Closed APN-Pucky closed 7 months ago
Yeah, this is intentional. I also was confused about this initially, but it is mentioned in the readme.
I think technically it might be possible to distribute on PyPI, but it would be quite inconvenient since it's 350+ MB of data. It only downloads the particular files that it needs from GitHub, so it's a lot more efficient than downloading all data.
As a consequence network-sandboxed tests fail even if scikit-hep-testdata is installed from pypi. I'll use the github sources for now, but wanted to let you know about it.
I guess this is intentional since the whole data would be too much to distribute over pypi?