​
Overview

State management in Davia provides a clean way to work with persistent data using Python’s type annotations. By leveraging the Annotated type and State class, you can easily access and modify state between different tasks.

​
Working with State

​
Initializing State

Initialize state during app creation:

from davia import Davia, State
import pandas as pd

products = pd.DataFrame({
    "name": ["Laptop", "Phone", "Headphones"],
    "price": [999.99, 699.99, 199.99],
    "quantity": [1, 1, 2],
})

app = Davia(
    title="Product Manager",
    description="Manage your shopping list with ease",
    state={"products": products},
)

​
Accessing State

Davia provides two ways to access state:

​
Method 1: Using Annotated Type

from typing import Annotated

@app.task
def add_product(
    products: Annotated[pd.DataFrame, State("products")],
    name: str,
    price: float,
    quantity: int = 1,
):
    new_product = pd.DataFrame({"name": [name], "price": [price], "quantity": [quantity]})
    products.loc[len(products)] = new_product.iloc[0]

​
Method 2: Accessing Complete State

@app.task
def get_products(state: State) -> dict:
    return state["products"].to_dict()

​
Modifying State

State modifications are applied directly to the parameter:

@app.task
def update_price(
    products: Annotated[pd.DataFrame, State("products")],
    product_name: str,
    new_price: float,
):
    products.loc[products["name"] == product_name, "price"] = new_price

​
Supported State Types

Davia supports:

  • Pandas DataFrames
  • Dictionaries
  • Lists
  • Custom objects
  • Primitive types (within a container)