This is a tool to quickly setup SQLAlchemy object relation mappings (ORMs). It uses SQLAlchemy's reflection tools to autoload table information from existing databases. ORM instances are configured using a dictionary describing which tables to map, what classes to map them to, and how they're related.
I've placed a commented demo in the "TestManyToManySelf" test case defined in "irrealis_orm/tests.py".
I have written bioinformatics tools to analyze data stored in Chado database tables, but the number and complexity of table definitions in the Chado schema is daunting to any programmer: the schema is highly normalized (see http://gmod.org/wiki/Chado_Tables) and reflects the complexity of the gene ontology (see http://www.geneontology.org). There are, literally, at least a hundred table definitions in the schema. This made for a great deal of repetitive and utterly boring grunt work before I could start accessing and using the data.
To save time I used SQLAlchemy's database introspection tools to automatically load table definitions from the database. Unfortunately, for a variety of reasons it is usually impossible to use database introspection to infer SQL relationships describing directed graphs, such as those used in Chado. Since this is usually futile, SQLAlchemy doesn't try to do it. So it must be done manually, and after doing it four or five times I wrote this tool to simplify the workflow.
I then found myself using the tool whenever I wanted a Python ORM to a database I didn't design myself.
To run tests, type "python -m irrealis_orm.tests". If you have "nose" and "coverage" installed, type "nosetests --with-coverage --cover-package=irrealis_orm".