This is a Flask based HATEOAS driven RESTful API to get and search data related to company and employees according to Paranuara Challenge - build a REST API to provide the desired information about people living in the Paranuara planet. Simply use a API tool like Postman. It uses MongoDB for data storage. pytest framework is used to write test cases to enhance data import and API behavior and stability. Travis CI (https://travis-ci.com/github/nrvikas/Paranuara-Challenge) is used to facilitate CI/CD for the ongoing development of the API.

The AWS instance also has the test screenshots about data import and sanitization. It also has screen shots of API usage and error scenarios.

• API
• Installation
• Start API
• Test Screenshots
• Config
  • System Config
  • Tests Config
• Data Import
• Schema
• Database
• Locking
• Code base
• CI-CD and Testing

The API supports HATEOAS The endpoints currently supported are:

Given an environment with Python and MongoDB (>= 3.6) installed, setting this application up is very easy and straight forward.

In the folder of the project:

pip install -r requirements.txt

From the project directory, execute: python run.py

After starting up the API with the above command, open your browser and open the address location: http://localhost:5000/

Alternatively please see the static folder for the screenshots of:

• Data Import
• Schema Validations
• API Test
• API Error Tests

I will make an AWS instance available for this shortly.

Below sections include the config options that can be defined in settings.py.

• DB_NAME
  • Database name for the API. Defaults to paranuara

• DB_NAME
  • Test Database name for the API. Defaults to paranuara_test

Data Import is supported my manage commands. We have 4 collections in this API instance.

• companies
• people
• fruits
• vegetables

In order to have this application evolve, fruits and vegetables were made collections and not just in-memory lists or data files. This way, just like people and companies even fruits and `vegetables`` can be added to the system by data imports and minimal or no code changes.

The data imports of people collections automatically creates ['fruits'] and ['vegetables'] lists in the people document deriving it from favouriteFood. This way we don't need to do DB look-ups while fetching data for API responses. Because of this, the order of importing data is important. fruits and vegetables must be imported before we import people

Following validations and sanitization procedures are covered during data import and there are test cases to verify these:

• Valid and allowed collection name only
• File extension to be only .json
• Must be a valid file path
• Various schema validations (mentioned in schema section below
• Collection dependency validation -to ensure fruits, vegetables and companies are imported before importing people

Format of the command to import data is 'collection_name' followed by 'absolute_file_path'

• Import fruits and vegetables
  • python manage.py fruits <absolute_path_to_the_folder>/fruits.json
  • python manage.py vegetables <absolute_path_to_the_folder>/vegetables.json
• Import people
  • python manage.py people <absolute_path_to_the_folder>/people.json
• Import companies
  • python manage.py companies <absolute_path_to_the_folder>/companies.json

Bulk import of data at once to all collections are also supported. Simple place all your .json files in one location and use the following command: python manage.py initialze_data_from_directory <abolute_directory_path_to_all_four_json>

As there is an index field in the given JSON files, we are using that as the ID of the documents exposed to public. The _id field created by MongoDB is not projected outside. During the data import, schema validations that take place are:

• Checking for 'type' validation to make sure number, boolean and list fields are as as per schema specification
• index field is a required field
• index field is verified for uniqueness during validation
• Fields that are not a part of the resource schema are discarded.
• During the import of 'people' resource, favouriteFood is split into fruits and vegetablesarray in the people document.

To enable efficient access of the people and companies documents, following MongoDB indexes are created. people collection indexes:

• index
• company_id
• eyeColor
• has_died

companies collection indexes:

• index field

Since we are now exposing only a READ-ONLY API without any POST facility, locking is not yet implemented.

The architecture of the code base enables the API to evolve. Modular approach taken enables us to develop other fully operational API endpoints for other collections with minimal effort. Simply create a Flask blueprint for the next API you want to develop and implement the specific API business logic. For this reason, there has been a specific effort to abstract Database operations into a db_layer module so the API business logic can be agnostic about the inner DB implementations.

• app : ___init initializes the Flask App, initializes DB, registers blueprints, add config variables and initializes core app modules
• db_layer: contains interfaces exposed for other Blueprints and future APIs to use to carry out any database operations.
• static: Static Image files/assets used in HTML templates. Ex. Main page of the API
• templates: HTML templates used by the application
• tests: pytest test cases for validation of data import, schema and API behavior.

To ensure proper standards and best practices for coding the API, we use flake8 as the styling guide.

Travis CI is integrated with this repository to run pytest test cases and uses flake8 to ensure proper code style and standards across Python files being pushed into the repository. Tests written cover:

• Data import validation.
• Schema and data dependency Validation
• API verification of all endpoints

It can be found here: https://travis-ci.com/github/nrvikas/Paranuara-Challenge