mathoptdev 0.1.5

Python CLI and SDK for mathopt.dev the platform for mathematical optimization

MathOpt

To interact with mathopt, we provide a CLI and a Python SDK. It contains tools to help developers solve combinatorial problems in the cloud.

Installation

Install with pip

pip install mathoptdev

Install with uv (recommended package manager):

uv add mathoptdev 

This installs a CLI that you can call with:

opt --help 

or

alias opt="uv run opt" # set and alias for uv
opt --help 

Login

This command will guide you through the login process:

opt login 

We store the API token in the MATHOPT_API_TOKEN variable in your .env file.

After logging in, check the user with

opt user 

REST API

You can call the API directly with a HTTP POST request. You need to specify the x-mathopt-api-token header.

curl -X POST https://www.mathopt.dev/api \
  -H "x-mathopt-api-token: YOUR_MATHOPT_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "action": "get_user"
  }'

We also have a helper function in python

import mathoptdev as opt 
body = {
  "action": "get_user"
}
opt.send_request(body) 

Create problem instances

To get you started, we provide the following helper command that generates a Travelling Salesman Problem (TSP):

opt tsp -c 5 

The option -c corresponds to the number of cities in the problem. The MPS file will be saved in the tsp folder in your current directory.

Once instances are generated, upload them with the following command:

opt instance create tsp 

To verify that we successfully processed your instances, you can query the backend with the command:

opt instance list 

or in python

import mathoptdev as opt 
opt.queries.get_instances() 

Solve instances

To solve an instance, you need to create a job with its id and the id of the strategy you want to use to solve it.

The list of available strategies can be queried with

opt strategy list 

We provide an example on how to create and queue a set of jobs in these scripts:

• create_jobs.py
• queue_jobs.py

Monitor your jobs

You can monitor your jobs with:

opt job list 

or

import mathoptdev as opt 
opt.queries.get_jobs() 

Download solutions

You can download the parquet file associated with a solution:

import mathoptdev as opt 
body = {
  "action": "download_solution",
  "solution_id": "my_solution_id"
}
opt.send_request(body) 

There is a full example at the link: download_solutions.py