This repository contains the code for an API that serves a restaurant daily ranking service. As well as code, it also includes some testing data and configuration information. These kind of files do not normally belong in a repository, but are distributed anyway for convenience.
This solution utilizes Django Rest Framework. It also uses Huey for task scheduling.
There is only one app, named "restaurant". It has a bunch of views, serializers, models, and some helper functions, which have a TODO to be moved to their own python files.
The code and this README have a few todos. You can find all of them by looking up "TODO" using your favorite grep tool (like visual studio code).
.vscode/launch.json - Vscode config for debugging django
convious/__init__.py
convious/asgi.py - Default asgi file
convious/settings.py - Django settings file
convious/urls.py - Urls file
convious/wsgi.py - Default asgi file
convious/restaurant/__init__.py
convious/restaurant/admin.py - Default admin file
convious/restaurant/apps.py - Default apps file
convious/restaurant/models.py - Restaurant models
convious/restaurant/serializers.py - Restaurant serializers
convious/restaurant/tasks.py - Restaurant Huey periodic jobs
convious/restaurant/tests.py - Restaurant tests... empty for now
convious/restaurant/views.py - Restaurant views
db.sqlite3 - Debug environment database
manage.py - Default manage file
readme.MD - You're standing here <-
requirements.pdf - This project's business requirements
requirements.txt - Python requirements file
settings.HUEY - Huey settings file
1. Everyone can add/remove/update restaurants
This is implemented using the /restaurants endpoint and the Restaurant model.
2. Every user gets X (hardcoded, but "configurable") votes per
day. Each vote has a "weight". First user vote on the same
restaurant counts as 1, second as 0.5, 3rd and all subsequent
votes as 0.25.
2.1. If a voting result is the same on multiple restaurants, the
winner is the one who got more distinct users to vote on it.
Configurable vote count: implemented using an environment variable and a default setting.
Vote count: Implemented with a dedicated model: VoteCount.
Vote weight: Calculated by the API on the fly using VoteCount.
Winner calculation: Implemented with the model DailyWinner, it is calculated every day by a periodic task in convious\restaurant\tasks.py.
The API endpoint is /votes
3. Every day vote amounts are reset. Unused previous day votes are
lost.
VoteCount is reset every day by a periodic task in convious\restaurant\tasks.py.
4. Show the history of selected restaurants per time period. For
example, the front-end should be able to query which restaurant
won the vote on a specific day.
This is implemented using the /dailywinners endpoint and the Restaurant model. Currently it only supports listing all restaurants and quarying specific dates. TODO: Add query support for arbitrary time periods (e.g. 2022-03-01 to 2022-03-30).
5. Do not forget that frontend dev will need a way to show which
restaurants users can vote on and which restaurant is a winner
This is implemented using the /restaurants endpoint and the Restaurant model. Currently, the API required user authentication, which we may not want. TODO: Remove user authentication requirement from parts of the API where it is not wanted, TBD.
6. Readme on how to use the API, launch the project, etc.
You're standing here <-
# Optional, making sure that we have the venv package installed, for example using Ubuntu:
apt-get install -y python3-venv
# Installing a new venv in the project directory:
python -m venv venv
# “Activating” the virtual environment. This shell session will now be using the virtual environment instead of the global interpreter:
. venv/bin/activate
# Installing all requirements in venv:
pip install -r requirements.txtpython manage.py migratepython manage.py runserverIf your vscode has a Python extension, open any file and debug. The repository includes the needed configuration.