warmstart.py`` for direct comparability.
Metric: mean (n_hard, n_soft, fitness) of ``driver.search``'s best individual
at a FIXED budget across several seeds, plus mean topologies explored (the
pre-filter's whole point is spending fewer evals per pruned topology, so more
topologies get tried at the same budget).
Usage: python experiments/ab_shapecurve_prune.py [budget] [n_seeds]
"""
from __future__ import annotations
import sys
import time
from pathlib import Path
from homemaker_layout import dom, driver
PROGRAMME_DIR = Path(__file__).parent.parent / "examples" / "harbor-house-l0"
def run_arm(seed_root: dom.Node, budget: int, seed: int, prune: bool):
t0 = time.perf_counter()
r = driver.search(
seed_root, PROGRAMME_DIR, budget=budget, pop_size=8, child_budget=80,
seed_budget=200, seed=seed, leaf_sharing=False,
feasibility_filter=True, feasibility_max_shape_fails=0,
shapecurve_prune=prune,
)
elapsed = time.perf_counter() - t0
return r, elapsed
def main() -> None:
budget = int(sys.argv[1]) if len(sys.argv) > 1 else 2000
n_seeds = int(sys.argv[2]) if len(sys.argv) > 2 else 5
seed_root = dom.load(str(PROGRAMME_DIR / "init.dom"))
rows = []
for seed in range(n_seeds):
off, t_off = run_arm(seed_root, budget, seed, prune=False)
on, t_on = run_arm(seed_root, budget, seed, prune=True)
rows.append((seed, off.best.n_hard, off.best.n_soft, off.best.fitness,
off.n_topologies, t_off,
on.best.n_hard, on.best.n_soft, on.best.fitness,
on.n_topologies, t_on))
print(f"seed {seed}: off hard={off.best.n_hard} soft={off.best.n_soft} "
f"fit={off.best.fitness:.4g} topo={off.n_topologies} {t_off:.1f}s | "
f"on hard={on.best.n_hard} soft={on.best.n_soft} "
f"fit={on.best.fitness:.4g} topo={on.n_topologies} {t_on:.1f}s",
flush=True)
n = len(rows)
mean = lambda idx: sum(r[idx] for r in rows) / n
print()
print(f"budget={budget} n_seeds={n_seeds} programme={PROGRAMME_DIR}")
print(f"OFF: mean hard={mean(1):.3f} soft={mean(2):.3f} fitness={mean(3):.6g} "
f"topo={mean(4):.1f} wall={mean(5):.1f}s")
print(f"ON : mean hard={mean(6):.3f} soft={mean(7):.3f} fitness={mean(8):.6g} "
f"topo={mean(9):.1f} wall={mean(10):.1f}s")
if __name__ == "__main__":
main()

