[pyclustering.utils] Random Sampling with a Reservoir

[annoviko]

Introduction
Algorithm for random sampling is required for CURE to work with big data.

Description
Implementation should be based on the article: 'Random Sampling with a Reservoir - JEFFREY SCOTT VITTER':

@article{Vitter:1985:RSR:3147.3165,
 author = {Vitter, Jeffrey S.},
 title = {Random Sampling with a Reservoir},
 journal = {ACM Trans. Math. Softw.},
 issue_date = {March 1985},
 volume = {11},
 number = {1},
 month = mar,
 year = {1985},
 issn = {0098-3500},
 pages = {37--57},
 numpages = {21},
 url = {http://doi.acm.org/10.1145/3147.3165},
 doi = {10.1145/3147.3165},
 acmid = {3165},
 publisher = {ACM},
 address = {New York, NY, USA},
} 

The implementation should be standalone and provide services to clients like CURE. Currently, CURE algorithm is only one client of this random samplier therefore it should provide all services that are required for CURE.