Separating `FFTFIT` into a separate Python package?

[astrogewgaw]

I just wanted to know if it would be possible to package the code for the FFTFIT algorithm bundled with PRESTO into a separate Python package under the MIT License? PRESTO already uses f2py to bind the code to Python, but I thought it would be nice to have it as a small Python package that is independent of PRESTO's Python installation. Is it fine if I go ahead with this? I thought this might also be useful for other software, like PINT (ref: nanograv/PINT#231; while they have implemented a native implementation for FFTFIT here: nanograv/PINT#777, I still think it would be nice to have a separate package, because it could be used to test their implementation against the original code as well).

[scottransom]

I think it is fine for you to go ahead with it! I'd be happy to re-license the parts of the code that I wrote.

[astrogewgaw]

That's great @scottransom 👍🏾 ! I have already started the work here: https://github.com/astrogewgaw/fftfit. Currently the code installs without a hitch (at least on my machine) and does what I think it should. I am using the same method that PRESTO uses to bind the code to Python, via f2py and numpy.disutils. There is a single method on the Python side, also called fftfit, that is just a renamed form of the measure_phase function from the get_TOAs.py script. Right now the code is just licensed in my name, but feel free to tell me how I should change that appropriately to give credit where it is due 😅 . The repo is open to pull requests, if you want to contribute any of the changes directly. Thanks for both the prompt response, and the re-licensing 😁 !