Skip to main content

Unit Circle Number System recursive factorization engine

Project description

ucns — Unit Circle Number System: Recursive Factorization Theory

Experimental sequence-theoretic factorization on the unit circle, with a witness-matrix recursive quotient solver.

This repository contains the UCNS (Unit Circle Number System) sequence theory and its implementation. The focus is recursive factorization: given a UCNS product object P, recover factors A and B such that A ⊠ B = P.

ucns is currently a research-stage Python package. It is suitable for experimentation, integration tests, and collaborator review. Public APIs may still change while the mathematics and implementation are being formalized.

v1.0 scope. UCNS v1.0 is a scoped, reproducible research release for catalogue-sufficient recursive factorization (Theorem N), not a claim of total general recursive primality. Carrier widening and general recursive completeness are explicitly out of v1.0 scope. See docs/ucns-spec-status-addendum-2026-05-16.md for the status vocabulary and ucns-spec.md for the reconciled spec.

Public API. ucns is the v1.0 public Python API. ucns.a0_safe is the A0-safe inspection facade. ucns_recursive is retained as a compatibility import path but is deprecated for direct user imports.


Installation

From PyPI, once released:

pip install ucns

From GitHub:

pip install git+https://github.com/The-Interdependency/ucns.git

For development:

git clone https://github.com/The-Interdependency/ucns.git
cd ucns
python -m pip install -e .[dev]
python -m unittest discover ucns_recursive/tests/ -v

Collaboration wanted

We are currently looking for mathematics collaborators to help formalize UCNS.

Start here:

The ask is bounded: help separate definitions, implemented algorithms, empirical results, proof sketches, conjectures, limitations, and counterexamples.

GPT generated; context, prompt Erin Spencer.


Status: Current Theorem Frontier

Status vocabulary (from docs/ucns-spec-status-addendum-2026-05-16.md): DEFENDED, IMPLEMENTED, TEST-BACKED, ORACLE-COMPLETE, FRONTIER, EXPERIMENTAL.

Layer Status
Flat kernel algebra DEFENDED
Depth-1 restricted completeness DEFENDED
Depth-2 oracle (Lemma 7) DEFENDED + ORACLE-COMPLETE
Full frozen depth-2 domain IMPLEMENTED + TEST-BACKED (not yet DEFENDED at spec level)
Depth-3 asymmetric (Theorem 9) TEST-BACKED (6/6 empirical)
Catalogue-sufficient completeness — all depths (Theorem N) DEFENDED — proof drafted, awaiting external formal review
Tractable sub-catalogues FRONTIER
Carrier widening FRONTIER / out of v1.0 scope
General recursive primality outside defended-complete domains out of v1.0 scope

factor_search_v08 (the witness-matrix recursive quotient solver) is the v1.0 factorization engine. It currently lives in the ucns_recursive package and is re-exported from ucns as the v1.0 public surface.

See ucns-theorem-n.md for the unified completeness theorem. The key implementation insight: factor_search_v08 is depth-agnostic — every step operates on == and plain catalogue scans. The catalogue is the only depth-sensitive input.

Prime quartet discontinuity. Cross-repo interoperability (ucns, edcmbone, a0, interdependent-lib) is not theorem continuity by default. See docs/prime-quartet-discontinuity.md and docs/edcm-edcmbone-bridge-checklist.md.

A0 rule. SEQ-PRIME is only absolute inside a defended-complete domain. A0-facing consumers should consult domain_status_metadata and treat SEQ-PRIME outside VERIFIED_DOMAIN_LABELS as non-absolute.


