To run the API, create a virtualenv with Python 3.6+.
Inside the virtualenv, run pip install -r requirements.txt from the root of the weightsapi directory.
IMPORTANT: Inside ./weightsapi/settings.py, add some random value to the SECRET_KEY const.
Then run ./manage.py migrate to create the SQLite database with necessary tables.
Finally, run ./manage.py runserver 0:8000 to launch the Python webserver.
The site will be accessible via http://localhost:8000/animals/
To run tests: ./manage.py test
A time limit of 3 hours was imposed on this task.
Goal: Build an API endpoint that records animal weights at a given point in time. Build a second endpoint that lets users query estimated weights for a given date (e.g. user has records for Jan 1 and Jan 5, but wants the estimated weight for Jan 3).
Uses Linear Interpolation / Extrapolation to calculate the estimated weight of an anmial for a given datetime timestamp.
-
Used scipy for calculating the Linear Interpolation/Extrapolation due to time constraints and its well-tested implementation.
-
Data is stored in SQLite.
-
The estimate_weight endpoint stores the calculated weight in the DB for future, repeated lookups. This acts a sort of cache for the estimated values.
- "Cache" busting was not implemented to remove or invalidate the stored estimated weights if new data BEFORE the estimated date was added. E.g if we have a stored estimated weight for Jan 6 and a recorded weight for Jan 5 is added, the Jan 6 estimated weight will still exist even though it could be affected by this new value.
-
Project set up is simplified due to 3-hour constraint. Typical installation would use separate settings files for different environments.