AIP Logic in Python?

Hi, I’ve tried to port some functioning Python business logic into AIP Logic, but my dev account only shows TypeScript and NoCode options. Has anyone had access to a Python option? Has that option been implemented yet?

are u talking about Code repository ? are not able to create python code repo’s ?

The way it seems to me, I have access to data transformations in python (code repository) but no to logic functions in python to operate over ontology objects. Am I talking nonsense?

So you want to create a AIP logic function that uses ontology object. Then use that logic function in python transform where you want to pass an object ?

is this because you want to run the AIP logic function on several objects as a transform ?

Yes, it is because I want to run a “complicated” business logic that combines many ontology objects and reflect the results on a workshop app. I tested a simple logic loop (workshop input >> logic function in TypeScript >> workshop output) and it worked, but I would feel much more comfortable doing my application logic functions in Python

What do you mean by
(workshop input >> logic function in TypeScript >> workshop output)

Did u create a AIP logic function that has input as an object output maybe as a string for simplicity, then publish it as a function. Then create a typescript code repo and call the AIP logic function from within the type script function ?

I dont see documentation that shows how to do it. I asked AIP assist and its not helping either.

As per this Doc Use a Logic function

  1. You can back an action with a Logic function, then call the action from Workshop.
  2. You can also call a Logic function to back a Markdown widget in Workshop; in this case, the output type from the Logic function must be a string.
  3. You can call a Logic function in other Logic functions, as well as in functions on objects, via the Ontology function tool in AIP Logic.

But the Doc doesn’t show how to use a AIP logic function in a type script function.

" it is because I want to run a “complicated” business logic that combines many ontology objects and reflect the results on a workshop app"

Does this involve using LLM ? I think based on the amount of features that are added to AIP logic, I feel Palantir’s goal is to make AIP logic a IDE where you can build complicated logics that can be executed without writing Python code to invoke LLM’s.

I have written pretty complicated logicx that use object query + link traversal + external function calling all in AIP logic and it has worked perfectly fine.

It would be interesting to understand the shortcoming of AIP logic that would force a use case to be written in Python transform as that would be a feedback for the platform team to improve the product.

I don’t want to use LLMs in my logic function, just want to use a string as input from workshop, a bunch of joins, groupby, aggregations, and transformations involving several ontology object, and output an object that can then be used for visualizations back in workshop. And I would like to do that in Python instead of TypeScript or NoCode

ok got it. Thanks for the clarification, so ur concern is Function only supports type script, Here is the Beta feature that support use of python with functions.

I havent tested it yet, might need to work with the platform admin or your Palantir team to get it enabled on your stack.

Yes, exactly, I’m sorry I didn’t make myself clear the first time — still quite new on this platform. Thank you so much.