Metadata-Version: 2.4
Name: harbr
Version: 0.0.0
Summary: Git-like dataset versioning and resumable transfer for your own S3 bucket.
Project-URL: Homepage, https://harbr.cloud
Project-URL: Source, https://github.com/harbr/harbr
Author-email: Harshit Verma <therealharshit014@gmail.com>
License-Expression: MIT
Keywords: datasets,machine-learning,s3,storage,versioning
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.9
Description-Content-Type: text/markdown

# harbr

Git-like dataset versioning and resumable transfer for your own S3 bucket.

The HuggingFace-Hub developer experience, pointed at storage you control. Push, pull,
and version ML datasets and models on any S3-compatible store (AWS S3, Cloudflare R2,
RustFS) with content-addressed dedup and resumable transfers. Open source, MIT.

> Under active development. This is an early placeholder release.
> Follow progress at https://harbr.cloud

## Planned

```python
import harbr

b = harbr.connect(bucket="my-data")     # uses your existing AWS credentials
b.push("imagenet/", name="imagenet", version="v2")
b.pull("imagenet", version="v2")        # -> local files, byte-identical
```

## License

MIT
