homemaker-layout/experiments/diag_leaf_sharing.py

167 lines
6.2 KiB
Python
Raw Normal View History

#!/usr/bin/env python3
"""Leaf-sharing floor probe (homemaker-py-erc.3, DESIGN.md §13.3).
Cheap de-risk BEFORE the full 20k A/B: does collapsing same-code rooms into
fewer, larger SHARED leaves actually lower the achievable fail floor, or does
the gain leak back as missing/size fails under the relaxed objective?
§13.1 found the per-leaf shape tax is ~1.8 and FLAT vs slicing density, so total
shape fails track leaf count linearly fewer leaves is the only floor-mover.
Leaf-sharing reduces ROOM-leaf count: a leaf sized to k×target counts as k
same-code rooms (graph._leaf_share_mult), so presence holds without a missing
fail and size is scored against k×target. This script builds the §12.2
constructive seed both ways (baseline OFF vs sharing ON, share_factor sweep),
scores each at its own seed geometry, and reports the fail breakdown.
DECISION RULE: if sharing-ON total fails drop well below baseline (and the drop
is in size+crinkliness, NOT bought back by missing) the floor moves proceed
to thread the flag through the driver for the staged 20k A/B. If missing fails
balloon or totals don't move → stop; same-code sharing cannot pay here.
Usage:
URB_NO_OCCLUSION=1 python3 experiments/diag_leaf_sharing.py
"""
from __future__ import annotations
import copy
import sys
from pathlib import Path
import numpy as np
sys.path.insert(0, str(Path(__file__).resolve().parents[1] / "src"))
from homemaker_layout import dom, fitness, innerloop, operators, programme # noqa: E402
PROGRAMMES = ["harbor-house", "maple-court"]
SEEDS = (0, 1, 2)
BUDGET = 80 # bootstrap child budget, as in Diagnostic B
ROOT = Path(__file__).resolve().parents[1]
class _force_sharing:
"""Context manager: make innerloop's NativeEvaluator build its fitness in
leaf_sharing mode (the dir's patterns.config has no such key), so the inner
loop optimises against the SAME relaxed objective the seed was scored under."""
def __init__(self, on: bool):
self.on = on
def __enter__(self):
self._orig = fitness.load_config
if self.on:
def patched(directory, overrides=None):
conf, cost = self._orig(directory, overrides=overrides)
conf = dict(conf)
conf["leaf_sharing"] = True
return conf, cost
fitness.load_config = patched
return self
def __exit__(self, *exc):
fitness.load_config = self._orig
# fail-string buckets (order matters: first match wins)
CATS = ("missing", "size", "width", "proportion", "crinkliness",
"adjacency", "access", "other")
def _bucket(fails) -> dict[str, int]:
out = {k: 0 for k in CATS}
for f in fails:
if "missing" in f or "too many" in f:
out["missing"] += 1
elif f.endswith(" size"):
out["size"] += 1
elif f.endswith(" width"):
out["width"] += 1
elif f.endswith(" proportion"):
out["proportion"] += 1
elif f.endswith(" crinkliness"):
out["crinkliness"] += 1
elif "adjacen" in f:
out["adjacency"] += 1
elif "access" in f or "inaccessible" in f:
out["access"] += 1
else:
out["other"] += 1
return out
def _build(seed_root, reqs, types, s, sharing, factor):
rng = np.random.default_rng(s)
return operators.constructive_topology(
seed_root, reqs, rng, types,
adjacency_aware=True, proportion_aware=True,
leaf_sharing=sharing, leaf_share_factor=factor)
def _measure(fit, pdir, seed_root, reqs, types, s, sharing, factor):
topo = _build(seed_root, reqs, types, s, sharing, factor)
n_leaves = sum(len(lvl.leaves()) for lvl in dom.levels(topo))
_score, fails = fit.score_with_fails(copy.deepcopy(topo))
before = {"n_leaves": n_leaves, "total": len(fails), **_bucket(fails)}
after_tree = copy.deepcopy(topo)
with _force_sharing(sharing):
innerloop.optimise(after_tree, str(pdir), x0=None, budget=BUDGET,