19
experiments/validate_shapecurve.py
View file

@ -34,9 +34,7 @@ import numpy as np
import yaml import yaml
from homemaker_layout import dom, driver, fitness as fit_mod, geometry, innerloop from homemaker_layout import dom, driver, fitness as fit_mod, geometry, innerloop
from homemaker_layout import shapecurve as sc
sys.path.insert(0, "experiments")
import shapecurve_spike as sc # noqa: E402
PROGRAMME_DIR = "examples/harbor-house-l0" PROGRAMME_DIR = "examples/harbor-house-l0"
_SHAPE_SUFFIXES = (" size", " width", " proportion") _SHAPE_SUFFIXES = (" size", " width", " proportion")
@ -114,7 +112,7 @@ def main(n_topologies: int = 200, nm_budget: int = 100, grid_n: int = 150,
seed = int(rng.integers(0, 2**31 - 1)) seed = int(rng.integers(0, 2**31 - 1))
trng = np.random.default_rng(seed) trng = np.random.default_rng(seed)
topo = driver.random_topology(seed_root, n_leaves, trng, types) topo = driver.random_topology(seed_root, n_leaves, trng, types)
dom._link(topo) dom.link(topo)
lvl = dom.levels(topo)[0] lvl = dom.levels(topo)[0]
if len(lvl.leaves()) < 2: if len(lvl.leaves()) < 2:
continue # undivided, nothing for the DP to do continue # undivided, nothing for the DP to do
@ -180,15 +178,20 @@ def main(n_topologies: int = 200, nm_budget: int = 100, grid_n: int = 150,
if __name__ == "__main__": if __name__ == "__main__":
# Usage: validate_shapecurve.py [n_topologies] [nm_budget] [grid_n] [rotate_deg] # Usage: validate_shapecurve.py [n_topologies] [nm_budget] [grid_n] [rotate_deg] [programme_dir]
# rotate_deg (optional, default 0): test on a scratch copy of the plot # rotate_deg (optional, default 0): test on a scratch copy of the plot
# rotated this many degrees about its centroid -- DESIGN.md §37.2's # rotated this many degrees about its centroid -- DESIGN.md §37.2's
# rotation-invariance check (0 => harbor-house-l0 unmodified). # rotation-invariance check (0 => plot unmodified).
# programme_dir (optional, default examples/harbor-house-l0): homemaker-py-wkh
# (DESIGN.md §37.5) uses this to sweep a genuinely non-rectangular plot
# (e.g. examples/programme-house's skewed parallelogram) rather than only a
# rotated copy of harbor-house-l0's near-rectangular one.
n = int(sys.argv[1]) if len(sys.argv) > 1 else 200 n = int(sys.argv[1]) if len(sys.argv) > 1 else 200
budget = int(sys.argv[2]) if len(sys.argv) > 2 else 100 budget = int(sys.argv[2]) if len(sys.argv) > 2 else 100
grid_n = int(sys.argv[3]) if len(sys.argv) > 3 else 150 grid_n = int(sys.argv[3]) if len(sys.argv) > 3 else 150
rotate_deg = float(sys.argv[4]) if len(sys.argv) > 4 else 0.0 rotate_deg = float(sys.argv[4]) if len(sys.argv) > 4 else 0.0
prog_dir = str(rotated_plot_dir(PROGRAMME_DIR, rotate_deg)) if rotate_deg else PROGRAMME_DIR base_dir = sys.argv[5] if len(sys.argv) > 5 else PROGRAMME_DIR
prog_dir = str(rotated_plot_dir(base_dir, rotate_deg)) if rotate_deg else base_dir
if rotate_deg: if rotate_deg:
print(f"(testing on {PROGRAMME_DIR}'s plot rotated {rotate_deg} deg -> {prog_dir})") print(f"(testing on {base_dir}'s plot rotated {rotate_deg} deg -> {prog_dir})")
main(n, budget, grid_n, prog_dir) main(n, budget, grid_n, prog_dir)

