- a way to search... things
- way to search data in terms of naturla language and mocuh more...
- distributed JSON API
- fancy cluster
- software library providing full=text search
- elasticsearch a wrapper around lucene
- Data is stored in an index
- each index can store multiple types, similar to SQL
- specifying attributes fora type is optional
- all data is sentas JSON
- connect to server
- PUT data (json format)
- GET data (query in json format)
- Analyisis is the process whereby words are transformed into tokens.
- [examples]
- 9 or 10 tokenizers
- write your own
- ie) snowball, NGram, Stop Word
- [filters]
- speciy something is more important than others based on score
- boost is a multiplier on the score
- have to tweak constants for accuracy
- report counts of common terms in documents
- ie) left side rail of linkedin & amazon
- Yes it has clustering and it just works
- queries run across all shards of cluster
- no built in security
- proxing nginx in fron works for security
- easy
- just need java
- .deb package available