mocobeta

# check
lucene $ ./gradlew check
BUILD SUCCESSFUL in 3m 59s

# packaging
lucene $ ./gradlew clean
lucene $ ./gradlew assembleRelease
BUILD SUCCESSFUL in 49s

# luke

mocobeta

@State(Scope.Benchmark)
public class SearchBenchmark {

  private static final String dirPath = System.getProperty("index.dir");
  private static final String[] terms1 = new String[]{"電車", "列車", "鉄道"};

  private Directory dir;
  private IndexReader reader;

  private Query query1;

mocobeta

use rand::prelude::*;

fn main() {
    let mut rng = thread_rng();
    let p: f32 = 0.00001;
    let max_doc: usize = 1_000_000;
    let mut postings: Vec<usize> = vec![rng.gen_range(1, 1000) as usize];
    loop {
        let next = postings.last().unwrap() + geo_random(p);
        if next > max_doc {

mocobeta

pub fn encode_vbyte(li: &[usize]) -> Vec<u8> {
    fn encode(k: usize) -> Vec<u8> {
        let mut vbytes = Vec::new();
        let mut tmp = k;
        while tmp >= 128 {
            vbytes.push(128 + (tmp & 127) as u8);
            tmp >>= 7;
        }
        vbytes.push(tmp as u8);
        vbytes

mocobeta

// https://crates.io/crates/bit-vec
use bit_vec::BitVec;

pub fn encode_rice(li: &[usize], m: u32) -> BitVec {
    fn encode_quotient(k: usize, m: u32) -> BitVec {
        let q: usize = (((k - 1) / m as usize) as f64).floor() as usize;
        // encode (quotient + 1) in unary code
        let mut bv = BitVec::from_elem(q + 1, false);
        bv.set(q, true);
        bv

mocobeta

// https://crates.io/crates/bit-vec
use bit_vec::BitVec;

pub fn encode_golomb(li: &[usize], m: u32) -> BitVec {
    fn encode_quotient(k: usize, m: u32) -> BitVec {
        let q: usize = (((k - 1) / m as usize) as f64).floor() as usize;
        // encode (quotient + 1) in unary code
        let mut bv = BitVec::from_elem(q + 1, false);
        bv.set(q, true);
        bv

mocobeta

// https://crates.io/crates/bit-vec
use bit_vec::BitVec;

pub fn encode_gamma(li: &[usize]) -> BitVec {
    fn encode(k: usize) -> BitVec {
        let body_len: usize = ((k as f64).log2().floor() as usize) + 1;
        let body = BitVec::from_bytes(&k.to_be_bytes());

        let mut bv = BitVec::from_elem(body_len * 2 - 1, false);
        // set selector bit

mocobeta

// https://crates.io/crates/bit-vec
use bit_vec::BitVec;

pub fn encode_delta(li: &[usize]) -> BitVec {
    fn encode(k: usize) -> BitVec {
        let body_len: usize = ((k as f64).log2().floor() as usize) + 1;
        let body = BitVec::from_bytes(&k.to_be_bytes());

        // set gamma encoded selector
        let mut bv = encode_gamma(&[body_len]);

mocobeta

FROM docker.elastic.co/elasticsearch/elasticsearch:7.7.0

ENV PATH /usr/share/elasticsearch/bin:$PATH

# switch user to elasticsearch
USER elasticsearch

# install plugins
RUN elasticsearch-plugin install analysis-kuromoji

mocobeta

# base image
FROM docker.elastic.co/elasticsearch/elasticsearch:7.7.0

# PATH
ENV PATH /usr/share/elasticsearch/bin:$PATH

USER elasticsearch

# copy configuration file
COPY elasticsearch.yml /usr/share/elasticsearch/config/