Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[Fork tracking] Chinese Segmenter enhancements #253

Draft
wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

[Fork tracking] Chinese Segmenter enhancements #253

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,3 +9,4 @@ charabia/target
/data.ms
Cargo.lock
.idea
.hypothesis
20 changes: 9 additions & 11 deletions charabia/Cargo.toml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,23 +24,21 @@ once_cell = "1.17.1"
serde = "1.0"
slice-group-by = "0.3.0"
whatlang = "0.16.2"
lindera-core = "=0.27.1"
lindera-dictionary = "=0.27.1"
lindera-tokenizer = { version = "=0.27.1", default-features = false, optional = true }
pinyin = { version = "0.9", default-features = false, features = [
"with_tone",
], optional = true }
lindera-core = "=0.27.2"
lindera-dictionary = "=0.27.2"
lindera-tokenizer = { version = "=0.27.2", default-features = false, optional = true }
character_converter = { version = "2.1.0", optional = true }
wana_kana = { version = "3.0.0", optional = true }
unicode-normalization = "0.1.22"
irg-kvariants = "0.1.0"
litemap = "0.6.1"
zerovec = "0.9.3"
litemap = "0.7.2"
zerovec = "0.10.1"

[features]
default = ["chinese", "hebrew", "japanese", "thai", "korean", "greek", "latin-camelcase", "latin-snakecase", "khmer"]

# allow chinese specialized tokenization
chinese = ["dep:pinyin", "dep:jieba-rs"]
chinese = ["dep:character_converter", "dep:jieba-rs"]

# allow hebrew specialized tokenization
hebrew = []
Expand Down Expand Up @@ -69,8 +67,8 @@ khmer = []
latin-snakecase = ["dep:finl_unicode"]

[dev-dependencies]
criterion = "0.3"
jemallocator = "0.3.0"
criterion = "0.5.1"
jemallocator = "0.5.4"
quickcheck = "1"
quickcheck_macros = "1"

Expand Down
65 changes: 19 additions & 46 deletions charabia/src/normalizer/chinese.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
use pinyin::ToPinyin;
use character_converter::traditional_to_simplified;

use super::CharNormalizer;
use crate::detection::{Language, Script};
Expand All @@ -23,14 +23,15 @@ impl CharNormalizer for ChineseNormalizer {
// Normalize to Pinyin
// If we don't manage to convert the kvariant, we try to convert the original character.
// If none of them are converted, we return the kvariant.
match kvariant.to_pinyin().or_else(|| c.to_pinyin()) {
Some(converted) => {
let with_tone = converted.with_tone();
Some(traditional_to_simplified(kvariant.to_string().as_str()).to_string().into())
// match kvariant.to_pinyin().or_else(|| c.to_pinyin()) {
// Some(converted) => {
// let with_tone = converted.with_tone();

Some(with_tone.to_string().into())
}
None => Some(kvariant.into()), // e.g. 杤
}
// Some(with_tone.to_string().into())
// }
// None => Some(kvariant.into()), // e.g. 杤
// }
}

fn should_normalize(&self, token: &Token) -> bool {
Expand Down Expand Up @@ -65,14 +66,6 @@ mod test {
language: Some(Language::Cmn),
..Default::default()
},
Token {
lemma: Owned("澚䀾亚㮺刄杤".to_string()),
char_end: 5,
byte_end: 15,
script: Script::Cj,
language: Some(Language::Cmn),
..Default::default()
},
]
}

Expand All @@ -81,70 +74,50 @@ mod test {
vec![
Token {
// lowercased
lemma: Owned("zūnyán".to_string()),
lemma: Owned("尊严".to_string()),
char_end: 2,
byte_end: 6,
char_map: Some(vec![(3, 4), (3, 4)]),
char_map: Some(vec![(3, 3), (3, 3)]),
script: Script::Cj,
language: Some(Language::Cmn),
..Default::default()
},
Token {
// lowercased
lemma: Owned("shēngérzìyóu".to_string()),
lemma: Owned("生而自由".to_string()),
char_end: 4,
byte_end: 12,
char_map: Some(vec![(3, 6), (3, 3), (3, 3), (3, 4)]),
script: Script::Cj,
language: Some(Language::Cmn),
..Default::default()
},
Token {
// It would be "yudǔyàběnrèn" without the kvariant normalization.
lemma: Owned("àoqìyàběnrènwàn".to_string()),
char_end: 5,
byte_end: 15,
char_map: Some(vec![(3, 3), (3, 3), (3, 3), (3, 4), (3, 4), (3, 4)]),
char_map: Some(vec![(3, 3), (3, 3), (3, 3), (3, 3)]),
script: Script::Cj,
language: Some(Language::Cmn),
..Default::default()
},
}
]
}

