Send notification content through Firebase and add an Arq scheduler

[foucblg]

No description provided.

[codecov]

Codecov Report

Attention: Patch coverage is 53.26087% with 86 lines in your changes missing coverage. Please review.

Project coverage is 80.99%. Comparing base (66e7531) to head (a0ce052).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
app/types/scheduler.py	45.94%	40 Missing ⚠️
app/utils/communication/notifications.py	55.81%	19 Missing ⚠️
app/core/notification/endpoints_notification.py	56.25%	7 Missing ⚠️
app/app.py	14.28%	6 Missing ⚠️
app/modules/ph/endpoints_ph.py	16.66%	5 Missing ⚠️
app/dependencies.py	57.14%	3 Missing ⚠️
app/utils/tools.py	57.14%	3 Missing ⚠️
app/modules/cinema/endpoints_cinema.py	50.00%	1 Missing ⚠️
app/modules/loan/endpoints_loan.py	91.66%	1 Missing ⚠️
app/types/exceptions.py	66.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #613      +/-   ##
==========================================
- Coverage   81.31%   80.99%   -0.33%     
==========================================
  Files         129      130       +1     
  Lines       10013    10116     +103     
==========================================
+ Hits         8142     8193      +51     
- Misses       1871     1923      +52     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[armanddidierjean]

We should add a test that tries to defer a method using the scheduler

[armanddidierjean]

Is there a real interest in keeping both queue_job_time_defer and queue_job_defer_to?
I think we are likely to calculate defer_seconds using a timedelta and datetime, and thus could easily use queue_job_defer_to instead.

It seems Arq will calculate a datetime in both cases, so there shouldn't be a performance gain from using a method or the other.

I would think only having one method will reduce code complexity and limit potentials mistakes