Modules & Imports

Modules let you organize code into separate files. Understanding imports is essential for working with Python libraries and your own code.

Import Basics

# Import entire module
import json
data = json.loads('{"key": "value"}')

# Import specific items
from json import loads, dumps
data = loads('{"key": "value"}')

# Import with alias
import numpy as np
arr = np.array([1, 2, 3])

# Import everything (avoid this!)
from json import *  # Hard to track what's imported

Your Own Modules

Any .py file is a module. Create helpers.py:

# helpers.py
def format_prompt(text):
    return f"User: {text}"

API_VERSION = "v1"

Use it in another file:

# main.py
from helpers import format_prompt, API_VERSION

prompt = format_prompt("Hello!")
print(f"Using API {API_VERSION}")

Relative Imports

Within a package, use relative imports:

my_agent/
├── __init__.py
├── core.py
└── tools/
    ├── __init__.py
    └── search.py

# In tools/search.py
from ..core import AgentCore  # Go up one level
from . import utils           # Same directory

Common Import Patterns

# Standard library
import os
import json
from pathlib import Path
from typing import List, Dict, Optional

# Third-party
import requests
from dotenv import load_dotenv

# AI libraries
from openai import OpenAI
from langchain.agents import create_react_agent

# Your modules
from .config import settings
from .tools import SearchTool

Avoiding Circular Imports

# BAD: a.py imports b.py, b.py imports a.py
# This causes errors!

# SOLUTION 1: Import inside function
def process():
    from .other_module import helper  # Import when needed
    return helper()

# SOLUTION 2: Restructure code
# Move shared code to a third module

Import Best Practices

Do	Don't
Group imports (stdlib, third-party, local)	Mix import styles randomly
Use absolute imports when possible	Use `from x import *`
Import at file top (usually)	Import same module multiple times
Use aliases for long names	Create confusing aliases