Go to file
2018-09-21 22:37:43 +02:00
raptor fix: Update the typos sort function to take the distance correctly 2018-09-21 22:25:59 +02:00
raptor-http feat: Change the query name to 'q' 2018-09-21 22:37:43 +02:00
raptor-indexer fix: Prefer using the group_id instead of the product_id 2018-09-15 21:03:18 +02:00
raptor-search doc: Add indexation and search usage instructions 2018-09-15 21:25:04 +02:00
.gitignore feat: Allow the user to specify options by argument 2018-09-20 13:01:08 +02:00
Cargo.lock feat: Allow the user to specify options by argument 2018-09-20 13:01:08 +02:00
Cargo.toml feat: Introduce a little http query server 2018-09-14 19:09:20 +02:00
LICENSE Initial commit 2018-05-05 10:16:18 +02:00
README.md doc: Add indexation and search usage instructions 2018-09-15 21:25:04 +02:00

raptor-rs

Raptor, the new RISE

Usage

First you need to generate the index files.

$ cargo build --release --bin raptor-indexer
$ time ./target/release/raptor-indexer products.json_lines

Once the command finished indexing you will have 3 files that compose the index:

  • The xxx.map represent the fst map.
  • The xxx.idx represent the doc indexes matching the words in the map.
  • The xxx.sst is a file that contains all the fields and the values asociated with it, it is passed to the internal RocksDB.

Now you can easily use raptor-search or raptor-http with only the prefix name of the files. (e.g. relaxed-colden).

$ cargo run --bin raptor-search -- relaxed-colden
$ cargo run --bin raptor-http -- relaxed-colden

Note: If you have performance issues run the searcher in release mode (i.e. --release).