I managed to get a few minutes to publish 0.9.9-beta today. It contains a slew of updates since 0.9.4-beta, including some corrections/bugfixes. Most importantly for me, it contains the new 'average' events that are emitted every time the engine sees a cycle from hour/day/week. These new events are now produced for each sensor. Originally they were just emitted for 'impulse' sensors, but now they are issued for normal 'sensors', since I re-factored some code to also take 'channel' into consideration. This therefore allows each 'channel' within a 'sensor' to emit these averages. Due to this change, 'impulse' sensors now carry the 'pseudo' channel of 1. Since the 'average' events were added, It was time to look again at the 'time' field and see what could be done, so that I could implement 'filesystem' support to re-read existing captures. I decided to leave the 'drifting time' alone, and just emit the re-captured events using the stored time, but allowing the user to supply a 'startdate' in the form of the 'vdate' parameter. My advice is actually to reset your basestation time regularly before the drift is more than 1 hour.
I will now try to update examples and provide a few more basic examples for use with the new features.
Its early days, please let me know how you get on and if this library is "hitting the spot" for you.
I do plan the include the 'history' data as produced within the (realtime) XML, and this is closer to reality now that I have the 'filesystem' support. I may implement this as a different type of 'mode'. For now, we have a sort of history in the way we can 'replay' xml files to reproduce these old events.
Expected EOD, 15-june-2013. 0.9.9-beta published.