Simplify search while improving performance

This commit is contained in:
Joakim Soderlund 2024-04-06 22:58:54 +02:00
parent d440aeab9e
commit 4ef6bc533d
5 changed files with 60 additions and 188 deletions

101
Cargo.lock generated
View file

@ -44,15 +44,6 @@ version = "1.0.96"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6b964d184e89d9b6b67dd2715bc8e74cf3107fb2b529990c90cf517326150bf4"
[[package]]
name = "arbitrary"
version = "1.4.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "dde20b3d026af13f561bdd0f15edf01fc734f0dafcedbaf42bba506a9517f223"
dependencies = [
"derive_arbitrary",
]
[[package]]
name = "arc-swap"
version = "1.7.1"
@ -299,17 +290,6 @@ dependencies = [
"serde",
]
[[package]]
name = "derive_arbitrary"
version = "1.4.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "30542c1ad912e0e3d22a1935c290e12e8a29d704a420177a31faad4a601a0800"
dependencies = [
"proc-macro2",
"quote",
"syn",
]
[[package]]
name = "derive_more"
version = "1.0.0"
@ -330,17 +310,6 @@ dependencies = [
"syn",
]
[[package]]
name = "displaydoc"
version = "0.2.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "97369cbbc041bc366949bc74d34658d6cda5621039731c6310521892a3a20ae0"
dependencies = [
"proc-macro2",
"quote",
"syn",
]
[[package]]
name = "downcast-rs"
version = "2.0.1"
@ -396,7 +365,7 @@ dependencies = [
"rayon",
"serde",
"serde_json",
"zip 0.6.6",
"zip",
]
[[package]]
@ -406,7 +375,6 @@ dependencies = [
"fimfareader",
"fimfareader-search",
"rustyline",
"zip 2.2.2",
]
[[package]]
@ -427,10 +395,8 @@ name = "fimfareader-search"
version = "0.1.0"
dependencies = [
"fimfareader",
"rayon",
"tantivy",
"thread_local",
"zip 2.2.2",
"zip",
]
[[package]]
@ -586,16 +552,6 @@ version = "1.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "b9e0384b61958566e926dc50660321d12159025e767c18e043daf26b70104c39"
[[package]]
name = "indexmap"
version = "2.7.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8c9c992b02b5b4c94ea26e32fe5bccb7aa7d9f390ab5c1221ff895bc7ea8b652"
dependencies = [
"equivalent",
"hashbrown",
]
[[package]]
name = "itertools"
version = "0.14.0"
@ -660,12 +616,6 @@ version = "0.4.15"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d26c52dbd32dccf2d10cac7725f8eae5296885fb5703b261f7d0a0739ec807ab"
[[package]]
name = "lockfree-object-pool"
version = "0.1.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9374ef4228402d4b7e403e5838cb880d9ee663314b0a900d5a6aabf0c213552e"
[[package]]
name = "log"
version = "0.4.25"
@ -1038,12 +988,6 @@ version = "1.3.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64"
[[package]]
name = "simd-adler32"
version = "0.3.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d66dc143e6b11c1eddc06d5c423cfc97062865baf299914ab64caa38182078fe"
[[package]]
name = "sketches-ddsketch"
version = "0.3.0"
@ -1261,16 +1205,6 @@ dependencies = [
"syn",
]
[[package]]
name = "thread_local"
version = "1.1.8"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8b9ef9bad013ada3808854ceac7b46812a6465ba368859a37e2100283d2d719c"
dependencies = [
"cfg-if",
"once_cell",
]
[[package]]
name = "time"
version = "0.3.37"
@ -1570,37 +1504,6 @@ dependencies = [
"flate2",
]
[[package]]
name = "zip"
version = "2.2.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "ae9c1ea7b3a5e1f4b922ff856a129881167511563dc219869afe3787fc0c1a45"
dependencies = [
"arbitrary",
"crc32fast",
"crossbeam-utils",
"displaydoc",
"flate2",
"indexmap",
"memchr",
"thiserror",
"zopfli",
]
[[package]]
name = "zopfli"
version = "0.8.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e5019f391bac5cf252e93bbcc53d039ffd62c7bfb7c150414d61369afe57e946"
dependencies = [
"bumpalo",
"crc32fast",
"lockfree-object-pool",
"log",
"once_cell",
"simd-adler32",
]
[[package]]
name = "zstd"
version = "0.13.2"

