gmpl-tex 0.1.1.dev1

Convert GMPL (GNU MathProg) models into LaTeX.

gmpl-tex

Convert GMPL (GNU MathProg) optimization models into LaTeX, so the constraints, sets, parameters, variables and objectives you wrote for a solver can go straight into a paper with readable, renamed symbols.

The tool works in two phases so you stay in control of the notation:

1. Lookup-table generation - parse model.mod into a JSON table listing every set, parameter, variable, constraint and objective name. Each name maps to a label you can edit.
2. LaTeX generation - render the model to LaTeX, substituting your edited labels from the JSON table.

Requirements

• Python 3.10 or newer. Check what you have:
  • Windows: py --version
  • macOS / Linux: python3 --version
  • If it's missing, install from python.org. On Windows, tick "Add python.exe to PATH" in the installer.
• lark - installed automatically.

Install

pip install gmpl-tex

After installing, the command gmpl-tex works the same on every platform.

Windows users: use py -m pip install gmpl-tex if pip isn't on your PATH.

Usage

gmpl-tex model.mod [lookup.json] --json
gmpl-tex model.mod [lookup.json] [output.tex] --latex

The example below illustrates a full run of the application.

1. Generate the editable lookup table (generates model.json)

gmpl-tex model.mod --json

For a model like

set I;
param d{i in I} >= 0;
var x{i in I} >= 0;
s.t. demand{i in I}: x[i] >= d[i];
minimize cost: sum{i in I} x[i];

model.json is:

{
    "sets":        { "I": "I" },
    "parameters":  { "d": "d" },
    "variables":   { "x": "x" },
    "constraints": { "demand": "demand" },
    "objectives":  { "cost": "cost" }
}

2. Open model.json and edit the labels you wish to rename

//Example: "cost" to "c"
{
      ...
    "objectives":  { "cost": "c" }
}

3. Render LaTeX using your edited labels (generates model.tex)

gmpl-tex model.mod model.json --latex

The following command generates a .tex file with mathematical statements in LaTeX, based on the labels in the model.json file.

Note: If you skip step 1 and run gmpl-tex model.mod --latex directly, a default table is created automatically (labels equal to the raw names) and used. If a model.json already exists next to the model, it is reused as-is and never overwritten.

An example model is included under examples/model.mod.

Troubleshooting

• Subscripted label values warning: LaTeX doesn't like multiple '_' in names. While the converter has no problem generating code with multiple name subscripts, LaTeX will eventually throw a Double subscript error on compilation. To generate valid LaTeX code, remove the '_' in any label values inside the json table. Then re-run the converter with --latex.

• A symbol shows up in red in the output: That symbol is missing from your model.json lookup table.

License

MIT - see LICENSE.