Requirements Files

The standard way to manage dependencies is using requirements.txt:

1
requests==2.28.0
2
numpy>=1.20.0,<2.0.0
3
pandas==1.5.3
4
matplotlib

Generating Requirements

Create a requirements file from installed packages:

1
pip freeze > requirements.txt

Installing from Requirements

1
pip install -r requirements.txt

Exact Version Pinning

1
requests==2.28.0
2
numpy==1.24.0

Pros: Reproducible builds Cons: May miss security updates

Minimum Version

1
requests>=2.25.0
2
numpy>=1.20.0

Pros: Gets updates automatically Cons: May break with incompatible updates

Version Range

1
requests>=2.25.0,<3.0.0
2
numpy>=1.20.0,<2.0.0

Pros: Balance between stability and updates Cons: Requires careful version management

requirements.txt (Production)

1
requests==2.28.0
2
numpy==1.24.0
3
pandas==1.5.3

requirements-dev.txt (Development)

1
-r requirements.txt
2
pytest==7.2.0
3
black==22.10.0
4
flake8==5.0.0

Install development dependencies:

1
pip install -r requirements-dev.txt

Project Structure

1
my_project/
2
    requirements.txt
3
    setup.py
4
    README.md
5
    src/
6
        my_package/
7
            __init__.py
8
            module1.py
9
            module2.py
10
    tests/
11
        test_module1.py
12
        test_module2.py

Check for Outdated Packages

1
pip list --outdated

Update All Packages

1
pip list --outdated --format=freeze | grep -v '^\-e' | cut -d = -f 1 | xargs -n1 pip install -U

Update Specific Package

1
pip install --upgrade package_name

Check for Vulnerabilities

Use tools like safety:

1
pip install safety
2
safety check

Keep Packages Updated

Regularly update packages to get security patches:

1
pip install --upgrade package_name