Hi, a noob question. Is there an easy way to reuse...

# datascience

[Mykola Gurov]

Hi, a noob question. Is there an easy way to reuse a bit of experimental code between the notebooks? So far I see two options:
- publish a (local) library. A bit cumbersome for quick adjustment loops. Good for when this code matures enough in the future.
- use a lib descriptor json with a bunch of init/initCell. Kinda OKish, but gets messy quick with multiline extension functions using string interpolation :)

[Mykola Gurov]

And while we’re on this topic, what’s the difference between init and initCell? They seem to work the same for me.

[altavir]

You can load local jar.

[holgerbrandl]

What about @file:Import ? Has this landed in the scripting APIs already, and could this be used in a jupyter notebook as well?

[altavir]

It is always problematic to trace dependencies for stand-alone files (they need to be kts scripts for that).

I just tested and @file:DependsOn("myJar.jar") is working perfectly fine.

[holgerbrandl]

Oh, I did not say that it would be good practice. I was just thinking about options.

[holgerbrandl]

Thanks for confirming the option! 🙂

[altavir]

Actually, the documentation says that class-directorles should work as well: https://github.com/Kotlin/kotlin-jupyter#dependencies-resolving-annotations

[holgerbrandl]

Nice

[altavir]

But I think it meanse compiled classes