In my experience it is best to avoid letting definitions get that big in the first place. Once a gh file gets really large, it’s pretty unlikely it will solve interactively, so there’s not a whole lot of benefit keeping it all together.
I’m a big fan of this approach, Elefront enables you to break projects up into multiple files. For example an architectural project might be structured like this:
- massing.gh
- panelization.gh
- rationalisation.gh
- fabrication.gh
where the input for each gh is the previous file. Since Elefront allows you to re-reference your data-structure as it was when it was baked in the previous step, the hassle of re selecting inputs disappears. You can think of each gh file as a node in your project graph as this paper describes.
I find breaking large projects up into multiple files makes maintenance,collaboration and re-use much easier.