Replace should have resource id parameter

[NaqGuug]

replace() expects that the resource id is included in the resource parameter. However, if you use scim2-models' contexts, like resource.model_dump(scim_ctx=Context.RESOURCE_REPLACEMENT_RESPONSE), the id will be set to None because it's read only. This combination means that the replace will never work. It only works if you don't use contexts or add the id manually.

My suggestion would be to add an id parameter to replace. Yes, users could manually add the id to the resource before calling it but honestly why not add that logic to replace itself. That way it's much clearer and easier for users to include the id, which comes from / goes to the PUT request.

[azmeuk]

replace() expects that the resource id is included in the resource parameter. However, if you use scim2-models' contexts, like resource.model_dump(scim_ctx=Context.RESOURCE_REPLACEMENT_RESPONSE), the id will be set to None because it's read only.

Actually you don't need to perform a model_dump before calling replace, it does this automatically with the right context:

scim2-client/scim2_client/client.py

Lines 549 to 551 in 315b306

	req.payload = resource.model_dump(
	scim_ctx=Context.RESOURCE_REPLACEMENT_REQUEST
	)

This way, IDs can be read automatically. However I don't like the current discrepancy about the ID parameter in replace and delete. I am not sure which direction is the good one: only read ID from the method params, or only read it from the object, or both. Related to #13

[NaqGuug]

Yes, it's great that you don't have to call model_dump before calling replace. However, the main problem here is that you CAN'T call model_validate beforehand without adding the id back manually.

Honestly I prefer the type + id method, it kinda makes more sense. It also would solve this issue where in some contexts the id is not in the resource itself. Just read the id from method params and ignore the object id.