The Mid-term is over. I am through the first half of the Google summer of Code! As far as the accomplishments are concerned, I have almost implemented the culmination of 3 papers in python and have kept on updating my pull request here: pull 420 .
My most efforts went on getting all the algorithm work so I was not able to work on finesse points. Abinash is not happy with the current code. There are many loop-holes and my coding style isn't good. Looks like the way I have spent my time understanding the algorithm from the implemented code isn't right. I needed to get a grip of the algorithm.
So I have spent a lot of time again on the theory. This time, I made meticulous notes. For notes check this out: http://kislaynotes.droppages.com/
My expectations for the next period is to completely rework the current implementation. This code can be thought more like a sample to understand what lies ahead.
One thing got clear. I need to get more object oriented. The current method that I was using was more of porting the C++ code to python disregarding the pgmpy's inherent classes.I now see, how sheepishly I read the already implemented exact inference algorithms in the library. I need to get more pythonic. My current code looks more like a C++ code which is very bad.
Anyways, so my focus now is to spend time with the already implemented library. See how the 1st paper fits in. Discuss a lot with the mentor. I have lost some precious time, but now since I know some of the things where I was doing wrong ... I hope to proceed swiftly
Bye! I hope to get some cool python coding done in the next few days!