This is an API for a simple book store. It allows customers to rent out books, return them and charges them appropriately.

More info on the API can be found in the docs

• Clone this repo. You might want to do this in a virtual environment (See more on virtualenv)

git clone https://github.com/dennynganga/koko_book_store.git

• Once cloned, and virtual environment activated (if using one):

pip install -r requirements.txt

• Run server, and open http://127.0.0.1:8000 on your browser

python manage.py runserver

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Please make sure to update tests as appropriate.

• First step involves running the tests on CCI. If all pass and we're on the master branch, we proceed to the publish job.
• In the publish job, a docker image is built with a unique tag (hash of the git commit that triggered the CI build).
• From Docker Hub, the application is then deployed to the VM. The latest image is pulled from Docker Hub, current one is stopped and is removed from the VM, then runs a container based on the latest image.

• Data Layer

For this test, a fairly simple database was used i.e SQLite. In order to be able to serve more requests efficiently in production, I'd switch to a more robust engine (preference to PostgreSQL) PostgreSQL supports query logging and this can be used to tell which queries are being run frequently or even which ones take long to execute. From analysing this info, we can decide which responses to cache (depending on business requirements) so as to decrease load on the database server. On the same note, I'd also bring in Elastic Search to provide a more scalable search. From our service, items that can be indexed include books, customers and rentals, so we can find them quicker. Also, as our database grows, we'd introduce a technique known as sharding. This is where the database is partitioned into smaller, faster parts.

• Hosting

This involves using a load balancer that routes requests to two or more application servers. This way, increased load is well handled.

• Session Storage

Moving session storage to a faster, in-memory caching tool like redis or memcached, to avoid database hits while reading and writing session data.

• Running long computations offline

Computations that take long to run should not be performed in the normal request-response cycle, so as to reduce page load times. In our case, an example would be sending a reminder email to a customer, or even generating a large CSV file that shows all active rentals. Such long-running tasks can be delegated to multiple workers. Celery (in combination with redis or rabbitmq) would come in handy in this.

• Caching Data on clients

It might also be useful to cache data on clients e.g smart phone app so as to reduce the number of requests sent to the servers. In turn, the servers will benefit from this technique.