Remove Optional typing for arrays.

[axelv]

Optional typing for an array can be annoying because it forces you to check for None each time you want to iterate over a array.
A better solution might be (which I implemented in this pull request):

• Remove the Optional[] typing for all arrays (by which I mean elements with max > 1)
• Use Field(default=...) for required elements
• Use Field(default_factory=list) for optional array elements
• use Field(default=None) for optional (non-array) elements

This makes it more convenient to iterate over array elements:

my_bundle = Bundle.parse_file("my_bundle.json")
assert my_bundle.entry is not None, "No entries in this bundle"
for entry in my_bundle_entry:
    pass # do stuff

becomes

my_bundle = Bundle.parse_file("my_bundle.json")
for entry in my_bundle_entry: # entry is always a list
    pass # do stuff

Note that for required elements the default=... will validate that a non-empty list is passed when parsing.
More info on this https://docs.pydantic.dev/latest/usage/models/#required-optional-fields

[m0rl]

Thanks for the proposal! I understand the issue (and wonder how much easier things like that would be should we have optional chaining available in python) but I'm afraid with the changes we would also have to have a different BaseModel (that will automatically exclude default_factory generated data on the resource serialization). Which in its turn will make things more complicated.

[ruscoder]

I like this idea, and it's not an issue to adjust automatic exclude of default factories during serialized since we already have a custom BaseModel

[ruscoder]

I merged main into your branch and changed the behaviour by getting rid of Ellipsis for literals - because it has broken the resource initialization.

Also, I've adjusted model_dump and model_dump_json by setting exclude_* to true, and exclude_unset removes all default values that were not explicitly set during initialization - it closes the concern of @m0rl

New PR - #21