Metadata-Version: 2.4
Name: constella-gpu-backend
Version: 0.1.2
Summary: Constella accelerator monitoring backend, agents, and HTTP/WebSocket APIs.
Project-URL: Homepage, https://github.com/kuma-loong/Constella
Project-URL: Repository, https://github.com/kuma-loong/Constella
Project-URL: Documentation, https://github.com/kuma-loong/Constella#readme
Project-URL: Issues, https://github.com/kuma-loong/Constella/issues
License-Expression: MIT
License-File: LICENSE
Keywords: api,ascend,fastapi,gpu,monitoring,nvidia
Classifier: Development Status :: 3 - Alpha
Classifier: Environment :: Console
Classifier: Framework :: FastAPI
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: POSIX :: Linux
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: System :: Monitoring
Requires-Python: >=3.10
Requires-Dist: fastapi>=0.116.0
Requires-Dist: pyyaml>=6.0.0
Requires-Dist: uvicorn[standard]>=0.35.0
Requires-Dist: websockets>=13.0
Description-Content-Type: text/markdown

# Constella Backend

`constella-gpu-backend` is the standalone Constella service package. It provides
the manager, local and remote accelerator agents, SQLite workload history,
high-resolution job curves, and the HTTP/WebSocket API without installing a
user interface.

Constella natively supports heterogeneous accelerator clusters. Version 0.1.2
supports NVIDIA GPUs through NVML with an `nvidia-smi` fallback, and Ascend NPUs
through DCMI with an `npu-smi` fallback.

## Install and run

```bash
pip install constella-gpu-backend
constella service start --no-local-agent
```

Start a local NVIDIA or Ascend agent with the service:

```bash
constella service start --device nvidia
constella service start --device ascend
```

The backend exposes its API documentation at `/docs` and OpenAPI schema at
`/openapi.json`. This package intentionally contains no Web or TUI frontend.

Other distributions are available for the [Web UI](https://pypi.org/project/constella-gpu-web/),
[TUI](https://pypi.org/project/constella-gpu-tui/), and
[full installation](https://pypi.org/project/constella-gpu/).

Source and documentation: [github.com/kuma-loong/Constella](https://github.com/kuma-loong/Constella)
