Optimize repository index page

[olenagerasimova]

Pypi repo has index page with all repo packages listed in text html format along with their sha256 sum. This page is generated by SliceIndex on each request. Such approach may work fine when repository is not large, but on big repo generating this page can become a very time-consuming operation. We should think of creating and supporting some repository index file with artifacts names and sha sums in order not calculate them on each request.