Update bba callback

[JeanLucPons]

This PR improve BBA callback using new pySC (1.5.3).

Example:

from pyaml.accelerator import Accelerator
from pyaml.common.constants import Action
import numpy as np

ios_x = {"H":[],"V":[]}
ios_y = {"H":[],"V":[]}

def bba_callback(action: int, cdata):

    # Action.MEASURE is triggered at each new IOS measurement
    if action == Action.MEASURE:
        plane = cdata["plane"]
        step = cdata["step"]
        ios_x[plane].append(cdata["bpm_pos"])
        ios_y[plane].append(cdata["ios"])
        if len(ios_x[plane])>=2:
            # Get slopes
            fit = np.polynomial.polynomial.polyfit(ios_x[plane], ios_y[plane], 1)
            # Extract small slopes
            std = np.std(fit[1])
            cfit = fit[:,np.where(np.fabs(fit[1])>=std)[0]]
            # Get average intersection point (-> offset)
            x = np.mean(-cfit[0] / cfit[1])
            print(f"step={step} plane={cdata["plane"]} pos={cdata["bpm_pos"]*1e6:7.3f} um offset={x*1e6:.3f} um")
    return True

sr = Accelerator.load("tests/config/EBSOrbit.yaml")
SR = sr.design
bba = SR.get_bba("BBA-BPM_C04-04")

# Add a misalignement
SR.get_bpm("BPM_C04-04").offset.set([20e-6,-15e-6])

bba.measure(callback=bba_callback)

print(f"HOffset: {bba.h_offset()}")
print(f"VOffset: {bba.v_offset()}")

outputs:

step=1 plane=H pos= 52.151 um offset=19.750 um
step=2 plane=H pos= 36.095 um offset=19.864 um
step=3 plane=H pos= 19.995 um offset=19.959 um
step=4 plane=H pos=  3.854 um offset=20.036 um
step=5 plane=H pos=-12.324 um offset=20.097 um
step=6 plane=H pos=-28.534 um offset=20.141 um
step=1 plane=V pos=  6.527 um offset=-15.002 um
step=2 plane=V pos= -4.236 um offset=-15.001 um
step=3 plane=V pos=-15.000 um offset=-15.000 um
step=4 plane=V pos=-25.764 um offset=-15.000 um
step=5 plane=V pos=-36.527 um offset=-15.000 um
step=6 plane=V pos=-47.290 um offset=-15.000 um
HOffset: 2.0149059939587564e-05
VOffset: -1.5000000000143988e-05

[JeanLucPons]

@gupichon-soleil why dtaacc is failing now ?

[gubaidulinvadim]

@gupichon-soleil why dtaacc is failing now ?

@JeanLucPons Guillaume is on vacation now. It's a bit strange if it fails. Normally, everything was tested before he went on vacation. @patrickmadela could you maybe check the reason of failure?
@JeanLucPons is your branch based on the latest pipeline version? #306 It looks like the path to apptainer image is different now

[JeanLucPons]

@gubaidulinvadim
My branch is based on main with 2 more commits.
No change in workflows.

[ubuntu20acu.pons] > git log
commit d6ea9a6ab8fb8f9541981a9a27c5771c7fc2668e (HEAD -> update-bba-callback, origin/update-bba-callback)
Author: PONS <pons@esrf.fr>
Date:   Thu Jul 30 10:06:41 2026 +0200

    Updated pySC release number

commit 8afb54277a1a0b240eac22524320a761d8dbab8d
Author: PONS <pons@esrf.fr>
Date:   Wed Jul 29 09:43:20 2026 +0200

    Update bba callback

commit ac22eaf8a75a1d3dc7cfcc22678f0359b8b72afc (origin/main, origin/HEAD, main)
Merge: ed3f6b52 4e08e56f
Author: Jean-Luc PONS <pons@esrf.fr>
Date:   Fri Jul 24 15:09:18 2026 +0200

    Merge pull request #279 from python-accelerator-middle-layer/schema-registry-rf
    
    Adding schema registry, validator and generator for RF

[kparasch]

I would be tempted to just merge since the changes are very small and the test that is failing does so for reasons irrelevant to this PR.

[patrickmadela]

@gupichon-soleil why dtaacc is failing now ?

@JeanLucPons Guillaume is on vacation now. It's a bit strange if it fails. Normally, everything was tested before he went on vacation. @patrickmadela could you maybe check the reason of failure? @JeanLucPons is your branch based on the latest pipeline version? #306 It looks like the path to apptainer image is different now

I checked the logs, but the error doesn't look familiar to me. It is probably related to the changes and improvements @gupichon-soleil made, but it will take some time to determine the root cause for me.

[gubaidulinvadim]

I would still suspect this commit 07f529d that is in Guillaume #306 but not in the main. But if there's no easy solution, I think it's best to wait for his return.