AI-GCS: Create indexing pipeline and REST API with index-time matching

[pombredanne]

This is to integrate the code indexing and searching/matching in ScanCode.io, PurlDB and MatchCode.io.

This means to create an indexing pipeline that is able to compute and index code for approximate fragment matching and also perform index-time matching to avoid indexing duplicates.

The outcome should be:

• Create new indexing pipeline in PurlDB for code fragments
  • Implement indexing-time matching to avoid indexing duplicates.
  • Then, make a new release of MatchCode.io/PurlDB as needed
• Create guide documenting how to use API for indexing