编写生产环境 Dockerfile

A Docker image packs your code, Python, and every dependency into one unit that runs the same on any machine. No more works on my laptop. 🐳

A Dockerfile is a recipe. Each line is a step Docker runs to build your image, from base layer to final run command.

Start from an official python image. The slim variant keeps the image small while still giving you a working Python runtime.

Use WORKDIR to choose where your code lives inside the image. Every later command runs relative to this folder.

Copy requirements.txt before your code. This lets Docker cache the install layer and skip it when only your code changes.

Run pip install with no cache to keep the layer lean. This bakes Flask and Gunicorn right into the image.

Now COPY the rest of your project in. Because it comes after the install, edits to your code do not bust the dependency cache.

Use EXPOSE to declare which port your app listens on. It is documentation plus a hint for tooling that wires things up.

The CMD line is what runs when the container starts. Point it at Gunicorn so the container boots straight into your server.

Add a .dockerignore file so caches, venvs, and git data never enter the image. Smaller builds are faster and safer.

Build with docker build, then run the image and map a port. Your packaged Flask app is now portable to any host.

Your Dockerfile builds from a slim base, installs deps, copies code, and launches Gunicorn. One image runs anywhere. Next: put Nginx in front. 🔧