Using ActionWeaver with Pydantic V2

This guide is designed to demonstrate use cases using Pydantic along with ActionWeaver.

%load_ext autoreload
%autoreload 2

import os

from typing import List
from uuid import UUID, uuid4

from actionweaver import action
from actionweaver.utils.tokens import TokenUsageTracker
from actionweaver.llms import wrap
from actionweaver.actions.factories.pydantic_model_to_action import  action_from_model

from pydantic import BaseModel, Field, PrivateAttr, validate_call

from openai import OpenAI

from datetime import datetime

Prerequisite: Patch OpenAI client with ActionWeaver

from openai import AzureOpenAI

MODEL="gpt-35-turbo-0613-16k"
client = AzureOpenAI(
    azure_endpoint = os.getenv("AZURE_OPENAI_ENDPOINT"),
    api_key=os.getenv("AZURE_OPENAI_KEY"),
    api_version="2023-10-01-preview"
)



client = wrap(client)

Use Case: Structure parsing using Pydantic V2

class User(BaseModel):
    _uid: UUID = PrivateAttr(default_factory=uuid4)
    _created_at: datetime = PrivateAttr(default_factory=datetime.now)
    name: str = Field(..., description="Name of the user")
    age: int = Field(..., description="Age of the user", gt=0)

class Users(BaseModel):
    users: List[User]

create_users = action_from_model(Users, stop=True)
create_users.json_schema()

{'$defs': {'User': {'properties': {'name': {'description': 'Name of the user',
     'title': 'Name',
     'type': 'string'},
    'age': {'description': 'Age of the user',
     'exclusiveMinimum': 0,
     'title': 'Age',
     'type': 'integer'}},
   'required': ['name', 'age'],
   'title': 'User',
   'type': 'object'}},
 'properties': {'users': {'items': {'$ref': '#/$defs/User'},
   'title': 'Users',
   'type': 'array'}},
 'required': ['users'],
 'title': 'Users',
 'type': 'object'}

users = create_users.invoke(client, messages=[{"role": "user", "content": "Niel and Tom are 31 years old, James is 25 years old"}], model=MODEL, stream=False, force=False)

users

Users(users=[User(name='Niel', age=31), User(name='Tom', age=31), User(name='James', age=25)])

Use Case: You can also use Pydantic models as function parameter.

The @validate_call decorator allows the arguments passed to a function to be parsed and validated using the function’s annotations before the function is called.

user_db = []

@action(name="InsertUsers", stop=True)
@validate_call
def insert_users(users: Users) -> str:
    """Inserts user data into the database and returns a success message."""
    # Add the provided users to the user database
    user_db.extend(users.users)

    return f"Inserted {len(users.users)} users ."

messages = [
    {"role": "user", "content": "Add following users: Niel and Tom are 31 years old, James is 25 years old"}
  ]

response = client.create(
    model="gpt-3.5-turbo",
    messages=messages,
    actions = [insert_users],
    stream=False,
    token_usage_tracker = TokenUsageTracker(500),
)

response, user_db