Richard Jones' Log: PyCon: "ATOP on BSDDB" looks like the Roundup hyperdb

People probably don't realise that the Roundup hyperdb (which is Atomic, Transactional, Object-Oriented and Persistent) may be used outside of Roundup the issue tracker. Utterly contrived example:

from roundup.hyperdb import String, Number, Multilink
from roundup.backends.back_bsddb import Database, Class

class config:
    DATABASE='/tmp/hyperdb_example'

db = Database(config, 'admin')
# define a simple schema
spam = Class(db, 'spam', name=String(), size=Number())
widget = Class(db, 'widget', title=String(), spam=Multilink('spam'))

# add some data
oneid = spam.create(name='one', size=1)
twoid = spam.create(name='two', size=2)
widgetid = widget.create(title='a widget', spam=[oneid, twoid])

# dumb, simple query
print widget.find(spam=oneid)
print widget.history(widgetid)

... which displays ...

['4']
[('4', <Date 2004-03-25.01:16:1.386365>, '1', 'create', {})]

And of course in the hyperdb, you've got multiple backends to choose from (anydbm, bsddb, bsddb3, metakit, sqlite, mysql and postgresql), full journalling of changes (may be turned off selectively), automatic behaviours (through detectors auditing and reacting), object querying and full-text indexing built-in.

Of course, having come up with this example, I've seen some very simple API changes that could make it even easier to use the hyperdb outside of Roundup. If there's interest, that is...