44
src/homemaker_layout/driver.py
View file

@ -174,7 +174,8 @@ def _evaluate(root: dom.Node, programme_dir, urb_root, x0, budget, inner_kw,
conn_grade: bool = False, conn_grade: bool = False,
collapse_insearch: bool = True, collapse_insearch: bool = True,
multi_use: bool = False, multi_use: bool = False,
shapecurve_warmstart: bool = False) -> tuple[Individual, int]: shapecurve_warmstart: bool = False,
shapecurve_prune: bool = False) -> tuple[Individual, int]:
# §12.3 shape-feasibility pre-filter (homemaker-py-9gp.1): if even the best # §12.3 shape-feasibility pre-filter (homemaker-py-9gp.1): if even the best
# achievable (proportion-aware) geometry of this topology already has at least # achievable (proportion-aware) geometry of this topology already has at least
# as many shape fails as the incumbent's TOTAL fails — and exceeds the tunable # as many shape fails as the incumbent's TOTAL fails — and exceeds the tunable
@ -191,17 +192,44 @@ def _evaluate(root: dom.Node, programme_dir, urb_root, x0, budget, inner_kw,
# point and write it onto the tree in place. `x0=None` below then picks it up # point and write it onto the tree in place. `x0=None` below then picks it up
# as the inner loop's start point. On infeasible or ineligible, `root` is left # as the inner loop's start point. On infeasible or ineligible, `root` is left
# untouched — falls through to today's cold/proportion-aware start exactly. # untouched — falls through to today's cold/proportion-aware start exactly.
if shapecurve_warmstart and x0 is None and shapecurve.eligible( dp_eligible = ((shapecurve_warmstart or shapecurve_prune)
root, leaf_sharing, superpose, max_share, multi_use): and shapecurve.eligible(root, leaf_sharing, superpose, max_share, multi_use))
shapecurve.solve(root, _fitness_for( dp_feasible = None
if dp_eligible and shapecurve_warmstart and x0 is None:
dp_feasible, _ = shapecurve.solve(root, _fitness_for(
str(programme_dir), leaf_sharing, superpose, max_share, str(programme_dir), leaf_sharing, superpose, max_share,
conn_grade, collapse_insearch, multi_use)) conn_grade, collapse_insearch, multi_use))
if (feasibility_max_shape_fails is not None and best_n_fails is not None): if (feasibility_max_shape_fails is not None and best_n_fails is not None):
# §37.5 DP-exact hard prune (homemaker-py-wkh, DESIGN.md §37.5): the
# shape-curve DP gives an EXACT feasible/infeasible verdict (0/200
# measured false negatives on harbor-house-l0, §37.2) for the same
# size/width/proportion family predicted_shape_fails only heuristically
# counts at one (proportion-aware) layout. Composed conservatively —
# DP feasible VETOES the heuristic prune outright (a real feasible
# point exists, so the heuristic's high count was a false signal from
# an unlucky single layout, never the true floor); DP infeasible only
# licenses an exact prune when the incumbent already has zero total
# fails (best_n_fails<=0) — infeasible proves the shape-fail floor is
# >=1, which alone beats a zero-fail incumbent, but does not by itself
# establish the floor reaches an arbitrary best_n_fails>0, so that case
# still defers to the heuristic count (unchanged behaviour).
if dp_eligible and shapecurve_prune and dp_feasible is None:
dp_feasible = shapecurve.is_feasible(root, _fitness_for(
str(programme_dir), leaf_sharing, superpose, max_share,
conn_grade, collapse_insearch, multi_use))
if shapecurve_prune and dp_feasible is True:
prune = False
pred = 0
elif shapecurve_prune and dp_feasible is False and best_n_fails <= 0:
prune = True
pred = max(1, best_n_fails)
else:
pred = operators.predicted_shape_fails( pred = operators.predicted_shape_fails(
root, _reqs_for(str(programme_dir)), root, _reqs_for(str(programme_dir)),
_fitness_for(str(programme_dir), leaf_sharing, superpose, max_share, _fitness_for(str(programme_dir), leaf_sharing, superpose, max_share,
conn_grade, collapse_insearch, multi_use)) conn_grade, collapse_insearch, multi_use))
if pred > feasibility_max_shape_fails and pred >= best_n_fails: prune = pred > feasibility_max_shape_fails and pred >= best_n_fails
if prune:
# predicted_shape_fails only counts the size/width/proportion/ # predicted_shape_fails only counts the size/width/proportion/
# crinkliness SOFT family (operators._SHAPE_FAIL_SUFFIXES), so the # crinkliness SOFT family (operators._SHAPE_FAIL_SUFFIXES), so the
# proxy carries no HARD information — tier it all soft. # proxy carries no HARD information — tier it all soft.
@ -284,6 +312,7 @@ def search(
seed_pop: list[dom.Node] | None = None, seed_pop: list[dom.Node] | None = None,
collapse_insearch: bool = True, collapse_insearch: bool = True,
shapecurve_warmstart: bool = False, shapecurve_warmstart: bool = False,
shapecurve_prune: bool = False,
) -> SearchResult: ) -> SearchResult:
"""Run the memetic loop from ``seed_root`` until ``budget`` oracle """Run the memetic loop from ``seed_root`` until ``budget`` oracle
evaluations are consumed. Returns the best individual found; its ``root`` evaluations are consumed. Returns the best individual found; its ``root``
@ -533,7 +562,7 @@ def search(
full = [ full = [
(root, programme_dir, urb_root, x0, budget_, kw_, lin, use_grade, (root, programme_dir, urb_root, x0, budget_, kw_, lin, use_grade,
mx, best_nf, leaf_sharing, superpose, max_share, conn_grade, mx, best_nf, leaf_sharing, superpose, max_share, conn_grade,
collapse_insearch, multi_use, shapecurve_warmstart) collapse_insearch, multi_use, shapecurve_warmstart, shapecurve_prune)
for root, x0, budget_, kw_, lin in tasks for root, x0, budget_, kw_, lin in tasks
] ]
if _pool is not None: if _pool is not None:
@ -620,7 +649,8 @@ def search(
conn_grade=conn_grade, conn_grade=conn_grade,
collapse_insearch=collapse_insearch, collapse_insearch=collapse_insearch,
multi_use=multi_use, multi_use=multi_use,
shapecurve_warmstart=shapecurve_warmstart) shapecurve_warmstart=shapecurve_warmstart,
shapecurve_prune=shapecurve_prune)
n_evals += used n_evals += used
admit(seed_ind, pop) admit(seed_ind, pop)

