Cool package! However, when I "pip installed" dfply, the diamonds.csv flie didn't download as well, and I got import errors, since the file gets read in as part of the package load process. I took a look at the packaging, and tried to help out by making some changes that would make it more pip-compatible and would (I hope--I'm still a newbie at packaging) simplify the growth of the project by making "six" and "pandas_ply" explicit global requirements that could be downloaded during installation.
This pull request:
Moves data out of python package,
added setup.py and MANIFEST.in links to data
made data.py a module that loads the data, to keep code working without API change.
Added pandas_ply and six to requirements, to remove vendor code from python package as well.
I'm sorry that I made all these into a single commit (stupid, I know). If there are any fixes you'd like me to make, please let me know!
Cool package! However, when I "pip installed" dfply, the diamonds.csv flie didn't download as well, and I got import errors, since the file gets read in as part of the package load process. I took a look at the packaging, and tried to help out by making some changes that would make it more pip-compatible and would (I hope--I'm still a newbie at packaging) simplify the growth of the project by making "six" and "pandas_ply" explicit global requirements that could be downloaded during installation.
This pull request:
I'm sorry that I made all these into a single commit (stupid, I know). If there are any fixes you'd like me to make, please let me know!