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

Mutualize char normalizers #309

Draft
wants to merge 2 commits into
base: main
Choose a base branch
from
Draft

Mutualize char normalizers #309

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
9eea14f
Mutualize char normalizers
ManyTheFish Sep 23, 2024
f8d8308
Use binary search
ManyTheFish Sep 24, 2024
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
263 changes: 263 additions & 0 deletions charabia/src/normalizer/character_converter.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,263 @@
use super::{CharNormalizer, CharOrStr};
use crate::{Script, Token};

const CHAR_PAIRS: &[(char, Option<(char, Option<char>)>)] = &[
('Æ', Some(('a', Some('e')))),
#[cfg(feature = "vietnamese")]
('Ð', Some(('d', None))),
('æ', Some(('a', Some('e')))),
#[cfg(feature = "vietnamese")]
('ð', Some(('d', None))),
#[cfg(feature = "vietnamese")]
('Đ', Some(('d', None))),
#[cfg(feature = "vietnamese")]
('đ', Some(('d', None))),
#[cfg(feature = "turkish")]
('ı', Some(('i', None))),
('Œ', Some(('o', Some('e')))),
('œ', Some(('o', Some('e')))),
('ة', Some(('ه', None))),
('ـ', None),
('ٱ', Some(('ا', None))),
('ى', Some(('ي', None))),
('‘', Some(('\'', None))),
('’', Some(('\'', None))),
('‛', Some(('\'', None))),
];

/// This module contains the implementation of the `CharacterConverterNormalizer` struct, which is a character normalizer

pub struct CharacterConverterNormalizer;

// All normalizers only need to implement the method `normalize_char` and the method `should_normalize` of the `CharNormalizer` trait.
impl CharNormalizer for CharacterConverterNormalizer {
// Creates the normalized version of the provided char.
fn normalize_char(&self, c: char) -> Option<CharOrStr> {
let mut normalized = c.to_lowercase();

// if the original character is converted in exactly one character,
// then we return the character directly instead of creating a string for it.
match (normalized.next(), normalized.next()) {
(Some(c), None) => normalize_char(c),
(Some(first), Some(second)) => {
let first = normalize_char(first);
let second = normalize_char(second);
match (first, second) {
(Some(first), Some(second)) => Some(first.merge(&second)),
(Some(first), None) => Some(first),
(None, Some(second)) => Some(second),
(None, None) => None,
}
}
(None, _) => None,
}
}

// Returns `true` if the Normalizer should be used.
fn should_normalize(&self, token: &Token) -> bool {
token
.lemma
.chars()
.any(|c| c.is_uppercase() || CHAR_PAIRS.binary_search_by(|(k, _)| k.cmp(&c)).is_ok())
}
}

fn normalize_char(c: char) -> Option<CharOrStr> {
match CHAR_PAIRS.binary_search_by(|(k, _)| k.cmp(&c)).map(|i| &CHAR_PAIRS[i].1) {
Ok(Some((first, Some(second)))) => {
Some(CharOrStr::Char(*first).merge(&CharOrStr::Char(*second)))
}
Ok(Some((first, None))) => Some(CharOrStr::Char(*first)),
Ok(None) => None,
_ => Some(c.into()),
}

// match c {
// 'œ' | 'Œ' => Some("oe".to_string().into()),
// 'æ' | 'Æ' => Some("ae".to_string().into()),
// 'ـ' => None,
// 'ٱ' => Some('ا'.into()),
// 'ى' => Some('ي'.into()),
// 'ة' => Some('ه'.into()),
// '’' | '‘' | '‛' => Some('\''.into()),
// #[cfg(feature = "turkish")]
// 'ı' => Some('i'.into()),
// #[cfg(feature = "vietnamese")]
// 'Ð' | 'Đ' | 'đ' | 'ð' => Some("d".to_string().into()),
// _ => Some(c.into()),
// }
}

// Test the normalizer:
#[cfg(test)]
mod test {
use std::borrow::Cow::Owned;

use crate::normalizer::test::test_normalizer;
use crate::normalizer::{Normalizer, NormalizerOption};
use crate::token::TokenKind;

// base tokens to normalize.
fn tokens() -> Vec<Token<'static>> {
vec![
Token {
lemma: Owned("œ".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
..Default::default()
},
Token {
lemma: Owned("Œ".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
..Default::default()
},
Token {
lemma: Owned("æ".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
..Default::default()
},
Token {
lemma: Owned("Æ".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
..Default::default()
},
// Taa Marbuta
Token {
lemma: Owned("النهاردة".to_string()),
char_end: 8,
byte_end: 16,
script: Script::Arabic,
..Default::default()
},
]
}

