XML2D - multivalue keys

[scleakey]

I think the except part might need some sort of warning in

floodmodeller-api/floodmodeller_api/xml2d.py

Lines 347 to 355 in 3ef6361

	if elem_key in self._multi_value_keys:
	if not list_idx_key == elem_key:
	list_idx_key = elem_key
	list_idx = 0
	try:
	self._recursive_remove_data_xml(new_dict[elem_key][list_idx], elem)
	list_idx += 1
	except (IndexError, KeyError):
	parent.remove(elem)

I was trying to set the link1d attribute but didn't realise it was in self._multi_value_keys so this would always fail silently, resulting in the links being completely absent in the saved xml file.

I realise this is kind of implied by:

floodmodeller-api/docs/source/user_guide/xml2d.rst

Lines 29 to 30 in 3ef6361

	- ``.link1d`` - A list of 1 or more 1D links, each containing a dictionary of parameters
	including the link shapefile and ief file.

and

floodmodeller-api/docs/source/user_guide/xml2d.rst

Lines 47 to 49 in 3ef6361

	Any option where multiple values can be provided (e.g. topography files) will be present
	as a list of dictionaries, whereas options where only one value can be required will be
	a single dictionary.

but it could be clearer.

[joe-pierce]

I've commited some changes that should capture when elements are not added as a list when they need to be, this needed to be done in the recursive update as well as the adding new keys in case an existing element is replaced with a non-list value.

I've also added a line that restores the xmltree in cases of errors. Also needed to update one of the tests to add elements as a list.

[scleakey]

I had to change my code from

floodmodeller-api/toolbox/Model Conversion/tuflow/component_converter.py

Lines 255 to 264 in b3b4a0e

	def edit_file(self) -> None:
	self._xml.domains[self._domain_name]["time"] = {
	"start_offset": self._start_offset,
	"total": self._total,
	}
	self._xml.domains[self._domain_name]["run_data"] = {
	"time_step": self._time_step,
	"scheme": self._scheme,
	}
	self._xml.processor = {"type": self._processor}

to

floodmodeller-api/toolbox/Model Conversion/tuflow/component_converter.py

Lines 255 to 264 in bf87626

	def edit_file(self) -> None:
	self._xml.domains[self._domain_name]["time"] = {
	"start_offset": {"value": self._start_offset},
	"total": {"value": self._total},
	}
	self._xml.domains[self._domain_name]["run_data"] = {
	"time_step": self._time_step,
	"scheme": {"value": self._scheme},
	}
	self._xml.processor = {"type": self._processor}

to get it to work after merging from main with your new commits. Is this intended?