text listlengths 2 2 |
|---|
[
[
"abs",
"float(\"inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"cos",
"float(\"inf\")"
],
[
"float(\"nan\")"
]
] |
[
[
"pow1_3",
"float(\"-inf\")"
],
[
"float(\"-inf\")"
]
] |
[
[
"mult2",
"float(\"-inf\")"
],
[
"float(\"-inf\")"
]
] |
[
[
"pow1_3",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"asinh",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"asinh",
"float(\"-inf\")"
],
[
"float(\"-inf\")"
]
] |
[
[
"inv",
"float(\"-inf\")"
],
[
"0"
]
] |
[
[
"tanh",
"float(\"inf\")"
],
[
"1"
]
] |
[
[
"div3",
"float(\"inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"pow1_5",
"float(\"inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"mult2",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"pow3",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"inv",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"cosh",
"np.pi"
],
[
"<constant>"
]
] |
[
[
"tanh",
"0"
],
[
"0"
]
] |
[
[
"atan",
"np.e"
],
[
"<constant>"
]
] |
[
[
"div5",
"float(\"-inf\")"
],
[
"float(\"-inf\")"
]
] |
[
[
"pow1_2",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"pow2",
"(-1)"
],
[
"1"
]
] |
[
[
"pow3",
"1"
],
[
"1"
]
] |
[
[
"pow1_3",
"1"
],
[
"1"
]
] |
[
[
"mult5",
"(-1)"
],
[
"<constant>"
]
] |
[
[
"pow2",
"np.pi"
],
[
"<constant>"
]
] |
[
[
"div2",
"0"
],
[
"0"
]
] |
[
[
"div4",
"np.e"
],
[
"<constant>"
]
] |
[
[
"pow3",
"float(\"-inf\")"
],
[
"float(\"-inf\")"
]
] |
[
[
"mult4",
"(-1)"
],
[
"<constant>"
]
] |
[
[
"abs",
"0"
],
[
"0"
]
] |
[
[
"sin",
"np.e"
],
[
"<constant>"
]
] |
[
[
"inv",
"1"
],
[
"1"
]
] |
[
[
"mult2",
"1"
],
[
"<constant>"
]
] |
[
[
"pow1_4",
"np.e"
],
[
"<constant>"
]
] |
[
[
"cos",
"0"
],
[
"1"
]
] |
[
[
"pow1_5",
"0"
],
[
"0"
]
] |
[
[
"pow1_2",
"1"
],
[
"1"
]
] |
[
[
"div3",
"0"
],
[
"0"
]
] |
[
[
"div2",
"np.pi"
],
[
"<constant>"
]
] |
[
[
"log",
"(-1)"
],
[
"float(\"nan\")"
]
] |
[
[
"sinh",
"float(\"inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"pow2",
"1"
],
[
"1"
]
] |
[
[
"pow1_2",
"float(\"-inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"div5",
"1"
],
[
"<constant>"
]
] |
[
[
"div5",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"mult5",
"np.pi"
],
[
"<constant>"
]
] |
[
[
"mult4",
"np.pi"
],
[
"<constant>"
]
] |
[
[
"abs",
"np.pi"
],
[
"np.pi"
]
] |
[
[
"neg",
"float(\"inf\")"
],
[
"float(\"-inf\")"
]
] |
[
[
"cos",
"np.pi"
],
[
"(-1)"
]
] |
[
[
"sinh",
"np.e"
],
[
"<constant>"
]
] |
[
[
"atan",
"float(\"-inf\")"
],
[
"<constant>"
]
] |
[
[
"tan",
"(-1)"
],
[
"<constant>"
]
] |
[
[
"tan",
"np.pi"
],
[
"0"
]
] |
[
[
"atanh",
"(-1)"
],
[
"float(\"-inf\")"
]
] |
[
[
"pow1_4",
"float(\"-inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"mult5",
"1"
],
[
"<constant>"
]
] |
[
[
"atanh",
"np.pi"
],
[
"float(\"nan\")"
]
] |
[
[
"div4",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"asin",
"(-1)"
],
[
"<constant>"
]
] |
[
[
"exp",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"div4",
"float(\"-inf\")"
],
[
"float(\"-inf\")"
]
] |
[
[
"div3",
"np.pi"
],
[
"<constant>"
]
] |
[
[
"pow1_4",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"log",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"acosh",
"np.e"
],
[
"<constant>"
]
] |
[
[
"exp",
"float(\"-inf\")"
],
[
"0"
]
] |
[
[
"pow4",
