The Hand of FuManChu

Dejavu, my pure-Python ORM, has a new Trac home at http://projects.amor.org/dejavu. As always, it's open to the public to download, use, or develop. It could use another contributor or two, and the snazzy new Trac front-end should help that immensely.

Just to remind you about the parts of Dejavu I think are cool:

1. It completely hides the storage layer, so your code need never know you're using a database. It supports MS Access, MS SQL Server, PostgreSQL, MySQL, SQLite, and shelve, all transparently.
2. Simple queries use simple syntax: Panda = sandbox.unit(zoo.Animal, Species='Ailurpoda', Name='Mei'). More powerful queries use the logic.Expression object and pure Python (no code strings!):

   red_filter = logic.Expression(lambda x: x.color == 'red')
   RedAnimals = sandbox.recall(zoo.Animal, red_filter)
   

3. Dejavu has been designed from the ground up to be used in a multithreaded environment.
4. New database adapters are easy to develop using the `db` module. The PostgreSQL module, for example, is 100 lines of code.
5. Model-layer Units are clean and easy, using modern, built-in Python types. They are also extendable to custom property classes and custom types:

   from dejavu import Unit, UnitProperty
   
   class MissionTrip(dejavu.Unit):
       """A Mission Trip experience."""
       
       DirectoryID = TripGroupProperty(int, index=True)
       FirstDate = NotifyProperty(datetime.date)
       GroupName = UnitProperty()
       LastDate = NotifyProperty(datetime.date)
       AmountDue = UnitProperty(fixedpoint.FixedPoint)
   

6. Persistent query-engines and analysis tools are included.