// expected result of the complete Normalizer pieline.
fn normalized_tokens() -> Vec<Token<'static>> {
vec![
Token {
lemma: Owned("zūnyán".to_string()),
lemma: Owned("尊严".to_string()),
char_end: 2,
byte_end: 6,
char_map: Some(vec![(3, 4), (3, 4)]),
char_map: Some(vec![(3, 3), (3, 3)]),
script: Script::Cj,
language: Some(Language::Cmn),
kind: TokenKind::Word,
..Default::default()
},
Token {
lemma: Owned("shēngérzìyóu".to_string()),
lemma: Owned("生而自由".to_string()),
char_end: 4,
byte_end: 12,
char_map: Some(vec![(3, 6), (3, 3), (3, 3), (3, 4)]),
char_map: Some(vec![(3, 3), (3, 3), (3, 3), (3, 3)]),
script: Script::Cj,
language: Some(Language::Cmn),
kind: TokenKind::Word,
..Default::default()
},
Token {
lemma: Owned("àoqìyàběnrènwàn".to_string()),
char_end: 5,
byte_end: 15,
char_map: Some(vec![(3, 3), (3, 3), (3, 3), (3, 4), (3, 4), (3, 4)]),
script: Script::Cj,
language: Some(Language::Cmn),
kind: TokenKind::Word,
..Default::default()
},
}
]
}

Expand Down
12 changes: 6 additions & 6 deletions charabia/src/normalizer/control_char.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -106,16 +106,16 @@ mod test {
fn normalized_tokens() -> Vec<Token<'static>> {
vec![
Token {
lemma: Owned("shēngérzìyóuoo".to_string()),
lemma: Owned("生而自由oo".to_string()),
char_end: 9,
byte_end: 17,
script: Script::Cj,
char_map: Some(vec![
(1, 0),
(3, 6),
(3, 3),
(3, 3),
(3, 4),
(3, 3),
(3, 3),
(1, 0),
(1, 1),
(1, 1),
Expand All @@ -125,16 +125,16 @@ mod test {
..Default::default()
},
Token {
lemma: Owned("shēngérzìyóuoo".to_string()),
lemma: Owned("生而自由oo".to_string()),
char_end: 9,
byte_end: 17,
script: Script::Cj,
char_map: Some(vec![
(1, 0),
(3, 6),
(3, 3),
(3, 3),
(3, 4),
(3, 3),
(3, 3),
(1, 0),
(1, 1),
(1, 1),
Expand Down
93 changes: 64 additions & 29 deletions charabia/src/segmenter/chinese.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,8 @@
use jieba_rs::Jieba;
use once_cell::sync::Lazy;
use std::fs::File;
use std::io::{self, BufRead};
use std::path::Path;

use crate::segmenter::Segmenter;

Expand All @@ -17,7 +20,39 @@ impl Segmenter for ChineseSegmenter {
}
}

static JIEBA: Lazy<Jieba> = Lazy::new(Jieba::new);
fn read_lines<P>(filename: P) -> Vec<String>
where
P: AsRef<Path>,
{
let path = filename.as_ref();
if !path.exists() {
println!("****");
return vec![];
}

if let Ok(file) = File::open(&path) {
let reader = io::BufReader::new(file);
let mut lines = Vec::new();

for line in reader.lines() {
if let Ok(line) = line {
lines.push(line);
}
}

return lines;
}
return vec![];
}

static JIEBA: Lazy<Jieba> = Lazy::new(|| {
let mut jieba = Jieba::new();
let lines = read_lines("./words.txt");
for line in lines {
jieba.add_word(line.as_str(), Some(99 as usize), None);
}
jieba
});

#[cfg(test)]
mod test {
Expand Down Expand Up @@ -65,37 +100,37 @@ mod test {

// Segmented and normalized version of the text.
const TOKENIZED: &[&str] = &[
"rénrén",
"shēngérzìyóu",
"人人",
"生而自由",
",",
"zài",
"zūn",
"yán",
"hé",
"quán",
"lì",
"shàng",
"yīlǜpíngděng",
"在",
"尊",
"严",
"和",
"权",
"利",
"上",
"一律平等",
"。",
"tā",
"men",
"fù",
"yǒu",
"lǐxìng",
"hé",
"liángxīn",
"他",
"们",
"赋",
"有",
"理性",
"和",
"良心",
",",
"bìng",
"yīng",
"yǐ",
"xiōngdì",
"guān",
"xì",
"de",
"jīngshén",
"hùxiāng",
"duì",
"dài",
"并",
"应",
"以",
"兄弟",
"关",
"系",
"的",
"精神",
"互相",
"对",
"待",
"。",
];

Expand Down
Loading