Remove the Perl oracle Owner's decision: "we need to abandon the perl oracle, this was only useful when initially porting, but I suspect many of the remaining problems have been carried in from the perl (such as the weird scoring of outdoor and circulation space, which definitely needs fixing)". 39 supports that second clause. Every defect the section found is inherited, not introduced: the two-sided crinkliness gaussian that double-charges surplus daylight (39.14), quality as a product over a variable number of factors (39.18), value_supported priced as value_inside so a terrace was worth more per m2 than a room (39.19), and circulation returning 0.07 per unit cost (hxi). So parity with the oracle was never a safety net -- it was a commitment to reproduce those defects. Each of 39.14, 39.18 and 39.19 would have been a parity failure had parity ever been checked, and keeping the tests would have meant reverting the fixes or explaining the failures away. Removed: oracle.py, test_oracle.py, the two parity tests and their fixture machinery in test_dom_corpus.py, innerloop.OracleEvaluator with its use_native and urb_root plumbing, the same plumbing through driver, and fourteen experiments/ scripts that could only run against Perl. Several of those are cited in earlier DESIGN sections; the citations now point into git history, which is the honest state -- they had been unrunnable since the oracle root (/home/bruno/src/urb) stopped being present. run_search is superseded by run_search_scaled, which does the same job natively. Kept: dump_areas.pl/.py, which validate GEOMETRY against Urb (4.1) rather than fitness, and the prose in fitness_cmd.py and dom.py explaining why the .score/.fails formats are shaped as they are. Provenance is worth keeping; a dead code path is not. CLAUDE.md updated: fitness.py is the only evaluator, and "Urb did it this way" is no longer an argument that a constant is right. 39.16 is the standing counterweight in the other direction -- the crinkliness target WAS right and twice looked wrong only because the code reading it was misunderstood. Inheritance is neither evidence for nor against. 410 passed. The 69 removed cases account exactly: 64 parity (all skipped, since no oracle .score was ever committed), 4 in test_oracle.py, and the guard test 39.20 added as a stopgap. Closes homemaker-py-118. Files homemaker-py-bk9 for the re-baseline that 39.19 made necessary. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01MJ84Feep79Hhm3E4zZJmnB
2026-09-06 07:56:24 +00:00
method="nm")
_s2, fails2 = fit.score_with_fails(copy.deepcopy(after_tree))
after = {"n_leaves": n_leaves, "total": len(fails2), **_bucket(fails2)}
return before, after
def _avg(rows, k):
return sum(r[k] for r in rows) / len(rows)
def main() -> int:
print("Leaf-sharing floor probe (§13.3)\n")
print("Seed geometry = constructive proportion-aware target (built per mode).")
print(f"Seeds: {SEEDS}. 'OFF' = baseline fitness; 'shareN' = leaf_sharing, "
"share_factor=N.")
print(f"seed = constructive seed; +il = after innerloop.optimise (nm, "
f"budget={BUDGET}) under the same objective.\n")
cols = ("leaves", "total", "missing", "size", "crink", "width", "prop",
"adj", "access", "other")
hdr = f"{'programme':<14}{'mode':>10}" + "".join(f"{c:>8}" for c in cols)
def _row(name, label, rows, k):
vals = [_avg(rows, "n_leaves"), _avg(rows, "total"),
_avg(rows, "missing"), _avg(rows, "size"),
_avg(rows, "crinkliness"), _avg(rows, "width"),
_avg(rows, "proportion"), _avg(rows, "adjacency"),
_avg(rows, "access"), _avg(rows, "other")]
print(f"{name:<14}{label:>10}" + "".join(f"{v:>8.1f}" for v in vals))
for name in PROGRAMMES:
pdir = ROOT / "examples" / name
reqs = programme.load_programme_dir(pdir)
types = sorted(reqs) + ["C", "O"]
conf, cost = fitness.load_config(pdir)
seed_root = dom.load(str(pdir / "init.dom"))
fit_off = fitness.Fitness(conf, cost)
conf_on = dict(conf)
conf_on["leaf_sharing"] = True
fit_on = fitness.Fitness(conf_on, cost)
print(hdr)
print("-" * len(hdr))
modes = [("OFF", fit_off, False, 1),
("share2", fit_on, True, 2),
("share3", fit_on, True, 3)]
for label, fit, sharing, factor in modes:
pairs = [_measure(fit, pdir, seed_root, reqs, types, s, sharing, factor)
for s in SEEDS]
befores = [b for b, _a in pairs]
afters = [a for _b, a in pairs]
_row(name, label, befores, "before")
_row(name, label + "+il", afters, "after")
print()
return 0
if __name__ == "__main__":
sys.exit(main())