searchengines-web/MeiliSearch
searchengines-web/MeiliSearch
1
0
Fork 0
mirror of https://github.com/meilisearch/MeiliSearch synced 2025-07-03 20:07:09 +02:00
Code Issues Releases Wiki Activity

fix: Remove stop-words from the serve examples

Browse source
This commit is contained in:
Clément Renault 2018-10-21 16:42:19 +02:00
parent cf41b20fbb
commit 26dcfe1e54
6 changed files with 56 additions and 41 deletions
Download patch file Download diff file Expand all files Collapse all files

18
examples/serve-console.rs
View file

@ -10,25 +10,18 @@ use pentium::{automaton, DocumentId, Metadata};
#[derive(Debug, StructOpt)]
pub struct CommandConsole {
/// The stop word file, each word must be separated by a newline.
#[structopt(long = "stop-words", parse(from_os_str))]
pub stop_words: PathBuf,
/// Meta file name (e.g. relaxed-colden).
#[structopt(parse(from_os_str))]
pub meta_name: PathBuf,
}
pub struct ConsoleSearch {
common_words: CommonWords,
metadata: Metadata,
db: DB,
}
impl ConsoleSearch {
pub fn from_command(command: CommandConsole) -> io::Result<ConsoleSearch> {
let common_words = CommonWords::from_file(command.stop_words)?;
let map_file = command.meta_name.with_extension("map");
let idx_file = command.meta_name.with_extension("idx");
let sst_file = command.meta_name.with_extension("sst");
@ -42,7 +35,7 @@ impl ConsoleSearch {
drop(db);
let db = DB::open_for_read_only(DBOptions::default(), rocksdb, false).unwrap();
Ok(ConsoleSearch { common_words, metadata, db })
Ok(ConsoleSearch { metadata, db })
}
pub fn serve(self) {
@ -52,20 +45,19 @@ impl ConsoleSearch {
let mut query = String::new();
io::stdin().read_line(&mut query).unwrap();
let query = query.trim().to_lowercase();
if query.is_empty() { break }
let (elapsed, _) = measure_time(|| search(&self.metadata, &self.db, &self.common_words, &query));
let (elapsed, _) = measure_time(|| search(&self.metadata, &self.db, &query));
println!("Finished in {}", elapsed);
}
}
}
fn search(metadata: &Metadata, database: &DB, common_words: &CommonWords, query: &str) {
fn search(metadata: &Metadata, database: &DB, query: &str) {
let mut automatons = Vec::new();
for query in query.split_whitespace().filter(|q| !common_words.contains(*q)) {
let lev = automaton::build(query);
for query in query.split_whitespace().map(str::to_lowercase) {
let lev = automaton::build_prefix_dfa(&query);
automatons.push(lev);
}

18
examples/serve-http.rs
View file

@ -19,10 +19,6 @@ pub struct CommandHttp {
#[structopt(short = "l", default_value = "127.0.0.1:3030")]
pub listen_addr: SocketAddr,
/// The stop word file, each word must be separated by a newline.
#[structopt(long = "stop-words", parse(from_os_str))]
pub stop_words: PathBuf,
/// Meta file name (e.g. relaxed-colden).
#[structopt(parse(from_os_str))]
pub meta_name: PathBuf,
@ -41,15 +37,12 @@ struct SearchQuery { q: String }
pub struct HttpServer {
listen_addr: SocketAddr,
common_words: Arc<CommonWords>,
metadata: Arc<Metadata>,
db: Arc<DB>,
}
impl HttpServer {
pub fn from_command(command: CommandHttp) -> io::Result<HttpServer> {
let common_words = CommonWords::from_file(command.stop_words)?;
let map_file = command.meta_name.with_extension("map");
let idx_file = command.meta_name.with_extension("idx");
let sst_file = command.meta_name.with_extension("sst");
@ -64,19 +57,18 @@ impl HttpServer {
Ok(HttpServer {
listen_addr: command.listen_addr,
common_words: Arc::new(common_words),
metadata: Arc::new(metadata),
db: Arc::new(db),
})
}
pub fn serve(self) {
let HttpServer { listen_addr, common_words, metadata, db } = self;
let HttpServer { listen_addr, metadata, db } = self;
let routes = warp::path("search")
.and(warp::query())
.map(move |query: SearchQuery| {
let body = search(metadata.clone(), db.clone(), common_words.clone(), &query.q).unwrap();
let body = search(metadata.clone(), db.clone(), &query.q).unwrap();
body
})
.with(warp::reply::with::header("Content-Type", "application/json"))
@ -86,15 +78,13 @@ impl HttpServer {
}
}
fn search<M, D, C>(metadata: M, database: D, common_words: C, query: &str) -> Result<String, Box<Error>>
fn search<M, D>(metadata: M, database: D, query: &str) -> Result<String, Box<Error>>
where M: AsRef<Metadata>,
D: AsRef<DB>,
C: AsRef<CommonWords>,
{
let mut automatons = Vec::new();
for query in query.split_whitespace().map(str::to_lowercase) {
if common_words.as_ref().contains(&query) { continue }
let lev = automaton::build(&query);
let lev = automaton::build_prefix_dfa(&query);
automatons.push(lev);
}