Test case project to explore and demonstrate using base-128 varint encoding on a list of document id's, such as those typically found in inverted index structures. The basic test case attempts to load an index with compressed data, and then perform a query on that data for a simple (but exhaustive) criteria.
There are several test cases run from the main entry point (TextCases); see the code comments for test goals.
There is a simple test case written that demonstrates the notion of using an inverted index to save and query multi-value conditions. This is accomplished by encapsulating the doc+position, as well as distinguishing between a term name (e.g., "author") and a term value (e.g., "Bill"). These individual pairs (doc + pos, and term + value) are then passed through an FNV-32A hash to yield a familiar list of integers which can then be fully represented via an inverted index.
This work is currently supporting development of an isolated datastore leveraging these techniques to allow for nested, multi-value "tables" of data within individual documents...more documentation to come.