Phase 1 · Python / Environment Primer (Skippable)

If you already know Python and the basics of LLM APIs, you can skip this chapter and return when needed.

1. Python Crash Course / Python 速补

Variables & functions

name = "LangGraph"
count = 3
items = [1, 2, 3]

def greet(concept: str) -> str:
    return f"Hello, {concept}!"

print(greet("StateGraph"))

Classes & imports

from dataclasses import dataclass

@dataclass
class Node:
    name: str
    next: str | None = None

Packages & modules

• A directory with __init__.py is a package and can be imported.
• All examples live under the examples/ package and import the model via examples.common.llm.

2. Environment Setup / 环境搭建

python -m venv .venv
source .venv/bin/activate        # Windows: .venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env             # then edit .env to choose LLM_PROVIDER

3. LLM API Primer / LLM API 速览

• Token: the smallest unit a model processes (roughly a piece of a word). Usage is billed per token.
• Chat Completion: send the model a set of messages (system sets the role, user is input); it returns a reply.
• API key safety: never hard-code keys or commit them to Git. Put them in .env (git-ignored) and read at runtime.

chain = prompt | model           # the most common LangChain shape
result = chain.invoke({"concept": "StateGraph"})

Next: Write your first Chain →