Dependency Management

One of the challenges of Python development is managing "Dependency Conflicts." If Project A requires Requests v2.0 and Project B requires Requests v1.0, installing them into your global Python system will break one of the projects.

Virtual Environments solve this by creating isolated, standalone directory trees that contain a specific Python installation and a distinct set of libraries.

1. Why Isolation is Mandatory

Without isolation, you risk:

1. System Instability: Overwriting critical libraries used by your Operating System's Python (on Mac/Linux).
2. "It Works on My Machine": Difficulty in sharing code since your global environment is unique.
3. Circular Dependencies: Trying to satisfy two conflicting requirements at once.

2. Using venv (The Built-in Standard)

As of Python 3.3, isolation is a built-in feature via the venv module.

# 1. Create the environment
python3 -m venv .venv

# 2. Activate it (Platform specific)
source .venv/bin/activate  # Mac/Linux
.venv\Scripts\activate     # Windows

# 3. Install dependencies
pip install requests

Once activated, your $PATH is modified so that the python and pip commands point to the bin/scripts folder inside the .venv directory.

3. The Dependency Manifest: requirements.txt

To ensure your project is reproducible, you should always track your dependencies.

# Save current environment state
pip freeze > requirements.txt

# Reconstruct environment on a new machine
pip install -r requirements.txt

4. Modern Tools: Poetry and Conda

While venv is the base, the industry often uses more advanced tools:

• Poetry: Handles dependency resolution, lockfiles (poetry.lock), and packaging in a single tool. It is the modern standard for new Python projects.
• Conda: A manager that handles not just Python, but also C-libraries and non-Python dependencies. It is essential for Data Science and Machine Learning.

Interview Pro-Tips: How does pip work?

If asked about dependency resolution:

• pip uses a backtracking resolver to find a set of package versions that satisfy all constraints.
• In professional CI/CD pipelines, always use Lockfiles (created by Poetry or pip-compile) to ensure you are installing the exact same bytes in production as you did in development.

Technical Summary

1. DotVenv: A common practice is to name your environment .venv and add it to your .gitignore.
2. Activation: Only a temporary terminal state; does not modify the filesystem outside the virtual environment folder.
3. Reproducibility: The goal is to make your environment disposable and reconstructible from a single command.