pyRete status 4

The public API is starting to come together and looks quite OK. I'll have to go through Clips user and programming manuals again to make sure that I haven't missed any obvious things but overall I'm quite pleased.

The "last" thing I need to do is to reimplement parts of the execution engine in order to get modules and focus working right. At the moment, if you set a focus and then call run() the agenda is created according to current focus and activations just as you'd expect. However, if you construct a focus stack of two or more modules only rules from the "last" one (current focus) will be executed.

Next step (make it right) according to my simple project plan (Make it work, make it right, make it fast ;-) is to polish up the structure and readability of the code. I've started a re-write according to PEP-8 (Style guide for Python code) but I haven't come that far yet.

There are also several things I'd like to re-implement from an architectural or algorithmic point of view. I'll have to figure out the best and most pythonic way to structure the modules and I'll probably want to re-write all of my tests using some form of framework. It would be nice to have another go at the rule compiler and the rete implementation (I *think* I can make it smaller, faster and more readable) so I'll probably stay busy with pyRete for a couple of months more.

Inga kommentarer: