Add an API to instantiate an instrument without configuration file

[gupichon]

Description:
Currently, an instrument can only be created from a YAML configuration file.
It should also be possible to instantiate an instrument programmatically using a direct API call.

Example use case:
Developers may want to:

• Build instruments dynamically (e.g., from in-memory definitions or database records)
• Use PyAML in scripts or tests without maintaining a YAML file on disk

Expected feature:
Provide a Python API such as:

from pyaml import Instrument

instr = Instrument.from_dict({
    "name": "ring",
    "devices": [...],
})

[TeresiaOlsson]

Why does it need to include a dict? I thought the user could just create objects for the required configuration dataclass/es and then use them to generate the objects with logic since the config classes already exist?

[gupichon]

Sounds good — we can support both approaches. I initially leaned toward a plain dict because it mirrors the YAML schema and is straightforward to implement. That said, building from typed objects (or a configuration object) is also feasible. We can keep the dict-based API for parity with YAML and add an object-based builder, both sharing the same validation layer.

[JeanLucPons]

It should be possible to write something like:

control = TangoControlSystem(name="live",tango_host="ebs-simu-3:10000")
qfCurve = CSVCurve(file="config/sr/magnet_models/QF1_strength.csv")
qAtt = Attribute(attribute="srmag/vps-qf1/c01-a/current")
qModel = LinearMagnetModel(name="QF1",curve=qfCurve,calibration_factor=1.00504,powerconverter=qAtt,unit="1/m")
quad = Quadrupole(model=qModel,peer=control)
quad.strength.set(0.8)

[TeresiaOlsson]

It should also include data validation in the same way as you get when using the yaml file.