Python API

gridcalc's engine is usable without either frontend. A Grid loads a workbook, recalculates it, and hands back cell values; the solver and goal-seek modules operate on that same Grid.

from gridcalc.engine import Grid

g = Grid()
g.jsonload("examples/example_lp.json")
g.recalc()
print(g.value(1, 3))

These pages cover the modules that make up that public boundary:

Module	What it holds
engine	Grid, Sheet, Cell, Vec, NamedRange -- the workbook model, both recalculation engines, and JSON load/save
formula	The Excel formula language: lexer, parser, AST, evaluator, dependency extraction, error values
opt	Building an LP, MIP, or QP from sheet cells and solving it, with sensitivity, sweeps, and diagnosis
goalseek	One-dimensional root finding over a Grid, with auto-bracketing
config	gridcalc.toml loading and the parsed Config
sandbox	AST validation, module classification, and workbook inspection without execution

Deliberately not documented here: gridcalc.tui (the curses layer), gridcalc.web (the webview bridge), and gridcalc.libs.xlsx (the ~310-entry function library). They are internal to their frontends or, in the library's case, better described by the function coverage audit than by a wall of one-line signatures.

The pages are generated from the source with mkdocstrings, so the docstrings here are the ones in the code.