ygbr / sqlasagna Goto Github PK

should we instance temporary objects and return them or use the sessions query / methods with execute ?

The group_by should return a grouped query object as in:

db.session.query(db.TableObj).group_by(db.TableObj.whatever)

Check how to map those kinds of relationships, and specially how do we identify association tables.

We should define how to configure SQLAlchemy configuration.
We should be able to externally configure the engine and connection. We also should be able to reload mapping / reconnect.
We should be able to change connection strings on the fly and reconnect.
This is required for turning SQLasagna into a python package for public usage.

Tables shall have a working count implementation.

We should get the relations based on ForeignKey analysis of the metadata reflected table lists.
Then we should update all relationship-linked objects and create the backref's.

This class will also allow external configuration. Something like:

from sqlasagna import sqlasagna

db = sqlasagna( engine_connection_params, mapper_customization_dict, relationship_customization_dict )

db.users.all()

db.users.query()

etc...

This way we can override relationships, backrefs, one-to-one, many-to-many configs.

Also this will allow us to disable relationship mappings which we do not want and set lazy or eager loading styles by default per relationship.

This should probably be a dict, past to the Lasagna initializer.

I cannot find sqlasagna in pypi, and therefore also not install it as described in the docs.

either the docs should be adjusted or the tool uploaded to pypi

After we have a class and configurations we can have a true python package :)