Weather API

Table of contents:

• Installation guide
  • Windows
  • Linux
• API Endpoints
  • City Weather
  • User
  • Token authorization
  • Swagger
• Celery
  • Run scheduled tasks
  • Run basic tasks
  • Add new scheduled tasks
• Technical decisions

Installation guide

Windows:

1. Install Python 3.8 , PostgreSQL , Erlang for rabbitmq , RabbitMQ

   • Start RabbitMQ:
     • Navigate to C:\Program Files\RabbitMQ Server\rabbitmq_server-3.10.5\sbin
     • Start rabbitmq-server.bat as Administrator

2. If you haven't already clone the repository and cd into it

   git clone https://github.com/antonarnaudov/weather_api.git
   cd weather_api

3. Create local database named weather-api-db using pgAdmin or SQL Shell (psql)
   NOTE: Make sure you set proper credentials in DATABASE settings

4. Create venv by running:

   python -m venv venv

5. Activate venv in:

   • cmd by running:

     call venv/scripts/activate

   • Powershell by running:

     & venv/scripts/activate

6. Install all dependencies by running:

   pip install -r requirements.txt

7. Run migrations with:

   python manage.py migrate

8. Try starting the app with:

   python manage.py runserver

9. Try running the tests:

   python manage.py test

Linux:

1. Install PostgreSQL
   Python3.8:

   sudo apt-get install python3.8

   RabbitMQ:

   sudo apt-get install rabbitmq-server
   sudo systemctl enable rabbitmq-server
   sudo systemctl start rabbitmq-server

   Optional (to make sure it works): systemctl status rabbitmq-server

2. If you haven't already clone the repository and cd into it

   git clone https://github.com/antonarnaudov/weather_api.git
   cd weather_api

3. Create local database named weather-api-db using pgAdmin or SQL Shell (psql)
   NOTE: Make sure you set proper credentials in DATABASE settings

4. Create venv by running:

   python3 -m venv venv

5. Activate venv by running

   source venv/bin/activate

6. Install all dependencies with:

   pip install -r requirements.txt

7. Run migrations with:

   python3 manage.py migrate

8. Try starting the app with:

   python3 manage.py runserver

9. Try running the tests:

   python3 manage.py test

API Endpoints

City Weather

Provides access to weather data for all cities

Read Only: api/cityweather/

Detailed: api/cityweather/{id}/
Search by city: api/cityweather/?city=sofia

User

Provides endpoints for user profile management

CRUD: api/auth/user/

{
  "id": 1,
  "extended": {
    "id": 1,
    "phone": "0888666754"
  },
  "last_login": null,
  "is_superuser": true,
  "username": "admin",
  "first_name": "Admin",
  "last_name": "Adminov",
  "email": "admin@cityweather.com",
  "is_staff": true,
  "is_active": true,
  "date_joined": "2022-07-09T18:53:58.243512Z",
  "groups": [],
  "user_permissions": []
}

Search: api/auth/user/?search=admin
Search fields: first_name, last_name, email, username

Register Users on POST
Required fields on POST: username, password, password2, email

Token authorization

Takes a set of user credentials and returns an access and refresh JSON web
token pair to prove the authentication of those credentials.

Get token

POST: api/auth/login/
Expects:

{
  "username": "admin",
  "password": "1234"
}

Returns:

{
  "refresh": "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJ0b2tlbl90eXBlIjoicmVmcmVzaCIsImV4cCI6MTY1NzY1NTA2NywiaWF0IjoxNjU3NTY4NjY3LCJqdGkiOiIxOTg4NWE1OGIzNTk0NzJiYjI2NmNjY2FjZGZmN2JkNyIsInVzZXJfaWQiOjF9.7zKhghN_UgQQW_7QnjfMckUOfaE3txwYnoXvEVoZQ8E",
  "access": "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJ0b2tlbl90eXBlIjoiYWNjZXNzIiwiZXhwIjoxNjU3NTY4OTY3LCJpYXQiOjE2NTc1Njg2NjcsImp0aSI6ImVjZTQ5YjdiZTdkNzQ2Y2NhZjlhNWEwZTNlOWQ4MjhjIiwidXNlcl9pZCI6MX0.HqXlOh1RrOxocD2kmpUJy2cqH1lZagtwCl-iJSsm4ro"
}