18
src/homemaker_layout/evolve.py
View file

@ -114,6 +114,22 @@ def _parse_args(argv=None) -> argparse.Namespace:
"superpose/max-share/multi-use — none of which the DP " "superpose/max-share/multi-use — none of which the DP "
"models). Falls through to today's start unchanged when " "models). Falls through to today's start unchanged when "
"ineligible or DP-infeasible (default: off)") "ineligible or DP-infeasible (default: off)")
p.add_argument("--shapecurve-prune", dest="shapecurve_prune",
action=argparse.BooleanOptionalAction,
default=_env_bool("HOMEMAKER_SHAPECURVE_PRUNE", False),
help="homemaker-py-wkh (DESIGN.md §37.5): use the shape-curve "
"DP's exact feasible/infeasible verdict alongside the "
"existing predicted_shape_fails pre-filter (a no-op "
"unless the driver.search()-level feasibility_filter is "
"also on -- not yet exposed as its own CLI flag). A "
"DP-feasible verdict vetoes a heuristic-triggered prune "
"outright (a real shape-feasible point exists, so the "
"heuristic's high count was a false signal); a "
"DP-infeasible verdict prunes immediately only when the "
"incumbent already has zero total fails (exact: "
"DP-infeasible proves the shape-fail floor is >=1); "
"otherwise falls through to today's heuristic-count "
"decision unchanged (default: off)")
p.add_argument("--conn-grade", dest="conn_grade", p.add_argument("--conn-grade", dest="conn_grade",
action=argparse.BooleanOptionalAction, action=argparse.BooleanOptionalAction,
default=_env_bool("HOMEMAKER_CONN_GRADE", False), default=_env_bool("HOMEMAKER_CONN_GRADE", False),
@ -247,6 +263,7 @@ def main(argv=None) -> int:
print(f"ruin recreate : {args.ruin_recreate}", file=sys.stderr) print(f"ruin recreate : {args.ruin_recreate}", file=sys.stderr)
print(f"collapse in-search : {args.collapse_insearch}", file=sys.stderr) print(f"collapse in-search : {args.collapse_insearch}", file=sys.stderr)
print(f"shapecurve warmstart : {args.shapecurve_warmstart}", file=sys.stderr) print(f"shapecurve warmstart : {args.shapecurve_warmstart}", file=sys.stderr)
print(f"shapecurve prune : {args.shapecurve_prune}", file=sys.stderr)
print(f"output : {out or 'stdout'}", file=sys.stderr, flush=True) print(f"output : {out or 'stdout'}", file=sys.stderr, flush=True)
anneal_ladder = None anneal_ladder = None
@ -303,6 +320,7 @@ def main(argv=None) -> int:
enable_ruin_recreate=args.ruin_recreate, enable_ruin_recreate=args.ruin_recreate,
collapse_insearch=args.collapse_insearch, collapse_insearch=args.collapse_insearch,
shapecurve_warmstart=args.shapecurve_warmstart, shapecurve_warmstart=args.shapecurve_warmstart,
shapecurve_prune=args.shapecurve_prune,
log=lambda m: print(m, file=sys.stderr, flush=True), log=lambda m: print(m, file=sys.stderr, flush=True),
) )
_finish_sharing = args.leaf_sharing _finish_sharing = args.leaf_sharing