View file

@ -13,8 +13,3 @@ path = "../search"
[dependencies.rustyline]
default-features = false
version = "14"
[dependencies.zip]
version = "*"
features = ["deflate"]
default-features = false

View file

@ -2,13 +2,10 @@
use std::env::args;
use std::error::Error;
use std::fs::File;
use std::io::BufReader;
use std::result::Result;
use std::time::Instant;
use rustyline::DefaultEditor;
use zip::ZipArchive;
use fimfareader::archive::Fetcher;
use fimfareader_search::Searcher;
@ -32,14 +29,7 @@ fn main() -> Result<(), Box<dyn Error>> {
println!("Finished loading in {finish:?}.");
println!("The archive contains {count} stories.");
let opener = || {
let file = File::open(&argv[1]).unwrap();
let reader = BufReader::new(file);
ZipArchive::new(reader).unwrap()
};
let searcher = Searcher::new(&fetcher, &opener);
let searcher = Searcher::new(&fetcher);
while let Ok(line) = editor.readline(">>> ") {
editor.add_history_entry(&line)?;

View file

@ -7,16 +7,10 @@ edition = "2021"
[dependencies.fimfareader]
path = ".."
[dependencies.rayon]
version = "*"
[dependencies.tantivy]
git = "https://github.com/quickwit-oss/tantivy.git"
[dependencies.thread_local]
version = "*"
[dependencies.zip]
version = "*"
version = "=0.6.6"
features = ["deflate"]
default-features = false

