[13.0] base_rest: New api

[lmignon]

Foward port of #48

[HviorForgeFlow]

Not sure if related with this PR but at start running an instance I've got this error:

2020-05-26 11:24:37,780 1 CRITICAL test odoo.service.server: Failed to initialize database `test`. 
Traceback (most recent call last):
  File "/opt/odoo/custom/src/odoo/odoo/service/server.py", line 1179, in preload_registries
    registry = Registry.new(dbname, update_module=update_module)
  File "/opt/odoo/custom/src/odoo/odoo/modules/registry.py", line 86, in new
    odoo.modules.load_modules(registry._db, force_demo, status, update_module)
  File "/opt/odoo/custom/src/odoo/odoo/modules/loading.py", line 535, in load_modules
    model._register_hook()
  File "/opt/odoo/auto/addons/base_rest/models/rest_service_registration.py", line 62, in _register_hook
    self._build_controllers_routes(services_registry)
  File "/opt/odoo/auto/addons/base_rest/models/rest_service_registration.py", line 66, in _build_controllers_routes
    for service in self._get_services(controller_def["collection_name"]):
  File "/opt/odoo/auto/addons/base_rest/models/rest_service_registration.py", line 98, in _get_services
    return work.many_components()
  File "/opt/odoo/auto/addons/component/core.py", line 438, in many_components
    return [comp(work_context) for comp in component_classes]
  File "/opt/odoo/auto/addons/component/core.py", line 438, in <listcomp>
    return [comp(work_context) for comp in component_classes]
  File "/opt/odoo/auto/addons/connector_search_engine/components/mapper.py", line 20, in __init__
    exporter = work.index.exporter_id
AttributeError: 'WorkContext' object has no attribute 'index'

[lmignon]

@HviorForgeFlow Thank you for the report. I'll take a look ASAP.This error is strange. The mapper component should not be returned by the lookup into the component registry since it's not registered for the rest service collection..
One I've found the issue I'll come back to you

[lmignon]

@HviorForgeFlow The encountered issue should be resolved by the 2 last commits.

[HviorForgeFlow]

It is working fine now thanks!

[lmignon]

@HviorForgeFlow Thank you for the feedback. If you need to develop new services, I encourage you to use the new api and datamodels. It's much more flexible and easier to maintain.

[lmignon]

@guewen @gurneyalex The new API is now also available for Odoo 13. It would be nice if you could use-it / test-it with your new WMS mobile app 😏

[gurneyalex]

hello @lmignon I see #48 is flagged as "in progress". Is it also the case for #71 ?

[lmignon]

hello @lmignon I see #48 is flagged as "in progress". Is it also the case for #71 ?
Good catch @gurneyalex. In fact, both of them needs review.

[kevinkhao]

@lmignon before I do a PR, I'd like to know what you think.
Here is some context: https://github.com/acsone/rest-framework/blob/37b69406818a82deedccbb0444c731cf12a2fc20/base_rest/restapi.py#L194
It's about functions from_params and to_response.
For our controllers, when we use Datamodel and because we most probably just use my_input_datamodel.dump() the on the input and my_output_datamodel.load() before returning our result:
Would it make sense to:

1. dump our datamodel so that it is directly available in the controller's arguments ? i.e: in function from_params
return ModelClass.load( params, many=self._is_list, unknown=marshmallow.EXCLUDE ).dump()
2. allow to return a dict and it is automatically load()ed on the datamodel ? i.e in function to_response
   remove lines 204 to 208
   errors = ModelClass.validate(result, many=self._is_list)
   What do you think ?

Currently it does seem to me as bringing unnecessary steps, but I am maybe missing something.

[lmignon]

@kevinkhao Sorry but I don't understand your question. The @Datamodel decorator is designed to decorate methods where the result of the call to the method is an instance of Datamodel and the param is also Datamodel instance. This decorator allows to only deal with datamodel objects on the service layer implementation.

[kevinkhao]

@kevinkhao Sorry but I don't understand your question. The @Datamodel decorator is designed to decorate methods where the result of the call to the method is an instance of Datamodel and the param is also Datamodel instance. This decorator allows to only deal with datamodel objects on the service layer implementation.

My idea was that when we define a route, we should just return a dict and the module will take care to do self.env.datamodel["my.output.datamodel"].load(resulting_dict) so that we can save 1 line at the end of our route definition. But after thinking about it, I'm still not sure about that idea.

Calling at the end of all my routes:
return self.env["my.datamodel"].load(myResultDict)
instead of
return myResultDict
is more verbose but more clear.

[lmignon]

My idea was that when we define a route, we should just return a dict and the module will take care to do self.env.datamodel["my.output.datamodel"].load(resulting_dict) so that we can save 1 line at the end of our route definition. But after thinking about it, I'm still not sure about that idea.

@kevinkhao But we don't have to declare a route by ourself. Routes are dynamically generated by the framework based on the @restapi.method decorator 262def2

[simahawk]

@lmignon sorry, was on holidays and missed your msg. I'll skim through your last changes in a while.

[lmignon]

@hparfr, @sebastienbeau Are you OK with this PR?

[hparfr]

The tests I had conducted so far on WMS shows that it seams to work.
I will do more testing when porting shopinvader soon and open issues or propose PR if needed.

[OCA-git-bot]

This PR has the approved label and has been created more than 5 days ago. It should therefore be ready to merge by a maintainer (or a PSC member if the concerned addon has no declared maintainer). 🤖

[simahawk]

@lmignon good to merge then?

[lmignon]

/ocabot merge major

[OCA-git-bot]

This PR looks fantastic, let's merge it!
Prepared branch 13.0-ocabot-merge-pr-71-by-lmignon-bump-major, awaiting test results.

[sbidoul]

[OCA-git-bot]

Congratulations, your PR was merged at e9f76f2. Thanks a lot for contributing to OCA. ❤️

[sebastienbeau]

youhou \o/