Collection of scripts or Desktop App?

[JoshCu]

GUI first or code first?

To expand on the title some more, I'm wondering what direction to take this going forward.
My personal goal is to make this as easy as possible to install and use. This means removing the docker image needed to build dependencies from source and eventually make it as easy as pip install ngiab_data_prep with pypi CICD.
This isn't technically tricky to do, but the more the inner workings are abstracted, the more this moves away from being a python notebook/module designed to be forked and modified.

Open questions

1. Is this going to become primarily a desktop tool?
2. How important is it to keep features simple and minimal to keep the code easy to fork and make modifications to?
3. Should this move to a pip tool? something like snakeviz where you pip install it and run it without ever cloning a repo.

Based on the answers to those questions

1. If it's going to be a desktop tool should we move to kivy/qt/tkinter to create a proper desktop application?
2. Possibly port to a QGIS module?
3. Should we separate the data-processing + cli and the front-end?
4. How do we decide what features to add and their priority?

Initial thoughts

Ultimately it's up the the users who get value from this tool and I'm not going to pretend to know what a good direction for this is. I'm a proponent of simplicity, perhaps overly so, and I like the unix philosophy:

Write programs that do one thing and do it well.

This goes outside the bounds of one thing as it is a frontend, subsetter, forcings engine, and configuration-generator all in one, but there's value in the ease of installing one tool that does all the ngiab data preprocessing rather than chaining 4 together.
However, the more this expands, the trickier it becomes to comprehend and modify for personalised use-cases. Additionally it will overlap more with existing tools like DMOD and others that may be better suited / already capable of handling a richer feature set.