Dronery

Python based REST API to handle a drones delivery service using the micro framework Bottle. The principal cargo for the drones to carry on are medications. The comunication interface from/to the drones is not provided and it's outside of the scope for this project. This is just a demo to test the features (and lack of) of the selected framework for RESTful API development. It is a prototype not intended for production at the moment.

Development restrictions

• No more than seven days to create a viable solution

General goals

• JSON based input/output
• Standards compliant endpoints with a versionated API interface '/api/v#' (althought just v1 is being developed at the moment)
• Dependencies at the minimum
• Portability (provided Python 3 is installed and available)
• Bare bones API features to keep the server implementation general enough (it imposes more processing on the client side as a side effect). This means that there are mostly generic CRUD capable endpoints with minimum "special cases" provided as "special endpoints". All in favor of predictability over specialization.
• Image files are not support in binary format, just Base64 encodings with an associated MIME type for the client to know how to handle it
• Protect sensitive API endpoints with authentication/authorization (just the very basic as a proof of concept more than security hardening)
• Provide some testing capabilities (using built-in unittest module and Boddle to mock requests)
• Provide some sort of endpoints self discovery when exploring with a web browser

General features

• Drones management through '/api/v#/drones' endpoint
• Medications management through '/api/v#/medications' endpoint
• Images management through '/api/v#/images' endpoint (Base64 encoding for the binary content)
• Loads management through '/api/v#/loads' endpoint
• CORS enabled for browser based clients
• Cache control headers support (just a general 'no-store' policy at the moment)
• Access control over management operations (plain old admin:admin http auth at the moment)
• Periodic execution of an audit task to log information about battery capacity of the drones (at the moment it is just a another thread running at a fixed interval)
• Basic filtering via GET parameters for the endpoints listing collections

Knwon issues

• When the server crashes due to an unexpected error, then the thread running the audit task keeps running on its own. This can be easily addressed by moving the audit task execution to an OS dependent scheduler alternative (like cron, systemd, etc) or running the server and the audit task in separate execution contexts. This approach was not used here to keep it simple and portable.

Improvements (given more available time for development)

• Do a more fine grained error handling to provide better hints and desciptive messages to the API consumer
• Dockerize the solution to have a more stable environment and simplify deployment and/or distribution
• Increment the number of unit tests to get more coverage and more resilience to regression bugs
• Improve the documentation of the API using more standards-based third party solutions (Swagger, etc)
• Provide support for PATCH operations

Note before reading the next sections

1. Most of the instructions given below that refers to directory locations use '/' just for the sake of simplicity (the only exception is for the virtual environment activation because it is rather different depending on the shell). On Windows '' is the correct directory separator to use.
2. When refering to the Python interpreter it is used the command python, again just for simplicity. On Windows systems python.exe, py.exe, or even py may be the ones available. Take the commands given below as a reference and adjust them if required. This may apply also to pip command.

Requirements

• Install Python 3.9.7+ (it may work with any Python 3 version but it is only tested against v3.9.7)
• Clone repository into a local directory (or simply unzip the downloaded archive)
• Start a terminal in that directory: cd /path/to/project/
• Create a virtual environment (optional but recommended): python -m venv .venv
• Activate the newly created virtual environment:
  • Bash source ./.venv/bin/activate
  • Windows CMD .\.venv\Scripts\activate.bat
  • Powershell .\.venv\Scripts\Activate.ps1

Test

• Make sure the previously listed requirements are met
• Install dependencies: pip install -r ./tests/requirements.txt
• Run tests: python -m unittest -v

Start the server

• Make sure the previously listed requirements are met
• Install dependencies: pip install -r ./requirements.txt
• Start the server: python -m bottle main:root
• Point your browser to http://localhost:8080
• Browse the built-in documentation to get a hang of how the API works
• Consume the API with a proper client tool:
  • ThunderClient is the recommended one. It is a VSCode extension comparable with Postman
  • If using ThunderClient, then import the provided thunder-collection_Dronery.json file for a head start