algo2code¶

algo2code transpiles algorithm boxes written in LaTeX algpseudocode into executable code (Taichi). Where mechdsl-core derives constitutive math from LaTeX, algo2code derives algorithms — the iterative solvers and return-mapping loops that don't come from differentiating an energy.

It is deliberately tiny: zero runtime dependencies (stdlib only), and it never imports mechdsl. The relationship is strict producer/consumer — mechdsl-core consumes algo2code-generated artifacts, not the other way around.

It is on PyPI under the MIT license and installs standalone — no mechdsl-core needed:

pip install algo2code

Why a separate package¶

Dissipative models (J2 plasticity, viscoplasticity) and linear solvers (PCG) are algorithms, not closed-form expressions. You can't sympy.diff your way to a return-mapping Newton loop. So MechDSL authors those algorithms once, in LaTeX algpseudocode, and transpiles them — keeping the algorithm in the paper identical to the algorithm that runs.

Because the package has no runtime dependencies, it's also usable on its own: point it at any algpseudocode block and get a deterministic Python/Taichi function back.

Pipeline¶

algpseudocode (LaTeX)
   │  algo_parser        parse \State / \For / \If / \Return into an Algorithm AST
   ▼
   │  expr_parser        parse the math expressions in each statement
   ▼
   │  type_inference     infer scalar/array types for declared args + scratch vars
   ▼
backends/taichi_codegen  emit a Taichi-compatible Python function

The emitted output is deterministic (golden-stable): transpiling the same source twice yields byte-identical code, which is what lets the codegen be regression-tested. The authoritative reference is dev/design_docs/11-ALGO2CODE.md.

Where to go next¶

Installation — every package, extra, and the from-source workflow.
Getting started — pip install, then transpile your first algorithm.
Usage — the transpile API, the algorithm library, and the solver seam.
Examples — runnable transpile snippets for the J2 family and PCG.