Tableau

Tableau is a collection of helper classes for building test fixtures and seed data.

• Model composition without any predefined schemas. The model object relationships are automatically deduced through the value annotations:

  from tableau import Datum
  
  foo = Datum(
    'Foo', ('id'), # schema name / identifiers
    id=1,
    field_a=1,
    field_b=2,
    collection=one_to_many([
      Datum(
        'Bar', auto('id'), # id will be automagically generated by the walker
        field_c=1
        ),
      Datum(
        'Bar', auto('id'),
        field_c=1
        ),
      ],
      referring_fields='foo_id'
      )
    )

• Export the model object graph to plain ANSI SQL statements, in the ORM-agnostic manner:

  import sys
  from tableau import Datum, DataSuite, DataWalker
  from tableau.sql import SQLGenerator
  
  # ...
  
  suite = DataSuite()
  DataWalker(suite)(foo)
  SQLGenerator(sys.stdout, encoding='utf-8')(suite)

  The above yields the following SQL statements:

  INSERT INTO `Foo` (`id`, `field_a`, `field_b`) VALUES
  (1, 1, 2);
  INSERT INTO `Bar` (`id`, `field_c`, `foo_id`) VALUES
  (1, 1, 1),
  (2, 1, 1);

• Automatically mapping the existing SQLAlchemy tables / declarative classes to the Data:

  from tableau.sqla import newSADatum
  
  # metadata = ...
  # Base = ...
  # session = ...
  
  class Foo(Base):
    __tablename__ = 'foos'
    id = Column(Integer, primary_key=True)
    field = Column(String)
  
    def some_model_specific_method(self):
      return self.field
  
  Datum = newSADatum(metadata, Base)
  datum = Datum(
    'foos', ('id'),
    field='test'
    )
  print datum.some_model_specific_method() # 'test'
  session.add(datum) # it can even be added to the session!