"float(\"-inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"sin",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"neg",
"0"
],
[
"0"
]
] |
[
[
"sin",
"float(\"-inf\")"
],
[
"float(\"nan\")"
]
] |
[
[
"tan",
"1"
],
[
"<constant>"
]
] |
[
[
"pow4",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"asin",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"log",
"1"
],
[
"0"
]
] |
[
[
"exp",
"1"
],
[
"np.e"
]
] |
[
[
"atanh",
"1"
],
[
"float(\"inf\")"
]
] |
[
[
"cosh",
"float(\"inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"pow2",
"float(\"inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"asin",
"1"
],
[
"<constant>"
]
] |
[
[
"mult3",
"np.e"
],
[
"<constant>"
]
] |
[
[
"pow4",
"1"
],
[
"1"
]
] |
[
[
"acos",
"(-1)"
],
[
"np.pi"
]
] |
[
[
"acosh",
"float(\"-inf\")"
],
[
"float(\"nan\")"
]
] |
[
[
"pow5",
"float(\"-inf\")"
],
[
"float(\"-inf\")"
]
] |
[
[
"pow1_3",
"(-1)"
],
[
"(-1)"
]
] |
[
[
"pow2",
"0"
],
[
"0"
]
] |
[
[
"pow1_3",
"np.pi"
],
[
"<constant>"
]
] |
[
[
"asinh",
"(-1)"
],
[
"<constant>"
]
] |
[
[
"pow5",
"(-1)"
],
[
"(-1)"
]
] |
[
[
"pow5",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"cosh",
"np.e"
],
[
"<constant>"
]
] |
[
[
"div5",
"0"
],
[
"0"
]
] |
[
[
"mult2",
"(-1)"
],
[
"<constant>"
]
] |
[
[
"inv",
"np.pi"
],
[
"<constant>"
]
] |
[
[
"inv",
"(-1)"
],
[
"(-1)"
]
] |
[
[
"acos",
"1"
],
[
"0"
]
] |
[
[
"pow3",
"(-1)"
],
[
"(-1)"
]
] |
[
[
"acosh",
"float(\"nan\")"
],
[
"float(\"nan\")"
]
] |
[
[
"mult5",
"float(\"inf\")"
],
[
"float(\"inf\")"
]
] |
[
[
"mult4",
"float(\"inf\")"
],
[
"float(\"inf\")"
]
] |
simplipy assets
Rule sets and engine configurations for simplipy, a fast, contract-sound simplification engine for symbolic mathematical expressions.
Install an engine by name (downloads this repository's files on first use):
pip install simplipy
import simplipy as sp
engine = sp.SimpliPyEngine.from_config(sp.get_path("acj-4-3")) # or "acj-3-2", "acj-2-1"
engine.simplify(('/', '<constant>', '*', '/', '*', 'x3', '<constant>', 'x3', 'log', 'x3'))
# ('<mul>', '<constant>', '<div>', 'log', 'x3', '</mul>')
engine.simplify('x3 * sin(<constant> + 1) / (x3 * x3)')
# '<constant>/x3'
Artifact-to-package compatibility: simplipy β₯ 0.12 (the AC engine with the clean
23-operator vocabulary) pairs with the acj-* artifacts and the base config below.
The pre-0.12 artifacts (2-1, 3-2, 4-3, dev_*) use the retired hyper-operator
vocabulary (mult3, pow2, ...) and load only on simplipy β€ 0.11 β pin
pip install "simplipy<0.12" to use them.
simplipy β₯ 0.12 artifacts (AC engine, clean vocabulary)
| Artifact | Sources β€ | Targets β€ | Rules | Mined with |
|---|---|---|---|---|
acj-2-1 |
2 | 1 | 8 | simplipy 0.12.0 |
acj-3-2 |
3 | 2 | 32 | simplipy 0.12.0 |
acj-4-3 |
4 | 3 | 520 | simplipy 0.12.0 |
base |
β | β | β | bare 23-operator config, the mining starting point |
Soundness correction (2026-07-29): acj-4-3 originally shipped 525 rules; the
independent rule-complete gate (simplipy.verify.verify_ruleset) subsequently killed 5
of them (rootn applied with an expression-position index that is a non-integer almost
everywhere β the left side is NaN a.e., so each rewrite is a positive-measure domain
extension). The published set is the filtered 520; the removal is recorded in the
provenance sidecar (post_filter), and the remaining rules gate clean
(486 CERTIFIED + 34 TOLERATED).