// expected result of the current Normalizer.
fn normalizer_result() -> Vec<Token<'static>> {
vec![
Token {
lemma: Owned("oe".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
char_map: Some(vec![(2, 2)]),
..Default::default()
},
Token {
lemma: Owned("oe".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
char_map: Some(vec![(2, 2)]),
..Default::default()
},
Token {
lemma: Owned("ae".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
char_map: Some(vec![(2, 2)]),
..Default::default()
},
Token {
lemma: Owned("ae".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
char_map: Some(vec![(2, 2)]),
..Default::default()
},
Token {
lemma: Owned("النهارده".to_string()),
char_end: 8,
byte_end: 16,
char_map: Some(vec![
(2, 2),
(2, 2),
(2, 2),
(2, 2),
(2, 2),
(2, 2),
(2, 2),
(2, 2),
]),
script: Script::Arabic,
..Default::default()
},
]
}

// expected result of the complete Normalizer pieline.
fn normalized_tokens() -> Vec<Token<'static>> {
vec![
Token {
lemma: Owned("oe".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
char_map: Some(vec![(2, 2)]),
kind: TokenKind::Word,
..Default::default()
},
Token {
lemma: Owned("oe".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
char_map: Some(vec![(2, 2)]),
kind: TokenKind::Word,
..Default::default()
},
Token {
lemma: Owned("ae".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
char_map: Some(vec![(2, 2)]),
kind: TokenKind::Word,
..Default::default()
},
Token {
lemma: Owned("ae".to_string()),
char_end: 2,
byte_end: 2,
script: Script::Latin,
char_map: Some(vec![(2, 2)]),
kind: TokenKind::Word,
..Default::default()
},
Token {
lemma: Owned("النهارده".to_string()),
char_end: 8,
byte_end: 16,
char_map: Some(vec![
(2, 2),
(2, 2),
(2, 2),
(2, 2),
(2, 2),
(2, 2),
(2, 2),
(2, 2),
]),
script: Script::Arabic,
kind: TokenKind::Word,
..Default::default()
},
]
}

test_normalizer!(
CharacterConverterNormalizer,
tokens(),
normalizer_result(),
normalized_tokens()
);
}
34 changes: 26 additions & 8 deletions charabia/src/normalizer/mod.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@ use self::swedish_recomposition::SwedishRecompositionNormalizer;
pub use self::turkish::TurkishNormalizer;
#[cfg(feature = "vietnamese")]
pub use self::vietnamese::VietnameseNormalizer;
use crate::normalizer::character_converter::CharacterConverterNormalizer;
use crate::segmenter::SegmentedTokenIter;
use crate::Token;

Expand All @@ -47,6 +48,7 @@ mod turkish;
mod vietnamese;

mod ae_oe_normalizer;
mod character_converter;

/// List of [`Normalizer`]s used by [`Normalize::normalize`] that are not considered lossy.
pub static NORMALIZERS: Lazy<Vec<Box<dyn Normalizer>>> = Lazy::new(|| {
Expand All @@ -62,21 +64,22 @@ pub static NORMALIZERS: Lazy<Vec<Box<dyn Normalizer>>> = Lazy::new(|| {
/// List of [`Normalizer`]s used by [`Normalize::normalize`] that are considered lossy.
pub static LOSSY_NORMALIZERS: Lazy<Vec<Box<dyn Normalizer>>> = Lazy::new(|| {
vec![
Box::new(LowercaseNormalizer),
Box::new(QuoteNormalizer),
Box::new(AeOeNormalizer),
// Box::new(LowercaseNormalizer),
// Box::new(QuoteNormalizer),
// Box::new(AeOeNormalizer),
Box::new(CharacterConverterNormalizer),
#[cfg(feature = "chinese-normalization")]
Box::new(ChineseNormalizer),
#[cfg(feature = "japanese-transliteration")]
Box::new(JapaneseNormalizer),
#[cfg(feature = "greek")]
Box::new(GreekNormalizer),
Box::new(ArabicNormalizer),
// Box::new(ArabicNormalizer),
Box::new(NonspacingMarkNormalizer),
#[cfg(feature = "vietnamese")]
Box::new(VietnameseNormalizer),
#[cfg(feature = "turkish")]
Box::new(TurkishNormalizer),
// #[cfg(feature = "vietnamese")]
// Box::new(VietnameseNormalizer),
// #[cfg(feature = "turkish")]
// Box::new(TurkishNormalizer),
]
});

Expand Down Expand Up @@ -226,6 +229,21 @@ pub enum CharOrStr {
Str(String),
}

impl CharOrStr {
pub fn merge(&self, other: &Self) -> Self {
let mut result = String::new();
match self {
Self::Char(c) => result.push(*c),
Self::Str(s) => result.push_str(s),
}
match other {
Self::Char(c) => result.push(*c),
Self::Str(s) => result.push_str(s),
}
Self::Str(result)
}
}

impl From<char> for CharOrStr {
fn from(c: char) -> Self {
Self::Char(c)
Expand Down
Loading