TSS Slide Approach

[rmhubley]

Is there a particular reason why this implementation of your TSS method uses a sliding approach that begins each sketch with an embedding of only stride-bp of sequence? Wouldn't it make more sense to reach the window length before emitting the first vector? In addition, it appears you don't encode the trailing sequence if it doesn't fit into a full window. Perhaps allowing the final stride to be smaller would allow for the last vector to cover the final bases with a tolerable amount of redundancy.