How would I go about creating a custom search index much like Lucene?

Question

I implemented a Lucene search solution awhile back, and it got me interested in compressed file indexes that are searchable. At the time I could not find any good information on how exactly you would go about creating a custom search index, so I wonder if anyone can point me in the right direction? My primary interest is in file formatting, compression, and something similar to the concept of Lucene's documents and fields. It should not necessarily be language specific, but if you can point me to online resources that have language specific implementations with full descriptions of the process then that is okay, too.

Answer 1

You may also try to look in the source code of excellent Sphinx search engine.

It is modern full-text open source search engine, and it uses smartly optimized indexes.