View file

@ -1,108 +1,90 @@
//! Main module.
use std::cell::RefCell;
use std::fs::create_dir_all;
use std::io::stdout;
use std::io::Cursor;
use std::io::Read;
use std::io::Seek;
use std::io::Write;
use std::path::Path;
use std::sync::atomic::AtomicUsize;
use std::sync::atomic::Ordering;
use std::time::Instant;
use rayon::iter::ParallelIterator;
use thread_local::ThreadLocal;
use zip::read::ZipArchive;
use tantivy::collector::TopDocs;
use tantivy::directory::MmapDirectory;
use tantivy::query::QueryParser;
use tantivy::schema;
use tantivy::schema::Schema;
use tantivy::schema::Value;
use tantivy::Index;
use tantivy::IndexSettings;
use tantivy::ReloadPolicy;
use tantivy::TantivyDocument;
use zip::read::ZipArchive;
use fimfareader::prelude::*;
pub struct Searcher {
schema: Schema,
index: Index,
}
impl Searcher {
pub fn new<T, F>(fetcher: &Fetcher<T>, f: &F) -> Self
pub fn new<T>(fetcher: &Fetcher<T>) -> Self
where
T: Read + Seek + Send,
F: Fn() -> ZipArchive<T> + Sync,
{
Searcher {
index: Self::load_index(fetcher),
}
}
fn schema() -> Schema {
let mut builder = Schema::builder();
builder.add_i64_field("sid", schema::INDEXED | schema::STORED);
builder.add_text_field("content", schema::TEXT);
let schema = builder.build();
let index = Self::load_index(schema.clone(), fetcher, f);
Searcher { schema, index }
builder.build()
}
fn load_index<T, F>(schema: Schema, fetcher: &Fetcher<T>, f: &F) -> Index
fn load_index<T>(fetcher: &Fetcher<T>) -> Index
where
T: Read + Seek + Send,
F: Fn() -> ZipArchive<T> + Sync,
{
let identity = fetcher.identity().unwrap();
let directory = Path::new("cache").join(identity);
let path = Path::new("cache").join(identity);
if !directory.exists() {
Self::make_index(schema.clone(), fetcher, f);
if path.exists() {
Index::open_in_dir(path).unwrap()
} else {
Self::make_index(&path, fetcher)
}
}
let store = MmapDirectory::open(&directory).unwrap();
Index::open_or_create(store, schema).unwrap()
}
fn make_index<T, F>(schema: Schema, fetcher: &Fetcher<T>, f: &F)
fn make_index<T>(path: &Path, fetcher: &Fetcher<T>) -> Index
where
T: Read + Seek + Send,
F: Fn() -> ZipArchive<T> + Sync,
{
let identity = fetcher.identity().unwrap();
let directory = Path::new("cache").join(identity);
let start = Instant::now();
print!("\r\rIndexing archive...\r\r");
create_dir_all(path).unwrap();
std::fs::create_dir_all(&directory).unwrap();
let store = MmapDirectory::open(&directory).unwrap();
let settings = IndexSettings::default();
let index = Index::create(store, schema, settings).unwrap();
let schema = Self::schema();
let index = Index::create_in_dir(path, schema).unwrap();
let mut writer = index.writer(1_073_741_824).unwrap();
let mut buffer = String::with_capacity(1_048_576);
let schema = index.schema();
let sid = schema.get_field("sid").unwrap();
let identifier = schema.get_field("sid").unwrap();
let content = schema.get_field("content").unwrap();
let mut writer = index.writer(536_870_912).unwrap();
let story_count = fetcher.iter().count() as f64;
let counter = AtomicUsize::new(0);
let total = fetcher.iter().count();
let start = Instant::now();
for (i, story) in fetcher.iter().enumerate() {
let progress = (i * 100) as f64 / story_count;
print!("\r\rIndexing archive... {progress:.2}%\r\r");
let local = ThreadLocal::new();
let cursor = Cursor::new(fetcher.read(story).unwrap());
let mut epub = ZipArchive::new(cursor).unwrap();
let mut document = TantivyDocument::default();
fetcher.par_iter().for_each(|story| {
let mut doc = TantivyDocument::default();
let mut arch = local.get_or(|| RefCell::new(f())).borrow_mut();
let mut file = arch.by_name(&story.archive.path).unwrap();
let mut data = Vec::with_capacity(file.size() as usize);
let mut text = String::with_capacity(1_048_576);
file.read_to_end(&mut data).unwrap();
let mut epub = ZipArchive::new(Cursor::new(data)).unwrap();
let count = counter.fetch_add(1, Ordering::SeqCst);
let percentage = (count as f64 / total as f64) * 100f64;
print!("\r\rIndexing archive... {:.2}%\r\r", percentage);
doc.add_i64(sid, story.id.into());
document.add_i64(identifier, story.id.into());
for i in 0..epub.len() {
let mut file = epub.by_index(i).unwrap();
@ -111,18 +93,24 @@ impl Searcher {
continue;
}
file.read_to_string(&mut text).unwrap();
doc.add_text(content, &text);
text.clear();
file.read_to_string(&mut buffer).unwrap();
document.add_text(content, &buffer);
buffer.clear();
}
writer.add_document(doc).unwrap();
});
writer.add_document(document).unwrap();
}
print!("\r\rCommitting archive index...\r\r");
stdout().flush().unwrap();
writer.commit().unwrap();
writer.wait_merging_threads().unwrap();
let finish = (Instant::now() - start).as_secs();
println!("Index generated in {} seconds.", finish);
println!("Index generated in {finish} seconds.");
index
}
pub fn search(&self, text: &str) -> Vec<(i64, f32)> {
@ -133,20 +121,22 @@ impl Searcher {
.try_into()
.unwrap();
let searcher = reader.searcher();
let identitfier = self.schema.get_field("sid").unwrap();
let content = self.schema.get_field("content").unwrap();
let parser = QueryParser::for_index(&self.index, vec![content]);
let schema = self.index.schema();
let identifier = schema.get_field("sid").unwrap();
let content = schema.get_field("content").unwrap();
let limit = TopDocs::with_limit(32);
let parser = QueryParser::for_index(&self.index, vec![content]);
let query = parser.parse_query(text).unwrap();
let searcher = reader.searcher();
let limit = TopDocs::with_limit(32);
let docs = searcher.search(&query, &limit).unwrap();
docs.into_iter()
.map(|(score, address)| {
let doc: TantivyDocument = searcher.doc(address).unwrap();
match doc.get_first(identitfier).map(|v| v.as_i64()) {
match doc.get_first(identifier).map(|v| v.as_i64()) {
Some(Some(value)) => (value, score),
_ => panic!("Invalid story key type!"),
}