Maritime schema integration

[minhemdnv]

These changes should allow the Types from maritime-schema to be used. The main changes are in types.py (now much shorter), as shared types live in maritime schema. Also some changes here and wherever an error was thrown. As far as I can see, the output generated is the same.

[ClaasRostock]

Great work. Approved with some comments.
Most comments are minor. The only one that needs to be resolved before merging is the one in pyproject.toml (how the maritime-schema package is referenced).

[ClaasRostock]

change to maritime-schema = "^0.0.3" once the maritime-schema package is published.
(This line is the reason why the GitHub workflow fails. The workflow does not have permission to clone the maritime-schema repository)

[minhemdnv]

also done :)

[ClaasRostock]

Installing the ship-traffic-generator package on my own machine I found out that we also need to limit the python versions we allow.

python = "^3.9"
needs to be changed to
python = ">=3.9, <3.13"

This is needed because "^3.9" effectively means ">=3.9, <4.0" , and hence includes 3.13
There are multiple other dependencies, though, that require python <3.13 (which makes sense. Also we ourselves would want to require that).
If we implicitly allow python 3.13 (which "^3,9" does), then dependency resolution fails.

[ClaasRostock]

we could think about applying the same improvements here that you introduced also in the caga.py module in maritime-schema. At least two minor things come to my mind; maybe there are more:
a) the to_camel function we could import from maritime_schema.utils.strings instead of implementing it here again
b) I remember you improved the way the model 'config' was handled, in maritime-schema. You created an additional base class that all models then inherited from. Not sure, but maybe this is an improvement that is worth to apply also here in ship-traffic-generator?

[minhemdnv]

sure, I will look into these!

[minhemdnv]

@tomarnepedersen I've fixed most of the issues that Claas pointed out, but perhaps you also want to take a look to make sure you are happy with the changes. I tried my best not to change anything too drastically! To get most things to work, I've had to modify the maritime-schema, and change a lot of fields to Optional. This is because in ship-traffic-gen the classes such as Initial() are created which are completely empty, and then subsequently populated. If the fields were not optional, Pydantic would throw an error and not allow the class to be created. But, setting all fields to optional makes it harder to use the built in Validation features that Pydantic provides, so in the long run I would consider refactoring some parts of ship-traffic-gen so that martime-schema classes are only created once all the values that are required actually exist.

[tomarnepedersen]

@tomarnepedersen I've fixed most of the issues that Claas pointed out, but perhaps you also want to take a look to make sure you are happy with the changes. I tried my best not to change anything too drastically! To get most things to work, I've had to modify the maritime-schema, and change a lot of fields to Optional. This is because in ship-traffic-gen the classes such as Initial() are created which are completely empty, and then subsequently populated. If the fields were not optional, Pydantic would throw an error and not allow the class to be created. But, setting all fields to optional makes it harder to use the built in Validation features that Pydantic provides, so in the long run I would consider refactoring some parts of ship-traffic-gen so that martime-schema classes are only created once all the values that are required actually exist.

See your point. Of course, possible to change the code in later release to do as you suggest. The reason for not having the fields optiona I guess is to not have to have checks if the parameter includes a value?
Will have a look at your changes.

[minhemdnv]

The reason for not having the fields optiona I guess is to not have to have checks if the parameter includes a value?

Exactly. But I agree, we can fix this later on

[tomarnepedersen]

Already merged a branch which was based on this one. This could be deleted