BS is a data structures library. It features fixed size, random access sequences. The sequences are constructed over B-trees. Complexity for most single item operations is about O(logN). The sequences are three to five times faster over random access BST-based sequences, while having better memory usage characteristics.
markov.sql implements third order (and lower) model Markov Chains, including training and generation, in pure SQL(ite). The source code archive also contains a pre-trained SQL dump with models based on the 1990 US Census Bureau data sets for first and last names, and an excerpt of the 2012 US tax office corporation names.