Feature: Add setpoint check before applying strengths

[JeanLucPons]

Description, motivation and use case
In order to avoid that a set of magnet are only partially set, pyAML should check that all power supply setpoint and achievable.

Proposed solution

• Add a valid range to each powerconverter as in the example.
• Add a setpoint_check() method to MagnetModel that will basically compute_hardware_values() and check that all values are in the expected range. If one setpoint is out of range, a PyAMLException should be raised.
• Modify control/abstract_impl.py and add setpoint_check() calls when needed.

- type: pyaml.magnet.cfm_magnet
  name: SH2B-C04
  mapping:
    - [B0, SH2B-C04-H]
    - [A0, SH2B-C04-V]
    - [A1, SH2B-C04-SQ]
  model:
    type: pyaml.magnet.linear_cfm_model
    multipoles: [B0,A0,A1]
    units: [rad,rad,m-1]
    pseudo_factors: [1.0,-1.0,-1.0]
    curves:
      - type: pyaml.configuration.csvcurve
        file: sr/magnet_models/SH2_h_strength.csv
      - type: pyaml.configuration.csvcurve
        file: sr/magnet_models/SH2_v_strength.csv
      - type: pyaml.configuration.csvcurve
        file: sr/magnet_models/SH2_sq_strength.csv
    matrix:
      type: pyaml.configuration.csvmatrix
      file: sr/magnet_models/SH_matrix.csv
    powerconverters:
      - type: tango.pyaml.attribute
        attribute: srmag/ps-corr-sh2/c04-b-ch1/current
        range: [-2.0,2.0]  # Power supply range
        unit: A
      - type: tango.pyaml.attribute
        attribute: srmag/ps-corr-sh2/c04-b-ch2/current
        range: [-2.0,2.0]  # Power supply range
        unit: A
      - type: tango.pyaml.attribute
        attribute: srmag/ps-corr-sh2/c04-b-ch3/current
        range: [-2.0,2.0]  # Power supply range
        unit: A

[gupichon]

Should the 'range' field be mandatory or not?

[gupichon]

Maybe the availability of the powerconverters should be tested too ?

[JeanLucPons]

I would say that the range is not mandatory. If the range exists the validity check can be done otherwise we consider that the setpoint can be sent to the PS.

Concerning, the state of underlying PS (is there a FAULT or not) ,I would say that this should also be handled but here some discussions are necessary to find an agreement on PS states management. Today the power converter is a DeviceAccess object, we may need an extra layer if we don't manage to define a common behavior concerning PS states.

[gupichon]

State management should be handled by each institute (and even within an institute, there can be debates about it). Maybe we only need a generic module to test the availability of the underlying hardware (for example, a simple ping for Tango)?

[gupichon]

Since device access is implemented in the control system layer and is generic for all devices (including BPMs, for instance), I will add check_write_value(float value) and check_device_availability() methods to the device access. This will give us the possibility to use the control system tools. For Tango, the min and max values are available through the device interface, so this field is not even necessary.

[JeanLucPons]

Yes if the range is not available in the config file, it can retrieved from the CS device itself.
I would not add a check_write_value() in DeviceAccess but rather a get_range() with a priority for the range defined in the config file.

I'm not really in favor of having a ping() in the setpoint check, this should be calculation only. For EBS, writing all the strengths is already long. If the writing fails due to a PS fault, a PyAML exception is returned.

[gupichon]

A ping is the fastest way to reach a device; it does not take long. Basically, it is a default command that does nothing. It is even faster than checking the state. I suggest either doing a simple ping and letting each institute implement its own behavior, or not implementing it yet.

[gupichon]

Ok for the get_range()

def get_range(self) -> list[float, float]:
    pass

With nan's if there ther is no boundary ([0.0, nan] -> 0 to infinite)