Metadata-Version: 2.4
Name: admiral-solver
Version: 2.0.3
Summary: Python SDK for Admiral — combinatorial optimization platform
Author-email: Admiral Platform <support@admiral-platform.tech>
License: Apache-2.0
Project-URL: Homepage, https://admiral-platform.tech
Project-URL: Documentation, https://api.admiral-platform.tech/v1/docs
Project-URL: Dashboard, https://admiral-platform.tech/dashboard
Keywords: optimization,qubo,ising,solver,combinatorial
Requires-Python: >=3.10
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: numpy>=1.24
Requires-Dist: httpx>=0.25
Provides-Extra: dev
Requires-Dist: pytest>=7.0; extra == "dev"
Requires-Dist: ruff; extra == "dev"
Dynamic: license-file

# Admiral

**Combinatorial optimization platform.** 9 problem types, 27 proprietary algorithms, solver racing, AI advisor, and natural language formulation — all behind a single API.

## Install

```bash
pip install admiral-solver
```

## Quick Start

```python
from admiral import Solver

solver = Solver(api_key="adm_sk_...")

# Solve a QUBO problem
result = solver.solve_qubo(Q=[[-5, 2, 4], [0, -3, 1], [0, 0, -8]])
print(result.energy)     # -13.0
print(result.solution)   # [1, 0, 1]
```

## Problem Types

| Type | Description |
|------|-------------|
| **QUBO** | Quadratic unconstrained binary optimization |
| **Ising** | Spin glass models |
| **HUBO** | Higher-order binary optimization |
| **DQM** | Discrete quadratic models |
| **CQM** | Constrained quadratic models |
| **MAX-SAT** | Maximum satisfiability |
| **PBO** | Pseudo-boolean optimization |
| **WCSP** | Weighted constraint satisfaction |
| **Potts** | Multi-state lattice models |

## Solver Methods

```python
solver.solve_qubo(Q, timeout=30, num_reads=100)
solver.solve_ising(h, J, timeout=30, num_reads=100)
solver.solve_hubo(terms, num_variables, timeout=30)
solver.solve_dqm(domains, linear, quadratic, timeout=30)
solver.solve_cqm(c, Q, A, b, var_types, bounds, timeout=30)
solver.solve_maxsat(clauses, num_variables, timeout=30)
solver.solve_pbo(objective, constraints, num_variables, timeout=30)
solver.solve_wcsp(num_variables, domains, functions, timeout=30)
solver.solve_potts(num_nodes, num_states, edges, timeout=30)
```

## Features

- **Solver Racing** — Race multiple solvers in parallel, return the best
- **AI Advisor** — Automatic solver recommendation
- **NL Builder** — Describe problems in English, get a formulation
- **Explainability** — Understand why a solution is optimal

## Links

- [Platform](https://admiral-platform.tech)
- [Dashboard](https://admiral-platform.tech/dashboard)
- [API Docs](https://api.admiral-platform.tech/v1/docs)

## License

Apache 2.0
