Use SIRI-ET and GTFS-RT TripUpdates at the same time

[leonardehrenfried]

Summary

Previously it was not possible to run SIRI-ET and GTFS-RT TripUpdates in the same instance. This PR changes it by cleaning up the relationships of the various real-time components.

A class diagram to show the new relationships is available in the issue #6348.

Issue

Closes #6348
Ref #4002

Unit tests

Some unit tests updated.

Documentation

The new class diagram is available in this PR in doc/dev/diagrams/TimetableSnapshotManager.svg.

Bumping the serialization version id

While no actual entities have changed, but the TimetableRepository has which is serialized into the graph. Therefore we need to bump it.

cc @fahrplaner

[codecov]

Codecov Report

Attention: Patch coverage is 69.04762% with 13 lines in your changes missing coverage. Please review.

Project coverage is 69.85%. Comparing base (12b5cb6) to head (b19e36e).
Report is 88 commits behind head on dev-2.x.

Files with missing lines	Patch %	Lines
...entripplanner/transit/configure/TransitModule.java	0.00%	5 Missing ⚠️
...ner/standalone/configure/ConstructApplication.java	0.00%	3 Missing ⚠️
...planner/updater/configure/UpdaterConfigurator.java	70.00%	3 Missing ⚠️
...oradapter/transit/mappers/TransitLayerUpdater.java	85.71%	0 Missing and 1 partial ⚠️
...planner/transit/service/DefaultTransitService.java	85.71%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@              Coverage Diff              @@
##             dev-2.x    #6363      +/-   ##
=============================================
+ Coverage      69.77%   69.85%   +0.07%     
- Complexity     18053    18115      +62     
=============================================
  Files           2060     2069       +9     
  Lines          77122    77213      +91     
  Branches        7856     7842      -14     
=============================================
+ Hits           53814    53938     +124     
+ Misses         20556    20522      -34     
- Partials        2752     2753       +1     

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

[vpaturet]

This looks good. Just minor comments.

[vpaturet]

You probably don't need to bump the serialization id since you change only a transient field.

[leonardehrenfried]

So removing fields doesn't count as breaking backwards-compatibility?

[abyrd]

So removing fields doesn't count as breaking backwards-compatibility?

I think the idea is that the transient field is not serialized, so the serialized graph is still compatible and will load without error.

[leonardehrenfried]

Alright then. The label is gone.

[abyrd]

Looking good so far. Just added a few questions that would reassure me I'm understanding what I'm seeing. In addition to those line comments, would it be correct to say that the refactoring described in the Javadoc on TimetableSnapshotManager has now been carried out within this PR? Should that Javadoc be updated?

[leonardehrenfried]

Looking good so far. Just added a few questions that would reassure me I'm understanding what I'm seeing. In addition to those line comments, would it be correct to say that the refactoring described in the Javadoc on TimetableSnapshotManager has now been carried out within this PR? Should that Javadoc be updated?

These observations are absolutely correct and I have updated the Javadoc.

[abyrd]

Thanks for confirming those points. This looks like a step in the right direction to me, thanks for all your work on this. Understood about any renaming / documentation updates coming in a subsequent step, which I'd be happy to contribute to.