Logging in `BaseModel.from_config` methods.

[davidorme]

Describe the bug

I've been going over the Variables PR #431 to try and push it along now that @vgro has broken the back of the update.

• I've updated all the variable descriptions
• I've updated all the model required vars sections.
• I'm going through the example log (which very helpfully prints out what data gets checked and set at each step - aren't we clever!) to validate all of the sections.

But as part of going through the log, I noticed a difference between the PlantsModel.from_config() and everything else. The other models do this (for example):

virtual_ecosystem/virtual_ecosystem/models/litter/litter_model.py

Lines 171 to 179 in a27518a

	LOGGER.info(
	"Information required to initialise the litter model successfully "
	"extracted."
	)
	return cls(
	data=data,
	core_components=core_components,
	model_constants=model_constants,
	)

Plants does this:

virtual_ecosystem/virtual_ecosystem/models/plants/plants_model.py

Lines 154 to 169 in a27518a

	# Try and create the instance - safeguard against exceptions from __init__
	try:
	inst = cls(
	data=data,
	core_components=core_components,
	flora=flora,
	model_constants=model_constants,
	)
	except Exception as excep:
	LOGGER.critical(
	f"Error creating plants model from configuration: {excep!s}"
	)
	raise excep
	LOGGER.info("Plants model instance generated from configuration.")
	return inst

Okay - so that is a little paranoid but there is a practical difference in the logging output. With plants, there is a message after all the init logging indicating model success. The other models report an intermediate step - they only claim to have extracted the information needed - but the messages read a bit like the __init__ has completed and then are followed by all the logging messages.

I think we should shift to using a Plants like setup for all model - it makes the logs clearer - but we should be consistent either way.

To Reproduce

Run ve_example and look at the log.