Class

Classes consist of a name, a list of properties, and their types. In the context of LLMs, classes describe the type of the variables you can inject into prompts and extract out from the response.

from pydantic import BaseModel
from path.to.bar import Bar
from path.to.my_enum import MyEnum

class Foo(BaseModel):
  property1: str
  property2: Optional[int]= None
  property3: List[Bar]
  property4: MyEnum

import z from "zod";
import { BarZod } from "./path/to/bar";
import { MyEnumZod } from "./path/to/my_enum";

const FooZod = z.object({
  property1: z.string(),
  property2: z.number().int().nullable().optional(),
  property3: z.array(BarZod),
  property4: MyEnumZod,
});

type Foo = z.infer<typeof FooZod>;

Field Attributes

When prompt engineering, you can also alias values and add descriptions.

<ParamField path="@alias" type="string"

Aliasing renames the field for the llm to potentially "understand" your value better, while keeping the original name in your code, so you don't need to change your downstream code everytime.

This will also be used for parsing the output of the LLM back into the original object. </ParamField>

<ParamField path="@description" type="string"

This adds some additional context to the field in the prompt. </ParamField>

class MyClass {
  property1 string @alias("name") @description("The name of the object")
  age int? @description("The age of the object")
}

Class Attributes

<ParamField path="@@dynamic"

If set, will allow you to add fields to the class dynamically at runtime (in your python/ts/etc code). See dynamic classes for more information. </ParamField>

class MyClass {
  property1 string
  property2 int?

  @@dynamic // allows me to later propert3 float[] at runtime
}

Syntax

Classes may have any number of properties. Property names must follow these rules:

• Must start with a letter
• Must contain only letters, numbers, and underscores
• Must be unique within the class
• classes can be recursively defined!

The type of a property can be any supported type

Default values

• Not yet supported. For optional properties, the default value is None in python.

Dynamic classes

See Dynamic Types.

Inheritance

Never supported. Like rust, we take the stance that composition is better than inheritance.