README

Dependencies

Ruby version: 2.7.0.
Run bundle install to install application dependencies.
Run rails s to start the application on http://localhost:3000.
Database included as a sqlite file.
Run rake db:setup to create the databases and load seeds.
Run test suite with bundle exec rspec

Domain Model

Notes

Availabilities

The availabilities table has a reference to both the employees, and but the days and hours in which an Employee is available are both stored as integers. Note: The day field represents the index of that day of the week.

id	monitoring_shift_id	employee_id	day	hour
1	1	1	0	23

This means that for the first monitoring shift, the Employee with ID 1 can work on Monday (index 0) at 23:00 hours.

Monitoring Schemas

The monitoring_schemas table has a JSON field named structure. This field stores the schedule delivered by a client for their service to be monitored. This schedule is a hash, in which the keys represents a day of the week by its index, and the values are arrays that represent a range of hours of the day.

For example:

  {
    0 => [19, 20, 21, 22, 23],
    1 => [19, 20, 21, 22, 23],
    2 => [19, 20, 21, 22, 23],
    3 => [19, 20, 21, 22, 23],
    4 => [19, 20, 21, 22, 23],
    5 => [6, 7, 8, 9, 10],
    6 => [6, 7, 8, 9, 10]
  }

This structure states that from Monday to Friday a service needs to be monitored between 19 hours and midnight, and the weekend the service needs to be monitored from 6 AM to 11 AM.

Monitoring Shifts

The monitoring_shits has references to both services and weeks tables. The structure field of the monitoring_shits table stores the confirmed monitoring schedule of a service in a particular week. Monitoring schedule is represented on a structure like the following:

  {
    0 => {
      19 =>1,
      20 =>1,
      21 =>1,
      22 =>2,
      23 =>2
    }
    ...
  }

Taking the structure of the monitoring_schema shown in the previous section, the structure of this monitoring_shit states that the Monday of that week the shift from 19 hours to 22 hours will be in charge of the Employee with ID 1, and the second shift, from 22 hours to midnight, will be in charge of the Employee with ID 2.

This field also stores the workloads assinged for each Employee for that particular Monitoring Shift. After the schedule has been calulated, a workload hash is included to the structure. Workloads look like this:

  {
    1 =>15,
    2 =>30,
    3 =>5,
    nil =>50
  }

This means that, for a particular week, the Employee with ID will work 15 time blocks, Employee with ID 2 will work 30 time blocks, and the Employee with ID 3 will work 5 time blocks. There will also be 50 time blocks unassigned.

Assumptions

Any Employee can be assigned to the Monitoring Shift of any Service.
A Service can only have one Monitoring Schema.
It is assumed that all Weeks start on a Monday.
It is assumed that structure hash field of a Monitoring Schema, is always sorted. Meaning that the first key of the hash is always zero (Monday), and the last key is always 6 (Sunday).

Additional Comments

Maintainers for Employees, Services, Weeks, and Monitoring Schemas were not considered in the development of the MVP.
The project seeds populates the database with:
- 2 Services called Recorrido, and Brutal.
- 2 Monitoring Schemas, one for each Service shown above.
- 5 Weeks. The current week + 4 weeks in the future.
- 1 Monitoring Shift, for the first week of the Recorrido service monitoring.
- 3 Employees called Mati, Benjamín, and Lennart.
- For the only monitoring shift created, the employees' availabilities are created randomly.
The algorithm to fill a Monitoring Shift works like this:
1. Shifts are build day by day, starting on Monday and ending on Sunday.
2. In each day, it starts by assigning time blocks to the employee that currently has less time blocks assigned (To give the chance that all employees have a balanced work load).
3. Whenever an employee isn't available on a given time block, it selects the employee who has the most consecutive time blocks starting that point (To reduce shift changing).

, this is explained in the ShiftCalculatorService as well.

matipacheco / maas-backend