Pyinterpolate: Development Note #1
Initial note is the hardest one but I feel that it must be done sooner than later so here I am with the first note about the development of Pyinterpolate Python package. Decision to start writing about whole process is related to the fact that two days ago package has hit 50 stars on Github and for me it was shocking and positively disturbing! Three years of work has given me more than I expected and project related to the one call of European Space Agency has grown more than a single block of statistical pipeline for epidemiological study. I won’t be falsely modest – I gave it a lot, much more than “working hours”. And I won’t stop.
The problem is that many things are in my head. I’ve realized that it is not the best idea from the cooperation perspective and from the user perspective. Especially now, when each change brings a lot of new features and enhancements. I will periodically release notes like this one to give you a taste about package development.
So, what are the plans for the nearest future?
- The first and most important: testing process documentation. I’m slightly overwhelmed by the things which could be done and I need help so the thing is to make contributions as easy as possible. And I need help with multiple aspects: from simple mathematical functions, documentation, tutorials and more complex algorithms. Even newbies in programming are welcome. For more experienced developers there are problems too… multiprocessing, code optimization, package versioning and CI/CD integrations. For me it is A LOT and my time is not as flexible as it should be 🙂 (My work is not related to the spatial interpolation at this moment).
- Project webpage. I’m so eager to make it so I’ve contracted it out to my friend which is an expert in the domain of accessibility and frontend. Hopefully the first version of the webpage will be presented in August.
- Universal Kriging. This is missing piece of package which must be included in the package. Why? Because many spatial features are correlated with other spatial features and it is extremely important to catch it.
If you’re an user and those points are not satisfying you then I’m glad to inform you that I’m open for any feedback 🙂 From the user requests a multiprocessing waits in the queue 😉 But it will be done after a big code refactoring. I want to put more emphasis to the GeoPandas data structures than it is done by now.