Shamseya Task

What's implemented

The Required:

api/core/reviews/ endpoint that returns all the reviews by default

in this shape:

{
  "2020-12-31": { // submitted_at field
      "count": 16,  // reviews count field
      "answer_count": 64, // answers count
      "answer_ids": [  // answers ids
          3149,
          3150,
          3151,
          3152,
          10153,
          10154,
          10155,
          10156,
          ...
      ]
  }
}

I collect by date, return their count, and merge all the answers for all the reviews in the specific date, and put their count
Since the shape of the data is not explicitly specified, I did what I saw suitable, and easier for me to implement.
I also return IDs only, because no need for more information explicitly required

filter reviews by from_date & to_date query params
- both are optional, this is tested, too.
I did not write units to test the shape of the data in different cases, did that manually

The Bonus:

The endpoint is protected. Allowed for logged-in (authenticated) users only.
- This is satisfied with this permission IsAuthenticated
There're three users in the attached dataset
- you can also create them using create_users command
- refer to their passwords here
Then Added another restriction
- A User must be an AdminUser (staff) or SuperUser
- This implemented using these permissions IsAdminUser | IsSuperUser
The previous points are satisfied with this python statement: permission_classes = (IsAuthenticated, IsAdminUser | IsSuperUser) in the Review APIView
- They're tested to make sure they work as required: test_views
I usually dockerize PostgreSQL, and that's what I did, since it's a bonus part I don't have the desire to do more configurations.

Extra:

Home view (/) to try different results using quick links to the api endpoint
- This was easier for me during the manual test, and might be useful for you
I also use Debug Tool Bar to monitor the performance
Plus some other helper commands to help me generate data
And lastly, some other boilerplate code from the Template project I'm using. Not harmful though.

Pre-requisites (On Ubuntu)

Python 3.6
docker & docker-compose

Run

clone the project
cd into project's directory
create a python virtualenv and activate it
$ python -m pip install poetry I use it to manage he dependencies of my projects
$ poetry install install project dependencies inside the venv
$ cp .env.example .env
$ docker-compose up to run postgres container
$ ./manage.py test to run tests
$ ./manage.py migrate to apply migrations
$ ./manage.py loaddata fixtures/4krev.db to load 4K Reviews Dataset
- There're other datasets you can use, too.
$ ./manage.py runserver 8005 to run server
Authenticate a user using any of the available users trough the login route
Open http://localhost:8005/ you'll find quick links to try different results
Or navigate to http://localhost:8005/api/core/reviews/ directly

Available users

usernames: super_user, super_only, staff_user, active_user
password (same for all): Awesome1

Authentication routes

/api/auth/login/
/api/auth/logout/

Thoughts on performance

I think I've reached the best performance for now,

Results from Debug Tool Bar (16K Reviews Dataset):

The Query is very fast, but the browser rendering (Browsable API) for the response takes time, we can solve that using Pagination.
Also we could use Postgres Materialized Views to make just one select statement (for larger datasets, 20K+)
- A helper for that would be this package: https://github.com/mypebble/django-pgviews
- we'll also need pagination for better user experience

More commands

$ ./manage.py populate_db add sample db records (takes time)
$ ./manage.py clean_db remove all db entries

shahwan42 / shamseya_task