Sections code for vector search

[Thejas-bhat]

• Introduces a new section called faissVectorIndexSection which maintains vector search related operations over a specific field - indexing a bunch of vectors and performing the similarity search on them.
• The section implements the section interface from section.go which basically ensures that the vector indexing related stuff that happens during the segment's lifecycle is agnostic of any other section that's present (for eg inverted index for textual data).
• The section itself uses the https://github.com/blevesearch/faiss (a fork from the facebookreasearch's faiss library) library to do the actual work of building the vector index and performing search over it.
• Currently, two types of index - flat or IVFSQ is built and maintained based on the volume of vectors present for a field within a segment.
• Currently during search time, the underlying faiss index is consulted to get the similar vectors, constructs a vector postings list which is iterable and each element within it is a pair of docNum in which the similar vector is present it and the corresponding vector's score.
• Caution: Still under active development.

[abhinavdangeti]

Looking forward to a lot of commentary here :)