Restaurant Project

This is a simple restaurant manager project.

Development

• Install docker and docker-compose.
• Run cp .env.sample .env and fill .env file.
• Run docker-compose -f docker-compose.dev.yml up --build. It may take some time (because seeding initial data).
• The service is run under 0.0.0.0:8000 and update after each file changes.

Deployment

• Install docker and docker-compose.
• Run cp .env.sample .env and fill .env file.
• Run docker-compose -f docker-compose.prod.yml up --build. It may take some time (because seeding initial data).
• The service is run under yourserverip or domain.

Documents

Rest API Documents

• You can see api documentations under /swagger or /redoc urls.
• Also, the pdf file rest_apis_doc.pdf is in the root of project.

Database Diagram

More Details

Database Models

As you see in the above picture, We have three primary models for our database: User, Restaurant and Review. Also, I developed two other models for better design: Category and BusinessHour. The main idea of the model's fields came from seed data, but I added some additional fields. For example, we needed the number of reviews each user has to update the average_stars value quickly.

SQL vs. NoSQL

I chose Postgres as a relational database because there is a connection between different entities we have. Also, the number of data isn't too much for a relational database. But maybe after scaling up the service, it's better to think about NoSQL databases.

APIs

We have twelve APIs in this project (You can see the details in swagger)

Two of them are for login and logout users:

• I assumed the user list was fixed and used a TokenAuthentication method.
• The username for all users sat as their user_id, and the password is initial_password.

Three of them need the user to be logged in:

• Get user reviews.
• Create a new review.
• Delete a review.

Other APIs are:

• Get the restaurants list.
• Search restaurants by category (adding query param ?category=food).
• Get all reviews of one specific restaurant.
• Get details of one specific restaurant.
• Get all reviews.
• Search reviews by keyword (adding query param ?keyword=good).
• Get details of one specific review.

The API flow is:

• API <-> View <-> Serializer <-> Model <-> Database

Test

• We have 90% test coverage at the project.

Future Works

• Add Statsd or Prometheus for monitoring APIs.
• Connect project logger to Sentry for seeing logs on the sentry dashboard.
• Switch to any CI/CD to have better deploying process (for example, Gitlab CI/CD).
• Deploy on a cloud platform and switch to Kubernetes for better scaling and maintaining features.
• Think of NoSQL for scaling up the database.