Python Essentials

To write code in LangGraph, you need a good understanding of Python’s type hinting system. Because LangGraph passes a shared State around, we must precisely annotate what kinds of data are allowed inside our State.

1. Dictionaries vs. TypedDicts

Regular Dictionary 📖

A standard Python dictionary allows any keys and values without enforcing any rules.

movie = {"name": "Avengers Endgame", "year": 2019}

Pros: Flexible and fast.
Cons: No type safety. The code does not check if key names are spelled correctly or if values are of the correct type, leading to potential runtime crashes.

Typed Dictionary ⌨

A TypedDict allows us to define a dictionary with a fixed set of keys and specific types for each key. This is the primary way we define Graph States in LangGraph.

from typing import TypedDict

class Movie(TypedDict):
    name: str
    year: int

# Correct usage
movie = Movie(name="Avengers Endgame", year=2019)

# Type checker will flag this because 'year' must be an integer:
# movie = Movie(name="Iron Man", year="2008")

Pros: Explicit, self-documenting, and validated by static analysis tools (like Ruff or Pyright) to catch bugs early.

2. Union Type 🤝

A Union indicates that a variable can hold one of several specified types.

from typing import Union

# x can be either an int or a float
def square(x: Union[int, float]) -> float:
    return x * x

x = 5       # Valid
x = 1.234   # Valid
# x = "text"  # Invalid (type checker will flag this)

💡 Note: In Python 3.10+, you can write int | float instead of Union[int, float].

3. Optional Type 🤔

An Optional type specifies that a variable can either be of a specific type, or None. It is shorthand for Union[Type, None].

from typing import Optional

def nice_message(name: Optional[str] = None) -> None:
    if name is None:
        print("Hey random person!")
    else:
        print(f"Hi there, {name}!")

4. Any Type 🎲

Any is a special type that matches any and all types. It turns off type checking for that variable.

from typing import Any

def print_value(x: Any):
    print(x)

print_value(123)
print_value("Batman")

Use Any sparingly, as it bypasses the safety benefits of type checking.

5. Lambda Functions ⏳

A lambda function is a small, anonymous function defined in a single line using the lambda keyword. In LangGraph, we often use lambdas as simple passthrough nodes or quick routers.

# A lambda function that squares a number
square = lambda x: x * x
print(square(10)) # Output: 100

# Used with map()
nums = [1, 2, 3, 4]
squares = list(map(lambda x: x * x, nums))
# squares is [1, 4, 9, 16]

LangGraph Basics

LangGraph Agents

1. Dictionaries vs. TypedDicts

Regular Dictionary 📖

Typed Dictionary ⌨

2. Union Type 🤝

3. Optional Type 🤔

4. Any Type 🎲

5. Lambda Functions ⏳

​1. Dictionaries vs. TypedDicts

​Regular Dictionary 📖

​Typed Dictionary ⌨

​2. Union Type 🤝

​3. Optional Type 🤔

​4. Any Type 🎲

​5. Lambda Functions ⏳

1. Dictionaries vs. TypedDicts

Regular Dictionary 📖

Typed Dictionary ⌨

2. Union Type 🤝

3. Optional Type 🤔

4. Any Type 🎲

5. Lambda Functions ⏳