esss / flask-restalchemy Goto Github PK

Currently, Swagger spec is generated using Flasgger. But to use Flasgger, each REST resource must have its unique class, leading to the hackish code at resourcefactory.py.

Find a better way to generate Swagger specs (maybe made an script available to compile the spec based on the models registered on the API).

Add a deploy stage to .travis.yml as in ESSS/cookiecutter-esss-pypackage.

Build and deploy by rever can be removed (conda-forge bot will do the work).

Currently, relationship endpoints doesn't support any query parameter (pagination, order by, limit). This should be implemented for these kind of endpoints.

Currently, we have to define a function and then call api.add_url_rule providing the function. It would be nice to have a decorator to add a function as an url_rule of the api, as flask does with the 'app.route' decorator.

Current behavior:

api = Api(flask_app)

def my_rule_fn():
    # do stuff 

api.add_url_rule('/my_rule', my_rule_fn)

After requested change:

api = Api(flask_app)

api.route('/my_rule')
def my_rule():
    # do stuff 

Hi all, I think this project is an excellent idea, since manually stitching together Flask + SQLAlchemy + Flask-restful + Marshmallow involves a lot of repetition and boilerplate.

However, I realise that for my project, I need to be able to add custom view endpoints, so I can add filter logic to them, and other behaviour that handles the GET args. There seems to be some already existing code that does this kind of thing, but is it possible to fully customize it? I can envisage the default filter builder not working exactly for my use-case. In addition, would it be possible to document these more advanced features?

Currently, serializers are instancialized in the model registartion (add_model, add_relationship).

This requires that register_column_serializer to be called before any mode registration, which is a burden for those that want to customize field serializers.