Each engines/acj-<S>-<T>/ directory is self-contained and reproducible with the same
one-command invocation shown below (deterministic β re-mining produces byte-identical
rules; for acj-4-3 the re-mine reproduces the original 525, from which the 5
post_filter rules listed in the provenance are removed). The
acj cells mine the complete source universe of the 23-operator vocabulary with the AC
engine itself as judge: a candidate rule is accepted only if it strictly descends the
engine's own reduction ordering, so nothing ships dead-on-arrival. The LLM proposal
channel is off. The provenance sidecar's core_build field pins the exact code commit
that produced the mine (<version>+g<sha>). Rule counts are small by design: the AC
engine's canonical constructors already perform arithmetic, collection, and ordering
natively, so rules only cover genuine mathematical identities.
Release gates (400-expression corpus, recorded at mining time): idempotence 0 failures, commutative-permutation invariance 0 failures on all three artifacts.
Legacy mined artifacts (simplipy β€ 0.11)
Each artifact directory engines/<S>-<T>/ is self-contained and reproducible from one
config and one command:
| Artifact | Sources β€ | Targets β€ | Rules | Mined with |
|---|---|---|---|---|
2-1 |
2 | 1 | 602 | simplipy 0.8.0 |
3-2 |
3 | 2 | 647 | simplipy 0.8.0 |
4-3 |
4 | 3 | 3,492 | simplipy 0.8.0 |
The (S, T) cell mines the complete universe of source patterns up to length S against
all candidate targets up to length T, augments the result with the certified survivors of a
fixed 740-candidate LLM proposal batch (LLM_PROPOSALS_v1.json, identical across
artifacts, sha256-pinned in each provenance sidecar), prunes rules whose effect the
remaining rules already achieve compositionally, and promotes every rule to the strongest
sound sort (_ arbitrary subtree, ! certified-finite subtree, ? variable-leaf) so each
rule fires as generally as it soundly can. Larger cells subsume smaller ones; the smaller
artifacts exist for fast loading and as fully-exhaustive reference points.
Every rule is independently verified: simplipy.verify.verify_ruleset gates each artifact
(gate-clean, 0 KILL / 0 ENGINE-MISALIGN) and simplipy.verify.monitor_ruleset sweeps the
deployed engine over an adversarial+sampled corpus (0 attributed violations) β a second
soundness authority independent of the miner.
Each artifact ships:
rules.jsonβ the mined + proposed, covered-pruned, sort-promoted rule setconfig.yamlβ the engine configuration (operator set; points atrules.json)mine.yamlβ the exact mining configurationrules.json.provenance.jsonβ everything that determines the mine: package and core build versions, host, parameters, seeds, the evaluation-matrix specification, per-length universe coverage, the proposals file's sha256 and per-outcome counts, per-length rule countsLLM_PROPOSALS_v1.jsonβ the co-located proposal batch (also published once atproposals/LLM_PROPOSALS_v1.json)
Reproduce any artifact byte-for-byte (mining is deterministic, single-command):
pip install "simplipy>=0.8.0,<0.12" # legacy artifacts; use "simplipy>=0.12" for acj-*
cd engines/2-1
simplipy find-rules -e config.yaml -c mine.yaml -o rules.json -v --reset-rules
The mine.yaml sets promote_sorts: true, so this single command mines, certifies the
proposals, prunes, and sort-promotes β the whole pipeline. Verify the result independently:
from simplipy.verify import verify_ruleset, monitor_ruleset
verify_ruleset("rules.json") # per-rule gate: only CERTIFIED/TOLERATED
monitor_ruleset("rules.json", "config.yaml", run_selftest=True) # deployed-corpus sweep
Development engines (simplipy β€ 0.11)
| Engine | Description |
|---|---|
dev_7-3 |
Development engine (sources β€ 7, targets β€ 3) used by flash-ansr |
dev_7-2 |
Development engine (sources β€ 7, targets β€ 2) |
The development engines predate the 0.5.0 certification overhaul and the 0.7.x special-point certification; they are kept for compatibility with published experiments. The mined artifacts above are the recommended starting point for new work.
LLM proposal batch
proposals/LLM_PROPOSALS_v1.json contains 740 candidate rewrite rules proposed by a
language model. Proposals carry zero authority: every candidate runs through the
identical certification chain as mined rules (find_rules(..., proposals=...)), and only
certified survivors join an artifact. The batch is a permanent, versioned input: artifacts
pin it by sha256 in their provenance.
- Downloads last month
- 17,336