Elementholder api refurbishment

[gupichon]

Description

Integration PR for epic #199, ElementHolder API refurbishment. Aggregates the sub-features that split the historically monolithic ElementHolder into typed sub-holders (.magnet(s), .bpm(s), .rf, .diagnostic, .tool, ...) reachable through Python properties, with typed get() and array access instead of the old untyped get_all*() and get_*s() list-returning methods.

This branch is merged incrementally as each sub-issue lands, rather than opened as a single large PR, to keep review scoped per sub-feature.

Related Issue

• Tracks ElementHolder API refurbishment #199

Sub-issues and their status on this branch:

• Feature: Add generic ElementHolder collection access #366, add generic ElementHolder collection access. Closed, merged to main before this branch was cut.
• Feature: Complete magnet holder navigation #373, complete magnet holder navigation. Merged via Complete magnet holder navigation (#373) #426, get_cfm() shortcut and dynamic attribute access on GenericArrayHolder based holders.
• Feature: Add a public diagnostic holder #372, add a public diagnostic holder. Not yet started.
• Feature: Add an RF masterclock facade #374, add an RF masterclock facade. PR Add an RF masterclock facade #427 was closed without merging. Reviewers flagged that masterclock.voltage is conceptually wrong, voltage needs its own facade, a cavity concept was proposed, instead of hanging off the default RF plant. Re-opening this requires a new design, not a resubmission of Add an RF masterclock facade #427's diff.
• Feature: Add a public tuning and measurement tool holder #375, add a public tuning and measurement tool holder. Implemented, ToolHolder, tool.get(name=None), typed tune, trm, orbit, orm, chromaticity, crm, dispersion properties with type validation. Tests added, full suite green. PR not yet opened.
• Feature: Unify name/wildcard access and miss-behavior across all holders #433, unify name and wildcard access and miss behavior across all holders, new issue. Surfaced while implementing Feature: Add a public tuning and measurement tool holder #375, sr.live.bpm[...] has no wildcard support at all, and every holder disagrees on what happens when a name is missing.

Changes to existing functionality

• None beyond what each merged sub-issue's own PR already documents, see Complete magnet holder navigation (#373) #426 for Feature: Complete magnet holder navigation #373, and Feature: Add a public tuning and measurement tool holder #375's own PR for the ToolHolder aliasing. This PR does not itself change behavior, it only aggregates already reviewed sub-PRs.

Testing

No tests added directly on this integration branch, each sub-issue's PR carries its own tests, see #426 for #373's tests/common/test_array_holder_navigation.py, and #375's PR for tests/tuning_tools/test_tool_accessors.py.

Verify that your checklist complies with the project

• New and existing unit tests pass locally
• Tests were added to prove that all features and changes are effective, covered per sub-PR
• The code is commented where appropriate, covered per sub-PR
• Any existing features are not broken

[JeanLucPons]

If bpm(s) moves to diagnostic holder, bpm.py should ne moved in diagnostics directory.
and empy bpm folder removed.

[gupichon]

If bpm(s) moves to diagnostic holder, bpm.py should ne moved in diagnostics directory. and empy bpm folder removed.

Yes, but @gubaidulinvadim suggested doing it later in another PR. Would you prefer to do it in this one? Personally, I don't mind.

[gupichon]

Everything has been merged here. I will rebase this afternoon and request a review.

@JeanLucPons, @TeresiaOlsson, @GamelinAl, @simoneliuzzo, @kparasch, @gubaidulinvadim and anyone else: feel free to start reviewing right now, as things won't change much after the rebase.

[kparasch]

Very minor comment. Could we simplify with:

return self._validate_type(name, self.get(name), Tune)

?

[kparasch]

I tend to prefer breaking backward-compatibility this early in development.

If we say that sr.live.tool.orm is the preferred way to access the orm tool, we should enforce it.

[kparasch]

Looks ok to me, some minor comments only

[GamelinAl]

Two issues:

1. Compared to #199 we still miss the support of ElementHolder get method for named arrays: sr.live.get("CELL08"). Only the old syntax sr.live.get_elements("CELL08") is working.
2. We should also clean the API, I would say we can remove:
   • get_element
   • get_elements
   • get_all_elements
   • get_*_tuning
   • get_betatron_tune_monitor

After this it should be good I think.