Repository Layout

ucns/                    # v1.0 public Python API (re-exports the engine)
  __init__.py            # from ucns import UCNSObject, multiply, factor_search_v08, ...
  a0_safe.py             # A0-safe inspection facade: identity, describe, canonical, factor
  core.py, embedding.py, epicycle.py, mobius.py, similarity.py
                         # v0.6.5-lineage modules (stable reference)

ucns_recursive/          # DEPRECATED for direct user imports; v1.0 engine implementation
  canonical.py           # UCNSObject, multiply, is_unit
  domains.py             # Frozen D' domain + payload catalogue
  domain_status.py       # Typed status vocabulary (DEFENDED, IMPLEMENTED, ...)
  host_recovery.py       # Recover host angle/face structure from P
  recursive_quotient.py  # find_left_factor, find_right_factor
  payload_system.py      # Coupled payload equation solver
  witness_matrix.py      # Witness, WitnessMatrix (global consistency)
  factor_search_v08.py   # Top-level factorization engine
  factorization_result.py  # A0-facing scoped factorization envelope
  object_record.py       # A0-facing object inspection record
  serialization.py       # Canonical JSON + stable hash
  tests/
    test_depth2_oracle.py          # Depth-2 oracle theorem (DEFENDED)
    test_depth2_full_domain.py     # Frozen depth-2 domain compact sweep
    test_failure_boundary_e109.py  # E10.9 regression tests

ucns-spec.md             # Reconciled core UCNS spec (canonical)
ucns-theorem-n.md        # Theorem N: catalogue-sufficient completeness (unified)
ucns-lemma8-depth3.md    # Depth-3 factor search (SUPERSEDED — see theorem-n)
ucns-spec-frontier-v090.md  # v0.9.0 frontier (partially superseded)
docs/
  ucns-spec-status-addendum-2026-05-16.md  # Status vocabulary + A0 rule
  pure-ucns-number-system.md
  coherence-primes-scarcity.md
ucns-code-v065.py        # Stable v0.6.5 snapshot (read-only reference)
code/                    # Exploratory artifacts (v0.8.0–v0.9.0)
code/sweeps/             # Empirical verification scripts
examples/visualization/  # Human-facing visualization boundary tests
  seed53.html            # 53-residue skip-star + heptagram + unwrap demo
  README.md              # claim linkage, non-proof boundary, open constraints

What belongs in this repo (and how to place visual demos)

This repository is primarily for UCNS mathematics, Python implementation, tests, and reproducible research artifacts.

Interactive front-end demos (including single-file HTML/JS/SVG sketches) can belong here when they function as research support artifacts: they must clarify a theorem/mechanism, document a failure boundary, or support a reproducibility workflow.

If you add one, place it under examples/visualization/ and include a short README that states:

  • the exact UCNS claim/theorem/domain status it illustrates;
  • what the demo does not prove (honest boundary of inference);
  • how it acts as a boundary object for unresolved constraints (what remains open, and what transition it marks between delivered artifact and ongoing research).

If the artifact is primarily outreach/showcase and not directly tied to UCNS verification or documentation, publish it as a separate GitHub Pages micro-app and link it from docs.

ucns_recursive is deprecated for direct user imports as of v1.0 canon reconciliation. New code should import from ucns and ucns.a0_safe. ucns_recursive remains a supported compatibility import path (no runtime warning yet); it is also where the engine implementation currently lives.


Core Algebra

Every UCNS object is a sequence of (angle, payload) pairs with a face-flip sequence:

from ucns import UCNSObject, multiply, is_unit
from fractions import Fraction

UNIT = None

# S2: the canonical depth-0 sequence object
S2 = UCNSObject(2, 2, [(Fraction(0), UNIT), (Fraction(1), UNIT)], [0, 0])

# Depth-1 object: A carries S2 as payload in its first cell
A = UCNSObject(2, 2, [(Fraction(0), S2), (Fraction(1), UNIT)], [0, 0])
B = UCNSObject(2, 2, [(Fraction(0), S2), (Fraction(1), UNIT)], [0, 0])

# Product
P = multiply(A, B)

Factorization: factor_search_v08

from ucns import factor_search_v08

result = factor_search_v08(P)
# Returns (A_recovered, B_recovered)  or  "SEQ-PRIME"

Returns a valid factorisation — the first one found under the loop ordering (balanced p ≥ 2 splits first, p = 1 last). Factorisation is not generally unique; other valid pairs may exist. Use store.factor_decompose with an explicit catalogue to enumerate all catalogue-bounded factorisations.

