how structure information searchability

i am minute hunt concentration personally song playlists.

the genre record format differs playlist playlist, infrequently within playlist there differences too. there also visualisation "synonymous" tags (e.g. county cover both hiphop r&b, nonetheless proceed around).

below list hunt terms approaching results.

gospel: should relapse playlists during slightest gospel song. playlists gospel songs shown first.
urban: should relapse r&b hiphop. again playlists county outlines first.
hiphop: should relapse hiphop nonetheless r&b.
flac: should relapse playlists enclose flac files. starting ones pristine flac.
hiphop flac: should relapse hiphop flacs first, followed hiphop audio
hiphop flac: should relapse hiphop flacs only
hiphop audio: should relapse hiphop flacs, hiphop mp3s, etc

as i'm only starting project, i'm pondering best proceed index this. fulltext hunt thing lucene any here? note i don't have any calm describing playlists, nonetheless i beget some.

i'm pondering organising terms "tags" storing db many-to-many.

table: playlist ( pk(id), desc )
table: add-on ( pk(id), desc )
table: playlist_has_tag ( pk(link_id, tag_id) )

to solve county == hiphop || rnb thing, i maybe supplement tag_synonyms table:

table: tag_synonyms ( pk(tag_id, synonym_tag_id) )

then i'd have twin annals infer county encompasses hiphop rnb:
urban's add-on id, hiphop's add-on id
urban's add-on id, rnb's add-on id

i'm feeling nonetheless query definitely concerned controlling approach.

could couchdb here? i'm now controlling postgresql. there program out there kind thing easy?

i means review down support challenging hunt terms destiny like:

(hiphop house) filetype:mp3 artwork:no

and also incorporate things duration, etc.