Support for input data validation

[bala-pitchuka]

Hi,

Just want to check if there is an existing mechanism to check that all mandatory fields are provided to zen engine without missing during evaluation.

like a sample rule if threshold > 23 then HOLD

if threshold is not provided as input to zen then them some validation error need to be thrown

Also how to check the json is a valid jdm json before execution currently facing following issue when run, just want to eliminate such issues during rule jdm json creation itself

RuntimeError: {"type":"NodeError","nodeId":"","source":"Graph did not halt. Missing output node."}

thnks

[stefan-gorules]

Hi @bala-pitchuka,

We have some initial plans of looking into type-safety in the future. Our foundation will rest on a dynamic language - ZEN Expressions that will be enhanced with type support at some point.

For now, most common pattern is to do validation of the data before sending it to the rules engine. Depending on the language used, there are some great options: Zod (TypeScript), Validator (Go), validator (Rust crate), etc.

[stefan-gorules]

Hi @bala-pitchuka

This is something that will be possible with Function v2 through usage of Zod. We will provide an example in docs soon.

[Aakarshan-369]

Hi @stefan-gorules !

I have been exploring the ZenEngine library in Node for a in-house BRMS prototype. Firstly great work on the library! it has intuitive semantics and is very performant as well.
I am having trouble figuring out how to validate a string as a valid Zen Expression syntactically. Before re-instantiating a engine with a new schema I wanted to make sure the changes in a rule row's inputs and outputs are valid Zen syntax.

if I'm approaching this incorrectly please suggest a way!

[bala-pitchuka]

Hi @bala-pitchuka

This is something that will be possible with Function v2 through usage of Zod. We will provide an example in docs soon.

@stefan-gorules docs would be great to explore the same, thnks

[Gabrieltay]

hi @bala-pitchuka , with Function v2, just an simple example here

import { z } from "zod";

/* input
{
  age: 20,
  salary: 3000.25,
  name: "John"
}
*/
export const handler = async (input) => {
  const schema = z.object({
    age: z.number()
      .min(15, { message: "Age must be at least 15" })
      .max(60, { message: "Age must be at most 60" }),

    salary: z.number()
      .nonnegative({ message: "Salary must be non-negative" })
      .refine(value => value.toFixed(2) === value.toString(), {
        message: "Salary must have up to 2 decimal places"
      }),
  });

  const result = schema.safeParse(input);
  if (result.success) {
    return result.data
  } else {
    console.log(result.error)
    return {}
  }
};