Core math
Visible precision and local calculus
Exact visible precision, canonical lift, hidden projector, and local visible calculus are part of the public kernel.
Python Kernel - Exact Identities - Reproducible Examples
A small Python kernel for observer geometry.
nomogeo focuses on a narrow, exact part of the stack: visible precision, local visible calculus, hidden-load geometry beneath a fixed ceiling, and a small quotient-side Gaussian layer. The goal is simple: make the core mathematics usable without pretending the package is broader than it is.
Python >= 3.11
BSD-3-Clause
PyPI
Minimal dependencies
Overview
What the package includes
Hidden layer
Fixed-ceiling hidden-load geometry
The package exposes the ceiling-conditioned inverse surface, hidden transport, determinant clock, and contraction-factor composition.
Public examples
Released demos with clear scope
Examples are published with commands, outputs, and boundaries so readers can see what each one does and does not claim.
Start Here
Documentation
Quick Example
import numpy as np
from nomogeo import visible_precision, canonical_lift
H = np.array([[3.0, 1.0], [1.0, 2.0]])
C = np.array([[1.0, 0.0]])
phi = visible_precision(H, C)
lift = canonical_lift(H, C)
print(phi)
print(lift)Examples
Released examples
These are the three public examples in the current release. Each one has a command, a declared scope, and frozen outputs.
Exact identity
Entanglement as hidden load
Shows the executable Gaussian identity tau = 2 I(A:B) on the stated families.
Structural separation
Bell / common-gluing
Shows that correlator compatibility can hold while full-law Gaussian common gluing fails.
Open exampleSynthetic worked example
Arrow rank deficiency
Compares how different voting observers erase different latent preference directions.
Open exampleImportant note
The inverse theorem is not global.
inverse_visible_class(T, Lambda, ...) is the fixed-ceiling inverse surface. It is not an inverse of the global map (H, C) -> Phi_C(H).
Important note
Long hidden composition uses contraction factors.
For long chains, use hidden_contraction(...) and load_from_hidden_contraction(...). Raw hidden loads are not the associative object.