The source of small dataset 'sample'

[ZorrowHu]

Thanks for sharing this awesome project! I want to know where this small dataset 'sample' with 310 nodes comes from? Is it a subset from yoochoose or diginetica, or is just generated manually?

[ZorrowHu]

And I also want to know some setups about the experiments like, is P@20 and MRR@20 the average performance over a fixed number of runs and if so how big is that number?