Docker running project: starting point

🇫🇷 Version française : branche fr.

This repository is your starting point for the hands-on project of the Docker module. By the end of the module, it will run three containers connected by Docker Compose:

• db: a MariaDB database, which holds 1,000 customer transactions;
• train: a one-off job that trains a fraud detection model on these transactions;
• api: a FastAPI web API that gives access to the transactions and to the model's predictions.

All the code (Python, SQL) is provided: you only write the Docker part. Follow the course step by step: it tells you when and how to complete each file.

Repository contents

DS-Docker-Course-Architecture/
├── step2-first-dockerfile/       the step 2 warm-up exercise (not part of the project)
│   ├── hello.py
│   ├── requirements.txt
│   └── Dockerfile                ✏️ fill-in-the-blanks template (step 2)
│
├── api/                          api service: the FastAPI application
│   ├── main.py                   the application (provided)
│   ├── requirements.txt          its dependencies (✏️ one line to add in step 8)
│   ├── Dockerfile                ✏️ to write (step 4)
│   └── .dockerignore             ✏️ to complete (step 4)
├── db/                           db service: the MariaDB database
│   ├── Dockerfile                ✏️ to write (step 5)
│   ├── init/
│   │   └── 01_transactions.sql   creates the table and loads the CSV into it (provided)
│   └── data/
│       └── transactions.csv      1,000 fictitious customer transactions (provided)
├── train/                        train service: model training
│   ├── train.py                  the training script (provided)
│   ├── requirements.txt          its dependencies (provided)
│   ├── Dockerfile                ✏️ to write (step 8)
│   └── .dockerignore             ✏️ to complete (step 8)
└── compose.yaml                  ✏️ to write (steps 6 and 8)

Each service has its own folder, named after it, which contains its Dockerfile. The step2-first-dockerfile/ folder is not part of the project: it is the step 2 exercise.

What you will complete

The files marked ✏️ only contain comments for now, which describe what they must contain.

File	Step	What to do
step2-first-dockerfile/Dockerfile	2	Replace each _____ with the right instruction
api/Dockerfile	4	Write, under each comment, the corresponding instruction
api/.dockerignore	4	List the files that must not be copied into the image
db/Dockerfile	5	Start from the official MariaDB image and add the scripts and data to it
compose.yaml	6	Describe the db and api services, their volume and their startup order
train/Dockerfile, train/.dockerignore	8	Write them, based on the specifications in the course
compose.yaml, api/requirements.txt	8	Add the train service and connect the model to the API

You do not need to modify the other files.

The provided programs

Program	How to start it	Configuration (environment variables)
api/main.py	uvicorn main:app --host 0.0.0.0 --port 8000	DB_HOST, DB_PORT, DB_USER, DB_PASSWORD, DB_NAME, MODEL_PATH
train/train.py	python train.py	The same DB_* variables, and MODEL_PATH

Each variable is described at the top of each program. Once the API is running, its interactive documentation is available at /docs.

Resources

• The Docker cheat sheet: a PDF of six sheets with the commands, file templates and solutions to the most common errors.
• The official Docker documentation: https://docs.docker.com
• The solution is published in the DS-Docker-Course-Solution repository: it is the final project, as it should be at the end of step 8. Use it to compare with your work once a step is done, not to avoid searching.