core: add ETCS SVL logic to ETCS braking simulator

[Erashin]

Fixes #10404.
Add SVL logic to ETCS braking simulator.
How: compute svl and eoa indications, compute minimum curve, and compare it against overlay.
Limit: approximate braking curves' calculation when SvL is after the end of the path, due to the missing tractive effort and slopes post path.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 80.64%. Comparing base (37c3b08) to head (574a3e7).
Report is 8 commits behind head on dev.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@            Coverage Diff             @@
##              dev   #10402      +/-   ##
==========================================
- Coverage   80.66%   80.64%   -0.03%     
==========================================
  Files        1098     1080      -18     
  Lines      111787   110825     -962     
  Branches      744      745       +1     
==========================================
- Hits        90176    89375     -801     
+ Misses      21568    21407     -161     
  Partials       43       43              

Flag	Coverage Δ
editoast	72.41% <ø> (+0.18%)	⬆️
front	89.93% <ø> (-0.01%)	⬇️
gateway	2.18% <ø> (ø)
osrdyne	2.53% <ø> (ø)
railjson_generator	87.58% <ø> (ø)
tests	?

Flags with carried forward coverage won't be shown. Click here to find out more.

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

[Erashin]

@bougue-pe : fixed in e3dbe47 and 1e7ce69.

[bougue-pe]

👏 Nice, thank you very much for the work !

Only nits and not-mandatory comments (or even just ideas for the future): no problem resolving them by yourself once decided/replied 👍

[eckter]

Good job on the tests

I only skimmed through the core changes, I trust PE's review on this

[bougue-pe]

As this SVL_STOP_OFFSET is not so "test generic", I would remove it and just explicitly use the named value:

Suggested change

	first_stop_offset = SVL_STOP_OFFSET
	final_stop_offset = 45_060_000
	stop_offsets = [
	0,
	first_stop_offset,
	svl_ph1_offset = 41_138_000
	final_stop_offset = 45_060_000
	stop_offsets = [
	0,
	svl_ph1_offset,

[bougue-pe]

Following the same idea:

Suggested change

	# Check that the curves does respect the EoA + SvL (EoA = stops), and that there is an
	# acceleration then deceleration in between (maintain speed when reach the MRSP).
	# Here, the EoA is at 41038m, and the SvL is at the switch at 41138m.
	distance_to_svl = 100_000
	first_stop_offset = SVL_STOP_OFFSET - distance_to_svl
	# Check that the curves does respect the EoA + SvL (EoA = stops), and that there is an
	# acceleration then deceleration in between (maintain speed when reach the MRSP).
	# Here, stop (EoA) is 100m before the PH1 switch (SvL).
	svl_ph1_offset = 41_138_000
	first_stop_offset = svl_ph1_offset - 100_000

[bougue-pe]

As mentioned, it looks to me like we shouldn't mutualize it, as it's only "lucky" that both EoA and SvL upper curve parts are under GUI influence and therefore separated by the same value as the EoA-SvL distance.
"Hard-documentation" though variable names seems too strong (and hides a lot of complexity about how the curves are computed) IMO.

Also, the values are not enough "test-generic" IMO (same as the SVL_STOP_OFFSET).

I wouldn't detail what's happening here at all (or just a comment, if you will).

Suggested change

	offset_start_first_brake = START_SVL_BRAKE_ON_PH1_SWITCH - distance_to_svl
	offset_start_first_brake = 33_361_530

🦜 Ditto for the "same location" test above.

[bougue-pe]

This is also part of the "end-of-the-path" hack (tracked it into #11036).
Yet, after thinking of it, we shouldn't have any of the 2 hacks in a separate commit, as the commit without the hacks will break: only one commit for the whole PR would be my pick on this 1️⃣.