31
src/homemaker_layout/shapecurve.py
View file

@ -353,6 +353,29 @@ def build_curves_with_children(
return Curve(w_of_h=w_of_h, h_of_w=h_of_w) return Curve(w_of_h=w_of_h, h_of_w=h_of_w)
def _check(level_root: dom_mod.Node, fit, grid_n: int) -> tuple[
Feasibility, dict[int, tuple[Curve, Curve]], np.ndarray, float, float]:
w_plot, h_plot = _dims(level_root)
grid = make_grid(max(w_plot, h_plot) * 1.2, n=grid_n)
curves_by_node: dict[int, tuple[Curve, Curve]] = {}
root_curve = build_curves_with_children(level_root, fit, grid, curves_by_node)
feas = check_feasible(root_curve, grid, w_plot, h_plot)
return feas, curves_by_node, grid, w_plot, h_plot
def is_feasible(level_root: dom_mod.Node, fit, grid_n: int = 150) -> bool:
"""Read-only DP feasibility verdict: does some equal-offset ratio
assignment clear the size/width/proportion FAIL_THRESHOLD for every leaf?
Unlike :func:`solve`, never writes ``division`` — the hard-prune caller
(``driver._evaluate``, homemaker-py-wkh) needs the boolean verdict alone,
without the warm-start's tree mutation (kept a strictly separate code path
so the two experimental flags, ``shapecurve_prune``/``shapecurve_warmstart``,
compose cleanly and can be A/B'd independently)."""
feas, *_ = _check(level_root, fit, grid_n)
return feas.feasible
def solve(level_root: dom_mod.Node, fit, grid_n: int = 150) -> tuple[bool, dict]: def solve(level_root: dom_mod.Node, fit, grid_n: int = 150) -> tuple[bool, dict]:
"""End-to-end: compute plot dims, build curves, check root feasibility, """End-to-end: compute plot dims, build curves, check root feasibility,
and (if feasible) write realising ratios in place. Returns (feasible, and (if feasible) write realising ratios in place. Returns (feasible,
@ -362,13 +385,7 @@ def solve(level_root: dom_mod.Node, fit, grid_n: int = 150) -> tuple[bool, dict]
notion of upper-storey ``below``-inherited fixed splits and will notion of upper-storey ``below``-inherited fixed splits and will
overwrite ``division`` unconditionally on every divided node it walks. overwrite ``division`` unconditionally on every divided node it walks.
""" """
w_plot, h_plot = _dims(level_root) feas, curves_by_node, grid, w_plot, h_plot = _check(level_root, fit, grid_n)
grid = make_grid(max(w_plot, h_plot) * 1.2, n=grid_n)
curves_by_node: dict[int, tuple[Curve, Curve]] = {}
root_curve = build_curves_with_children(level_root, fit, grid, curves_by_node)
feas = check_feasible(root_curve, grid, w_plot, h_plot)
if feas.feasible: if feas.feasible:
realise(level_root, curves_by_node, grid, w_plot, h_plot) realise(level_root, curves_by_node, grid, w_plot, h_plot)
geometry.clear_cache() geometry.clear_cache()

