We have successfully abstracted datastructure of Bayes which allowed us to implement an alternate Redis storage backend (#81) while making it possible to easily add more backends (such as ORM based). However, I found it little difficult to follow along the flow of the LSI implementation to understand all the datastructures needed for that. Can someone give a high-level overview of datastructures of LSI, their relationship, and desired operations?
As a side note, can we please make sure to abstract the datascructure away from the logic from the day one of every new algorithm we might implement as indicated in #88.
The LSI functionality/implementation is a pretty tricky to follow. Most of it hasn't been changed since it was first written. I'm not sure I could adequately explain how to go about abstracting the pieces.
We have successfully abstracted datastructure of Bayes which allowed us to implement an alternate Redis storage backend (#81) while making it possible to easily add more backends (such as ORM based). However, I found it little difficult to follow along the flow of the LSI implementation to understand all the datastructures needed for that. Can someone give a high-level overview of datastructures of LSI, their relationship, and desired operations?
As a side note, can we please make sure to abstract the datascructure away from the logic from the day one of every new algorithm we might implement as indicated in #88.