The solver implements the full witness-matrix pipeline:

  1. Host recovery — extract candidate A/B angle sequences from P
  2. Payload system construction — build the p×q coupled equations
    multiply(S_A[k], S_B[j]) == P_payloads[k][j]
  3. Witness-matrix consistency — verify one globally consistent payload assignment explains every cell
  4. Face recovery — enumerate valid face-bit assignments
  5. Exact recomposition — final truth test: multiply(A_cand, B_cand) == P

For depth-3+ targets, extend the catalogue with the deep payloads of the expected factors (see ucns-theorem-n.md §4.2–4.3):

# Theorem 9 example: depth-3 A × depth-2 B
from ucns import depth_of

def catalogue_from(*objs):
    """Minimal catalogue: recursive payload closure of given objects."""
    cat = [None]
    def collect(o):
        if o is None: return
        for _, p in o.A_plus:
            if p is not None and p not in cat:
                cat.append(p); collect(p)
    for o in objs: collect(o)
    return cat

result = factor_search_v08(P, catalogue_from(A, B))

Running the Tests

python -m unittest discover ucns_recursive/tests/ -v

Build and release validation

python -m pip install -e .[dev]
python -m build
python -m twine check dist/*
python -m unittest discover ucns_recursive/tests/ -v

Clean wheel install smoke test:

python -m venv /tmp/ucns-wheel-test
. /tmp/ucns-wheel-test/bin/activate
pip install dist/*.whl
python - <<'PY'
from ucns import UCNSObject, multiply, factor_search_v08
print('ucns import ok')
PY

Root Cause Fixed (E10.9)

The v0.8.0 failure analysis identified three root causes now corrected in factor_search_v08:

  1. No false atomicity — depth-1 payloads such as S2 are descended into recursively, not treated as atomic
  2. Global witness consistency — a single assignment of all payload factors must explain every cell simultaneously
  3. Staged reconstruction — host recovery → payload system construction → witness verification

Completeness Frontier

UCNS has a DEFENDED flat kernel, a DEFENDED depth-1 restricted completeness theorem, and a DEFENDED + ORACLE-COMPLETE depth-2 oracle theorem (Lemma 7). These are instances of Theorem N (catalogue-sufficient factorization, ucns-theorem-n.md): if the catalogue contains every payload of the true factors, factor_search_v08 finds a factorization. Depth enters only through catalogue selection, not through the algorithm.

The full frozen depth-2 domain is IMPLEMENTED + TEST-BACKED in factor_search_v08 but not yet DEFENDED at the spec level. Theorem 9 (asymmetric depth-3) is TEST-BACKED empirically; see code/sweeps/t9_minimal_cat.py. Carrier widening remains FRONTIER and out of v1.0 scope.


Accreditation: GPT generated from context provided by Grok, Claude as prompted by Erin Spencer.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ucns-0.8.3.tar.gz (79.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

ucns-0.8.3-py3-none-any.whl (83.2 kB view details)

Uploaded Python 3

File details

Details for the file ucns-0.8.3.tar.gz.

File metadata

  • Download URL: ucns-0.8.3.tar.gz
  • Upload date:
  • Size: 79.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.13

File hashes

Hashes for ucns-0.8.3.tar.gz
Algorithm Hash digest
SHA256 3bd09bd3de014485c67ff44fc162aee994bd83e860597aa9a9424dcfcf2b3677
MD5 f9d1b37c76434c45a8e892e17e726cbf
BLAKE2b-256 11cead756e83a0da1002f2b3032b4a35e77df82936f6296a4fb7107e16bf0ef1

See more details on using hashes here.

File details

Details for the file ucns-0.8.3-py3-none-any.whl.

File metadata

  • Download URL: ucns-0.8.3-py3-none-any.whl
  • Upload date:
  • Size: 83.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.13.13

File hashes

Hashes for ucns-0.8.3-py3-none-any.whl
Algorithm Hash digest
SHA256 ffc12f9448f830fcf87c2ecd21ddf94d0be6b80b271babd01aea1ef18ac22f09
MD5 20de3ae765bccacff810741946670066
BLAKE2b-256 7ff43f26f764f6d1f6126946bc223ce5773a5708f67e4d1adc7131b0e41f3c29

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page