Permission to use Cython files for proposed pandas I/O rdata module

[ParfaitG]

Hello pyreadr authors,

Given the growing use of Python and R as open source data science stacks, many users on various mailing lists, forums, and StackOverflow over the years have regularly asked how to import and write R data files (.rds and .rda) into and from Python pandas. Current solutions for this need is your convenient package, pyreadr, and rpy2. To streamline this data exchange directly into pandas, I recently opened a pandas ticket for a proposed IO method to read and write R data files with a small demo.

Since pandas aims to avoid optional dependencies as much as possible and since pyreadr is a Python pandas/numpy wrapper to the librdata C library, I proposed bypassing your package's Python code and only borrow your Cython code (librdata.pxd and librdata.pyx) with slight modification and topline credit to you as original authors. This IO module would then directly run librdata from its C sources. What do you think of this proposal? If users require any more extensive R data handling, then they can be referred to your package. Let me know your thoughts.

Regards,
ParfaitG

[ofajardo]

We had a similar request for my another package pyreadstat, pandas is using it for their funcition read_spss. There they are just importing pyreadstat instead of using the cython code. That would be my suggestion for pyreadr.

Of course this project is open source, so anyone is welcome to take the code and use for other purposes. But then they have to maintain that code ... I am afraid I don't have capacity to maintain a duplicated code base.

Also take into account that the license of pyreadr is very restrictive, I am not sure what kind of license pandas has, but you have to ensure that the restrictions for these pieces of code, even if they become detached from pyreadr stay as strict as they are now. You will also need to distribe the pyreadr license and attached licenses toghether with pandas license.

Regarding authorship I would also ask you to do the fist commit with my github handle so that I appear as contributor to the repo.

[ofajardo]

The discussion happened in the pandas issue.

In summary the restrictive AGPL license of pyreadr clashes with the more permissive license of pandas (BSD-3). Therefore pyreadr, neither totally or partially may be included or vendored inside pandas. If pandas would include a similar module in the future, it must be guaranteed that it is a clean-sheet implementation that directly wraps the C library without using code from pyreadr.

[ParfaitG]

Understood @ofajardo. I will do better to read licenses going forward. Glad I reached out before extensive development work. I hope we can work together in the future on any open source project. See you around the Git and Stack world!

[ofajardo]

Yep, glad to have met you! I wish you all the best in your coming projects, I am sure they are going to be interesting.