cessen/furigana_gen
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Use dictionary form of word when checking against known word list.

Browse Source
This commit is contained in:
Nathan Vegdahl 2025-09-04 06:17:59 +09:00
parent 40ed2fd955
commit 2cd80a3dce
1 changed files with 8 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
src/lib.rs
View File

@ -321,7 +321,7 @@ fn add_html_furigana(
let mut new_text = String::new(); let mut new_text = String::new();
for i in 0..worker.num_tokens() { for i in 0..worker.num_tokens() {
let t = worker.token(i); let t = worker.token(i);
let (surface, kana, pitches) = { let (surface, kana, word, pitches) = {
let surface = t.surface(); let surface = t.surface();
let feature = t.feature(); let feature = t.feature();
@ -343,7 +343,7 @@ fn add_html_furigana(
&[] &[]
}; };
(surface, kana, pitches) (surface, kana, word, pitches)
}; };
let needs_help = learner.process(surface); let needs_help = learner.process(surface);
@ -362,6 +362,7 @@ fn add_html_furigana(
let furigana_text = apply_furigana( let furigana_text = apply_furigana(
surface, surface,
&kana, &kana,
word,
exclude_kanji, exclude_kanji,
exclude_words, exclude_words,
pitches, pitches,
@ -397,6 +398,7 @@ fn add_html_furigana(
fn apply_furigana<'a>( fn apply_furigana<'a>(
surface: &'a str, surface: &'a str,
kana: &'a str, kana: &'a str,
dictionary_form: &'a str,
exclude_kanji: &FnvHashSet<char>, exclude_kanji: &FnvHashSet<char>,
exclude_words: &FnvHashSet<String>, exclude_words: &FnvHashSet<String>,
pitches: &[u8], pitches: &[u8],
@ -405,7 +407,10 @@ fn apply_furigana<'a>(
) -> Vec<(String, String)> { ) -> Vec<(String, String)> {
let mut out: Vec<(String, String)> = Vec::new(); let mut out: Vec<(String, String)> = Vec::new();
if furigana_unneeded(surface, exclude_kanji, exclude_words) || !is_kana_str(kana) { if furigana_unneeded(surface, exclude_kanji, exclude_words)
|| furigana_unneeded(dictionary_form, exclude_kanji, exclude_words)
|| !is_kana_str(kana)
{
return Vec::new(); return Vec::new();
} }