103
tests/test_driver.py
View file

@ -360,6 +360,109 @@ def test_shapecurve_warmstart_skips_multistorey(monkeypatch):
assert not solve_calls assert not solve_calls
HARBOR_L0 = Path(__file__).parent.parent / "examples" / "harbor-house-l0"
def _fake_optimise_ok(root, programme_dir, x0=None, budget=200, urb_root=None, **kw):
for _, b in innerloop.free_with_keys(root):
b.division = [0.25, 0.25]
return innerloop.Result(
x=np.array([0.25]), fitness=0.5, n_fails=0, fail_lines=(),
x0_fitness=0.25, x0_n_fails=1, n_evals=budget, n_oracle_calls=1,
)
def test_shapecurve_prune_off_matches_baseline(fake_inner):
"""homemaker-py-wkh: with the flag off (default), the run is identical to
one that omits the param — the same clean A/B control as the existing
feasibility-filter/shapecurve-warmstart control tests."""
init_root = dom.load(str(INIT_FILE))
base = driver.search(init_root, CORPUS, budget=600, pop_size=4,
child_budget=60, seed_budget=100, seed=9)
off = driver.search(init_root, CORPUS, budget=600, pop_size=4,
child_budget=60, seed_budget=100, seed=9,
shapecurve_prune=False)
assert off.best.sig == base.best.sig
assert off.n_topologies == base.n_topologies
assert off.n_evals == base.n_evals
def test_shapecurve_prune_vetoes_heuristic_when_dp_feasible(monkeypatch):
"""homemaker-py-wkh (DESIGN.md §37.5): a DP-feasible verdict is a real
certificate that some ratio point clears every leaf's shape threshold, so
it must veto a heuristic-triggered prune outright — even one predicted
from a bad (999-fail) proxy layout — and skip the ``predicted_shape_fails``
eval entirely rather than just override its verdict."""
from homemaker_layout import operators, shapecurve
monkeypatch.setattr(shapecurve, "is_feasible", lambda root, fit, grid_n=150: True)
pred_calls = []
monkeypatch.setattr(operators, "predicted_shape_fails",
lambda root, reqs, fit: pred_calls.append(1) or 999)
monkeypatch.setattr(innerloop, "optimise", _fake_optimise_ok)
if not HARBOR_L0.is_dir():
pytest.skip("harbor-house-l0 not available")
root = dom.load(str(HARBOR_L0 / "init.dom"))
ind, used = driver._evaluate(
root, HARBOR_L0, None, x0=None, budget=100, inner_kw={}, lineage="child",
feasibility_max_shape_fails=0, best_n_fails=5, leaf_sharing=False,
shapecurve_prune=True)
assert not pred_calls, "heuristic proxy must be skipped when DP proves feasibility"
assert not ind.lineage.startswith("pruned/")
assert used == 100
def test_shapecurve_prune_hard_prunes_when_dp_infeasible_and_incumbent_perfect(monkeypatch):
"""homemaker-py-wkh: DP-infeasible proves the shape-fail floor is >=1
(exact, 0/200 measured false negatives — DESIGN.md §37.2), which alone
beats a zero-total-fail incumbent — an exact prune, no heuristic count
needed."""
from homemaker_layout import operators, shapecurve
monkeypatch.setattr(shapecurve, "is_feasible", lambda root, fit, grid_n=150: False)
pred_calls = []
monkeypatch.setattr(operators, "predicted_shape_fails",
lambda root, reqs, fit: pred_calls.append(1) or 0)
if not HARBOR_L0.is_dir():
pytest.skip("harbor-house-l0 not available")
root = dom.load(str(HARBOR_L0 / "init.dom"))
ind, used = driver._evaluate(
root, HARBOR_L0, None, x0=None, budget=100, inner_kw={}, lineage="child",
feasibility_max_shape_fails=0, best_n_fails=0, leaf_sharing=False,
shapecurve_prune=True)
assert not pred_calls, "the exact DP verdict makes the heuristic proxy redundant here"
assert ind.lineage.startswith("pruned/")
assert used == 1
def test_shapecurve_prune_defers_to_heuristic_when_incumbent_nonzero(monkeypatch):
"""homemaker-py-wkh: DP-infeasible only proves the shape-fail floor is
>=1, not that it reaches an arbitrary best_n_fails>0, so that case must
still fall through to today's heuristic-count decision unchanged."""
from homemaker_layout import operators, shapecurve
monkeypatch.setattr(shapecurve, "is_feasible", lambda root, fit, grid_n=150: False)
pred_calls = []
monkeypatch.setattr(operators, "predicted_shape_fails",
lambda root, reqs, fit: pred_calls.append(1) or 999)
if not HARBOR_L0.is_dir():
pytest.skip("harbor-house-l0 not available")
root = dom.load(str(HARBOR_L0 / "init.dom"))
ind, used = driver._evaluate(
root, HARBOR_L0, None, x0=None, budget=100, inner_kw={}, lineage="child",
feasibility_max_shape_fails=0, best_n_fails=5, leaf_sharing=False,
shapecurve_prune=True)
assert pred_calls, "heuristic proxy must still be consulted when best_n_fails>0"
assert ind.lineage.startswith("pruned/")
assert used == 1
def test_search_parallel_smoke(): def test_search_parallel_smoke():
"""n_workers>1 runs without error and produces valid results.""" """n_workers>1 runs without error and produces valid results."""
init_root = dom.load(str(INIT_FILE)) init_root = dom.load(str(INIT_FILE))

22
tests/test_shapecurve.py
View file

@ -88,3 +88,25 @@ def test_solve_is_deterministic():
divisions_2 = [tuple(b.division) for b in solver.free_branches(root)] divisions_2 = [tuple(b.division) for b in solver.free_branches(root)]
assert f1 == f2 is True assert f1 == f2 is True
assert divisions_1 == pytest.approx(divisions_2) assert divisions_1 == pytest.approx(divisions_2)
def test_is_feasible_agrees_with_solve_but_never_writes(monkeypatch):
"""homemaker-py-wkh: the hard-prune caller needs the boolean verdict
without solve()'s tree mutation, so ``is_feasible`` must (a) agree with
``solve``'s own verdict and (b) never write ``division`` -- verified on
both the feasible and infeasible fixtures already exercised above."""
fit = _fit()
feasible_root = _small_feasible_topology()
before = [tuple(b.division) for b in solver.free_branches(feasible_root)]
assert shapecurve.is_feasible(feasible_root, fit) is True
after = [tuple(b.division) for b in solver.free_branches(feasible_root)]
assert before == after
seed = dom.load(str(HARBOR_L0 / "init.dom"))
rng = np.random.default_rng(0)
infeasible_root = driver.random_topology(seed, 60, rng, ["k1", "l1", "b1", "C", "O"])
before = [tuple(b.division) for b in solver.free_branches(infeasible_root)]
assert shapecurve.is_feasible(infeasible_root, fit) is False
after = [tuple(b.division) for b in solver.free_branches(infeasible_root)]
assert before == after