search

python-microscope / microscope

Python library for control of microscope devices, supporting hardware triggers and distribution of devices over the network for performance and flexibility.
https://www.python-microscope.org
GNU General Public License v3.0
73 stars 42 forks source link

Harmonise the enable/disable interface #195

Closed dstoychev closed 3 years ago

dstoychev commented 3 years ago

I feel like disable() and enable() have too broad use at the moment. Some devices (e.g. DeformableMirror) take it as "prepare the device for operation", whereas others (e.g. LightSource) directly operate the device. I think this is a huge discrepancy in usage and I propose we stick to the former definition. As such, light sources will need a new pair of methods, something along the lines of switch_on() and switch_off(). These new methods will not be able to guarantee that the device will switch to the desired state because often hardware (TTL) signals override the software control. Example signatures:

@abc.abstractmethod
def switch_on(self): -> bool:
    # Return true on successful communication, not if the device has switched on
    pass

@abc.abstractmethod
def switch_off(self): -> bool:
    # Return true on successful communication, not if the device has switched off
    pass

This is related to #171.

mickp commented 3 years ago

With continuous wave light sources, turning them in is preparing them for operation. Generally, switching untriggered light sources on requires a settling time until the output power is stable. Modulation is achieved by some other device, such as a shutter or an AOM. I think the proposed methods are redundant: enabling a CW light source will involve calling switch_on. If you call switch_off, it's then no longer ready for use in a measurement, so should not be considered enabled.

dstoychev commented 3 years ago

I agree. I raised this issue while working on an implementation of a compound light source, such as the ones from CoolLED and Lumencor, where the device may be enabled yet not emitting light because it's waiting for a hardware trigger to modulate it. But it turns out that such light sources can still be implemented with the current LightSource class, so this issue is not longer relevant.