Introduce module tests for GTFS-RT

[leonardehrenfried]

Summary

As discussed in the realtime working group meeting, this introduces module tests for the GTFS-RT realtime code. They are ports of the existing tests but have been converted to use @habrahamsson-skanetrafiken's test environment code. This results in faster test execution and a less confusing debugging experience.

To make this possible, I extracted the bulk of the test environment into a base class which is subclassed by both the Siri and GTFS-RT test to add a few specific methods.

Unit tests

Well... yes.

Documentation

Javadoc for each test case.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 68.54%. Comparing base (f535524) to head (410284c).

Additional details and impacted files

@@              Coverage Diff              @@
##             dev-2.x    #5871      +/-   ##
=============================================
+ Coverage      68.51%   68.54%   +0.02%     
- Complexity     16725    16732       +7     
=============================================
  Files           1918     1918              
  Lines          72734    72734              
  Branches        7456     7456              
=============================================
+ Hits           49834    49854      +20     
+ Misses         20329    20310      -19     
+ Partials        2571     2570       -1     

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

[habrahamsson-skanetrafiken]

@leonardehrenfried I think the idea looks good! But in general I would prefer composition in favor of inheritance. Specifically if we ever want to test gtfs and siri together it would get clunky with an abstract base class.

[leonardehrenfried]

@habrahamsson-skanetrafiken I've changed the inheritance to composition. It's not super-elegant either but less bad than some other options. My intention is for the split between Siri and GTFS test environment to become unnecessary once I've extracted a common timetable snapshot manager.

[habrahamsson-skanetrafiken]

If we would want to merge the SiriEnv, the GtfsEnv and the RealtimeTestData into one class, then we could put two separate snapshot sources in the merged environment but make sure only one was initialized. And then throw a runtime exception if you try to call the gtfs methods on an environment where the siri snapshot source was initialized. I think this would work?

But maybe it's nice with the separation that we get now. I don't really have a strong opinion. So you can leave it as it is if you want.

I like the structure that the moduletests add!

Otherwise just some minor comments.

[habrahamsson-skanetrafiken]

Maybe this should be called NOT_FULL_DATASET to make it clearer that we are setting the param to false in the call sites?

[t2gran]

What does "not full" mean? Partial ? I guess the underlying problem is that the input param i a boolean and not an enum.

[leonardehrenfried]

Thomas is correct. I wanted to turn this boolean into an enum for many years. Do you want me to do that in this PR or another one?