2018-05-05 16:18:23 +08:00
|
|
|
# raptor-rs
|
|
|
|
|
Raptor, the new RISE
|
2018-09-16 03:25:04 +08:00
|
|
|
|
|
|
|
|
## Usage
|
|
|
|
|
|
|
|
|
|
First you need to generate the index files.
|
|
|
|
|
|
|
|
|
|
```bash
|
|
|
|
|
$ 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).
|
|
|
|
|
|
|
|
|
|
```bash
|
|
|
|
|
$ 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`).
|
