A ternary spreadsheet in Python — a grid of cells holding balanced ternary values (−1, 0, +1) with formula evaluation, sort engine, and mutation-based autofill. This package provides the same functionality as the C implementation with the ergonomics of Python: dataclasses, type hints, and __init__.py package structure.
Python is the lingua franca of data science and machine learning. A ternary spreadsheet in Python enables interactive exploration of ternary computation via Jupyter notebooks, integration with NumPy/Pandas for data pipelines, and use as a teaching tool for balanced ternary logic. The Python version emphasizes readability and composability over raw speed — making it ideal for prototyping ternary algorithms before implementing them in C or Rust. The sort engine and mutation-based autofill make ternary genetic algorithms accessible as spreadsheet operations, bridging the gap between visual spreadsheet thinking and programmatic optimization.
The Grid class is an N×M matrix of Cell objects:
classGrid:
def__init__(self, rows: int, cols: int, default: TernaryValue=TernaryValue.NEUTRAL):
self._cells= [[Cell(default) for_inrange(cols)] for_inrange(rows)]Each Cell holds a TernaryValue (NEGATIVE, NEUTRAL, POSITIVE) and optional formula.
classTernaryValue(Enum):
NEGATIVE=-1NEUTRAL=0POSITIVE=+1Supports arithmetic: POSITIVE + NEGATIVE → NEUTRAL (truncated addition), −POSITIVE → NEGATIVE (negation).
Formulas are evaluated in dependency order:
grid.set_formula(2, 2, Formula(type=FormulaType.SUM, r1=0, c1=0, r2=1, c2=1))
grid.evaluate() # evaluates all formulas via topological sort| Formula | Semantics | Example |
|---|---|---|
| SUM | Σ values clamped to [−1, +1] | SUM(A1:B2) = clamp(sum, −1, +1) |
| PRODUCT | Π values clamped | PRODUCT(A1:A3) = clamp(prod, −1, +1) |
| THRESHOLD | +1 if Σ > t, −1 if Σ < −t, else 0 | THRESHOLD(A1:A3, 1) |
Sorts a range by ternary fitness — rows with more positive values rank higher:
sort_engine.sort(grid, r1=0, c1=0, r2=4, c2=4)Fitness per row: Σ (+1) − Σ (−1). O(N log N).
Mutated copy from a source range to a destination:
autofill.fill(grid, src=(0,0,1,4), dst=(3,0,4,4), mutation_rate=10)Each cell copies with mutation_rate% chance of a one-step ternary shift. This is a genetic operator — repeated autofill cycles evolve the grid's content.
pip install -e src/fromternary_spreadsheetimportGrid, TernaryValuegrid=Grid(5, 5)
grid.set(0, 0, TernaryValue.POSITIVE)
grid.set(0, 1, TernaryValue.NEGATIVE)
print(f"Cell(0,0) = {grid.get(0, 0)}") # POSITIVEprint(f"Cell(0,1) = {grid.get(0, 1)}") # NEGATIVE# Evaluate formulasgrid.evaluate()
# Sort by fitnessfromternary_spreadsheet.sort_engineimportsort_rangesort_range(grid, 0, 0, 4, 4)| Class/Module | Key Methods | Description |
|---|---|---|
Grid | set, get, evaluate, set_formula | Grid container and evaluator |
Cell | value, formula, has_formula | Single ternary cell |
TernaryValue | NEGATIVE, NEUTRAL, POSITIVE | Enum for ternary states |
Formula | type, range coords, threshold | Formula definition |
sort_engine | sort_range(grid, ...) | Fitness-based sort |
autofill | fill(grid, src, dst, rate) | Mutation-based fill |
The Python Ternary Spreadsheet mirrors the C version and serves as the high-level interface for ternary grid computation. In the γ + η = C framework, each cell is a ternary decision point: +1 (γ, constructive), −1 (η, avoidant), 0 (neutral). Formula evaluation propagates these values through dependency graphs, computing emergent competence C. The sort and autofill engines apply evolutionary dynamics — selection and mutation — making the spreadsheet a visual laboratory for γ/η dynamics. See ARCHITECTURE.md.
- Knuth, D. E. (1981). TAOCP Vol. 2. — Balanced ternary arithmetic.
- Chambers, J. M. (2016). Extending R. CRC Press. — On extending data manipulation systems with new types.
- McKinney, W. (2017). Python for Data Analysis, 2nd ed. O'Reilly. — Pandas DataFrame as a grid computation model.
MIT