A CRUD rest api app for books fetched from Openlibrary

Database models were created according to the logic indicated in the below diagram

Database diagram

(Folder models)

• book
• author
• work
• book_author
• book_work
• user

As indicated from openlibrary data,

• A book may have multiple authors and belong to multiple works
• A work may contain multiple books and
• An author may have written multiple books

Implementation with flask-sqlalchemy

User authentication implemented with flask-jwt-extended

(Currently protected endpoint: /books (and /v2/books, see versioning below) → GET. Uncomment decorator @jwt_required() to protect others)

Implemented with marshmallow

(Folder schemas)

• olib_book
• book
• author
• work
• book_author
• book_work
• user

(Folder resources)

• olib_book
• book
• author
• work
• user

In openlib there are books without registered authors, which suggests that authors should be handled by separate endpoints, decoupled from book endpoints implementation.

So the endpoints are the below:

• /olib-books → GET
  • Fetch book codes defined in book_codes.py from open-library, extracts info, clean tables and populate database
  • Accepts query async=<bool> based on which the get requests to open library are performed sequentially (async=False) or concurrently (async=True) using requests module in combination with asyncio. Second option is significantly faster.
• /books/<book_id> → GET, PUT, DELETE a book with a specific book_id
• /books → POST
  • Create a book with {'code': <book_code_str>, 'title': <book_title_str>}
• /books → GET
  • Get list of all books in database. May filter results by adding query contains=<string> (Filters by <string> in title, case insensitive)
• /books/<book_id>/authors/<author_id> → POST
  • Link a book to an author. Add a row to books_authors table
• /books/<book_id>/works/<work_id> → POST
  • Link a book to a work. Add a row to books_works table

• /authors/<author_id> → GET, PUT, DELETE an author with a specific author_id
• /authors → POST
  • Create an author with {'code': <author_code_str>}
• /authors → GET
  • Get list of all authors in database

• /works/<work_id> → GET, PUT, DELETE a work with a specific work_id
• /works → POST
  • Create a work with {'code': <work_code_str>}
• /works → GET
  • Get list of all works in database

• /users/register → POST register a user with {'username': <username>, 'password': <password>}
• /users/login → POST login a user with {'username': <username>, 'password': <password>}
• /users/logout → POST logout a user
• /users/<user_id> → GET, DELETE a user by user_id

• /rules → GET returns info about all available endpoints/methods and their versions

Versioning with Blueprints is implemented in file versioning.py

• Endpoints with prefix /v{i} are exposed for all resources/methods where {i} corresponds to the implemented resources/methods versions
• Endpoints without /v{i} prefix correspond to the latest implemented versions
• Implemented versions are:
  • v1, v2 for /books → GET (so we have /v1/books → GET | unprotected, and /v2/books(=/books) → GET | protected)
  • v1 for all other resources/methods
• /rules has no versions

Make .env file as in .env.example

WORK_ENV=prod/dev/test. If empty, defaults to dev. See config.py/run.py for details.

>>> python run.py

With Docker (Desktop) installed, on the folder where Dockerfile is:

• >>> docker build -t books-api-image .
• >>> docker run --rm -it --name books-api-container -p 5000:5000 -w /app -v ${PWD}:/app books-api-image (Powershell) OR
• >>> docker run --rm -it --name books-api-container -p 5000:5000 -w /app -v "%cd%":/app books-api-image (Terminal)

In root folder:

• >>> pytest -s to run all tests
• >>> pytest --strict-markers -m <marker-name> -s to run tests by marker, where marker-name s can be found in pytest.ini file

• Implement versioning with Blueprints ✓
• Documentation with flask-smorest/swagger
• Use flask-migrate for migrations
• Use gunicorn as a server
• Unit testing ✓