russell@unturf.com
70ecda3d6c
arborist/qa/nli/ — SHADOW ONLY (never an audit_mode input; manifest not yet in governance_policy_hash per §7 #2). manifest.json pins cross-encoder/nli-MiniLM2-L6-H768 @ a fixed HF revision + the bench-validated θc 0.5/θe 0.9 + 2 alternates + the Phase-3 TODO; shadow.py = ShadowNLI/shadow_check (lazy transformers+torch behind a new [nli] extra, clauses() segmenter, the §7 #5 clause-level Demote() decision, degrades to available=False when [nli] absent); bench/scripts/nli_shadow_sweep.py + make bootstrap-nli / bench-nli-shadow (the gate-item-4 instrument); 16 tests. First sweep (116 records — 5f-falsification packs + the arborist-nli-bench eval sets): 28/28 synth recombination demoted, 0/26 FP on legit summaries, 0/9 fires on already-STRICT_SPAN records, 25/50 on UNGROUNDED (the contradiction half; quiet on non-sequiturs). Gate items 1/2/3/5/6 clear on available data; item 4 — shadow FP rate on a real live-bench-qa sample — remains the open measurement. Production verifier unchanged; falsification-hard stays 10/12.
117 lines
3.6 KiB
TOML
117 lines
3.6 KiB
TOML
[build-system]
|
|
requires = ["setuptools>=68"]
|
|
build-backend = "setuptools.build_meta"
|
|
|
|
[project]
|
|
name = "arborist"
|
|
version = "0.0.1"
|
|
description = "An arborist for trees and forests of cross-linked information"
|
|
readme = "README.md"
|
|
license = { text = "AGPL-3.0-only" }
|
|
requires-python = ">=3.10"
|
|
authors = [
|
|
{ name = "Russell Ballestrini", email = "russell@unturf.com" },
|
|
{ name = "foxhop" },
|
|
{ name = "TimeHexOn" },
|
|
]
|
|
dependencies = [
|
|
"httpx>=0.27",
|
|
"zstandard>=0.22",
|
|
"cryptography>=42",
|
|
]
|
|
|
|
[project.optional-dependencies]
|
|
html = [
|
|
"selectolax>=0.3",
|
|
]
|
|
wikitext = [
|
|
"mwparserfromhell>=0.6",
|
|
]
|
|
mesh = [
|
|
# httpx is already in core deps; mesh wire only depends on stdlib +
|
|
# cryptography (also core). This extras block exists as the documented
|
|
# opt-in surface even though no extra packages are required today.
|
|
]
|
|
math = [
|
|
# Symbolic algebra/calculus π* substrate (ticket #000030). SymPy is
|
|
# ~30 MB installed; pulling it into core deps would inflate every
|
|
# fresh checkout. Tests skip via pytest.importorskip when absent.
|
|
"sympy>=1.13",
|
|
]
|
|
hessian = [
|
|
# Phi_alignment_probe (ticket #000034 Phase 1a). Lanczos top-k +
|
|
# bottom-k eigendecomposition for measuring whether v7's frozen
|
|
# linear projection W aligns with the loss Hessian's low-eigenvalue
|
|
# subspace. Numpy + scipy together ~80 MB; gated separately from
|
|
# core to keep the default install lightweight. Tests skip via
|
|
# pytest.importorskip when absent. Install with:
|
|
# pip install 'arborist[hessian]'
|
|
"numpy>=1.26",
|
|
"scipy>=1.11",
|
|
]
|
|
crawler = [
|
|
# Verbatim lift from agents.ai.unturf.com/core. Off by default — the
|
|
# default test suite never imports the crawler. Install with:
|
|
# pip install 'arborist[crawler]'
|
|
# then run `make test-crawler`.
|
|
"aiohttp>=3.8",
|
|
"beautifulsoup4>=4.11",
|
|
"lxml>=4.9",
|
|
"html5lib>=1.1",
|
|
"html2text>=2024.2.26",
|
|
"miniuri>=1.1",
|
|
"feedparser>=6.0",
|
|
"Pillow>=10.0",
|
|
"cairosvg>=2.7",
|
|
"pypdf>=4.0",
|
|
]
|
|
vec = [
|
|
# Optional sqlite-vec semantic retrieval backend (ticket #000039).
|
|
# sqlite-vec ships only the loadable SQLite extension (~1 MB);
|
|
# fastembed pulls onnxruntime + tokenizers + huggingface-hub
|
|
# (~150 MB) and downloads the bge-small-en-v1.5 ONNX model
|
|
# (~130 MB) on first use. Gated separately so a fresh checkout
|
|
# stays python3.12 + venv + sqlite3. CLI surfaces `arborist embed`
|
|
# / `--backend vec` only when `sqlite_vec` imports. Install with:
|
|
# pip install 'arborist[vec]'
|
|
# (sentence-transformers is the heavier "official" embedder path
|
|
# the ticket §5 names; fastembed is the lightweight ONNX one.)
|
|
"sqlite-vec>=0.1.9",
|
|
"fastembed>=0.4",
|
|
]
|
|
nli = [
|
|
# Sentence-pair NLI for the #000049 Phase-2 *shadow* path
|
|
# (arborist/qa/nli/) — measures whether a clause-level contradiction
|
|
# veto would demote a weakly-grounded answer; never touches
|
|
# audit_mode. transformers + a CPU torch is ~600 MB installed, so it
|
|
# is gated hard out of core / dev — a fresh checkout stays
|
|
# python3.12 + venv + sqlite3, and the default test suite skips the
|
|
# NLI tests via pytest.importorskip when this extra is absent.
|
|
# Install with:
|
|
# pip install 'arborist[nli]'
|
|
# Phase 3 (if it happens) should ONNX-export the pinned checkpoint
|
|
# and switch this to onnxruntime-cpu to drop torch (cf. [vec]).
|
|
"transformers>=4.40",
|
|
"torch>=2.2",
|
|
"sentencepiece>=0.2",
|
|
"protobuf>=4.0",
|
|
]
|
|
dev = [
|
|
"pytest>=8",
|
|
"pytest-asyncio>=0.23",
|
|
"pytest-xdist>=3.5",
|
|
"arborist[html]",
|
|
"arborist[wikitext]",
|
|
"arborist[mesh]",
|
|
"arborist[crawler]",
|
|
"arborist[math]",
|
|
"arborist[hessian]",
|
|
"arborist[vec]",
|
|
]
|
|
|
|
[project.scripts]
|
|
arborist = "arborist.cli:main"
|
|
|
|
[tool.setuptools.packages.find]
|
|
where = ["."]
|
|
include = ["arborist*"]
|