How to import data from other files within a .k file?

[kubernegit]

How can I import data from another .k-file within a .k-file?

The normal import seems to target schemas only. I haven't figured out yet how to import other data that is defined in kcl.
Using multi file compilation may lead to nicely decoupled kcl files but it also shifts the problem to another layer, so that the calling instance needs to know which files do belong together. My target would be to do something like
kcl run -S myconfig mykclfile.k and do all imports within the .k file.

[He1pa]

import can import data
a.k

foo = 1

b.k

import a
b = a.foo

Is this what you want? If you have any other questions, please give me more information.

[kubernegit]

Thanks! I guess I didn't consider accessing it via dot notation and assumed it should work by using "foo" in your example directly, like it would in multi-file compilation.

I might just need some help structuring the code. Let's say I have a big data object called config, which is stored in a separate config.k file (in the same directory as the importing file). It needs to be a single, large object so that I can create a schema for it to enforce types. After importing it, I would need to call it with config.config, which looks confusing.
If I split all the parameters up, I can call them with config.myparam, but then I lose the ability to create a schema that ensures all parameters are present. When using multi-file compilation, I can just call the data with config, but then I have to rely on the calling instance to pass the correct data. Renaming like config = config.config doesn't seem to work unless I invent a new name for config, which I also find confusing.

Do you have any suggestions?

[He1pa]

If you want to enforce data type with a schema, you can define the data object

config: Config = Config {...} # Config is a schema

In the design of KCL, normal "variables" are all exported variables, and exported variables are immutable. So you can't config = config.config.
In addition, if you don't want to output config in the result, but reference it elsewhere, you can add the prefix '_'. So you can modify the code

_config: Config = Config {} # Config is a schema

and

import config as cfg
config = cfg._config

Does this solve your problem?

[kubernegit]

Thank you. I think your method is something I can get by with.