Refresh token

POST: api/auth/refresh/
Expects:

{
  "refresh": "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJ0b2tlbl90eXBlIjoicmVmcmVzaCIsImV4cCI6MTY1NzY1NTA2NywiaWF0IjoxNjU3NTY4NjY3LCJqdGkiOiIxOTg4NWE1OGIzNTk0NzJiYjI2NmNjY2FjZGZmN2JkNyIsInVzZXJfaWQiOjF9.7zKhghN_UgQQW_7QnjfMckUOfaE3txwYnoXvEVoZQ8E"
}

Returns:

{
  "access": "eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJ0b2tlbl90eXBlIjoiYWNjZXNzIiwiZXhwIjoxNjU3NTY5MTIwLCJpYXQiOjE2NTc1Njg2NjcsImp0aSI6IjRlZThkZTk5ZDYzYzRiNjdiODAyYmU3ODhlZGY0ZWFjIiwidXNlcl9pZCI6MX0.6i0qPK0KLjRwFMkoO6R-s5IrWBImT5Rt86Xu6GntQc8"
}

IMPORTANT:

To use the access token in request, simply add it to request headers as: Authorization: Bearer ${access}

Token Lifespan:

• Access token lifespan is 5 minutes
• Refresh token lifespan is 1 day
• To retain an ongoing session you should refresh the access token on every request

Swagger

Auto-generated documentation for all accessible endpoints

GET: api/swagger/

Celery Tasks

Make sure you have started the RabbitMQ Server

NOTE: You can skip those steps if you've already done them

On Linux you should execute the following commands:

sudo systemctl enable rabbitmq-server
sudo systemctl start rabbitmq-server

On Windows you should:

• Navigate to C:\Program Files\RabbitMQ Server\rabbitmq_server-3.10.5\sbin
• Start rabbitmq-server.bat as Administrator

Run scheduled tasks:

Open any terminal and execute this line to start the celery workers:
For Windows:

celery -A weather_api worker -l info --pool=solo

For Linux:

celery -A weather_api worker -l info

In a separate terminal run this line to start the celery beat and all scheduled tasks:

celery -A weather_api beat -l INFO --scheduler django_celery_beat.schedulers:DatabaseScheduler

Run basic tasks:

Open any terminal and execute this line to start the celery workers:
For Windows:

celery -A weather_api worker -l info --pool=solo

For Linux:

celery -A weather_api worker -l info

NOTE: Go to weather_api > celery and change update_city_weather_data_per_day schedule to 60 seconds to see instant result.

Add new scheduled tasks:

1. Create @app.task decorated function in any file or app
   • Import app from weather_api.celery
2. Open weather_api > celery.py and register your worker inside app.conf.beat_schedule following this structure

"example_worker_runs_every_minute": {
    "task": "app.file.worker_function",
    "schedule": 60.0 # schedule expects seconds
}

Technical decisions

1. I used RabbitMQ instead of Redis for message broker because I aimed to make the project executable on both Windows and Unix based OS. Redis requires a WSL to run under windows and RabbitMQ can work directly under it, although it requires to manually start the server.

2. I used a custom filter instead of overriding the list method for the search by city, thus keeping the ViewSet super clean and simple

3. I used JsonFiled for storing the third party api data, which worked really nice with the Django ORM, allowing me to tap into the nested fields as if they were written in the model

4. I used created_at/updated_at DateTime fields to keep track of time ;-D, so I know when to refresh the data

5. I extended the user model (although it wasn't necessary) as a matter of good practice, because its nightmare to do it later in the development process, when the app expands

6. I moved all non-weather specific functionality in app named core, and all utility functions in core > utils

7. I added a custom mixin which allows me to dynamically change the serializer, based on the request type

8. The default pagination class is overridden, so it utilizes size parameter, for dynamic single page pagination

9. I added a custom exception handler, to minimize the chance for Server Error 500 response

10. All Open Weather Map API credentials are set as settings

11. Added a blank page on base rout, so it does not throw errors

12. I used PR-s for these projects and did not delete the branches, so you can track the full development history

13. I used WritableNestedSerializers to create Extended objects from the Users serializer

14. I provided console logging on the celery task which logs the time of execution and the next time of execution

15. I added a few model and api tests, for the user, token authentication and city weather using django tests and django-factory_boy