fix: update large number samples

This commit is contained in:
ThetaDev 2023-05-06 01:22:13 +02:00
parent e94de9a0f6
commit 72d817edd7
8 changed files with 33785 additions and 16936 deletions

View file

@ -339,22 +339,24 @@ pub(crate) fn entry(lang: Language) -> Entry {
},
comma_decimal: false,
number_tokens: ::phf::Map {
key: 10121458955350035957,
key: 12913932095322966823,
disps: &[
(0, 5),
(6, 0),
(0, 7),
(9, 8),
(0, 0),
],
entries: &[
("\u{9be}", 5),
("কোঃট\u{9be}", 9),
("নিঃট\u{9be}", 6),
("\u{9be}\u{9be}", 3),
("\u{9be}খট\u{9be}", 5),
("নিয\u{9c1}", 6),
("\u{9be}", 3),
("\u{9be}", 5),
("\u{9be}\u{9be}ৰট\u{9be}", 3),
("নিয\u{9c1}তট\u{9be}", 6),
("\u{9be}\u{9be}", 3),
("\u{9be}", 5),
("\u{9be}খট\u{9be}", 5),
("কোঃট\u{9be}", 9),
("নিঃট\u{9be}", 6),
("নিয\u{9c1}", 6),
("নিঃ", 6),
("\u{9be}", 5),
],
},
album_types: ::phf::Map {
@ -851,14 +853,14 @@ pub(crate) fn entry(lang: Language) -> Entry {
},
comma_decimal: true,
number_tokens: ::phf::Map {
key: 12913932095322966823,
key: 7485420634051515786,
disps: &[
(2, 0),
],
entries: &[
("mM", 9),
("M", 6),
("m", 3),
("kM", 9),
("M", 6),
],
},
album_types: ::phf::Map {
@ -3181,18 +3183,14 @@ pub(crate) fn entry(lang: Language) -> Entry {
},
comma_decimal: false,
number_tokens: ::phf::Map {
key: 2980949210194914378,
key: 12913932095322966823,
disps: &[
(1, 3),
(5, 0),
(1, 0),
],
entries: &[
("억명", 8),
("천명", 3),
("만회", 4),
("천회", 3),
("억회", 8),
("만명", 4),
("", 3),
("", 4),
("", 8),
],
},
album_types: ::phf::Map {
@ -3964,18 +3962,18 @@ pub(crate) fn entry(lang: Language) -> Entry {
},
comma_decimal: false,
number_tokens: ::phf::Map {
key: 10121458955350035957,
key: 12913932095322966823,
disps: &[
(5, 1),
(2, 0),
(3, 0),
(0, 2),
],
entries: &[
("ထောင\u{103a}", 3),
("သန\u{103a}", 6),
("က\u{102f}ဋေထ", 10),
("", 3),
("က\u{102f}ဋေ", 7),
("သောင\u{103a}", 4),
("သန\u{103a}", 6),
("ထောင\u{103a}", 3),
("\u{102d}\u{103a}", 5),
("သောင\u{103a}", 4),
],
},
album_types: ::phf::Map {

View file

@ -335,7 +335,7 @@ where
_ => dict_entry.number_tokens.get(token).map(|t| *t as i32),
};
if dict_entry.by_char {
if dict_entry.by_char || lang == Language::Ko {
exp += filtered
.chars()
.filter_map(|token| lookup_token(&token.to_string()))
@ -511,7 +511,7 @@ pub(crate) mod tests {
fn t_parse_large_numstr_samples() {
let json_path = path!(*TESTFILES / "dict" / "large_number_samples.json");
let json_file = File::open(json_path).unwrap();
let number_samples: BTreeMap<Language, BTreeMap<u8, (String, u64)>> =
let number_samples: BTreeMap<Language, BTreeMap<String, (String, u64)>> =
serde_json::from_reader(BufReader::new(json_file)).unwrap();
number_samples.iter().for_each(|(lang, entry)| {
@ -540,12 +540,17 @@ pub(crate) mod tests {
// in the string.
let rounded = {
let n_significant_d = string.chars().filter(char::is_ascii_digit).count();
let mag = (expect as f64).log10().floor();
let factor = 10_u64.pow(1 + mag as u32 - n_significant_d as u32);
(((expect as f64) / factor as f64).floor() as u64) * factor
if n_significant_d == 0 {
expect
} else {
let mag = (expect as f64).log10().floor();
let factor = 10_u64.pow(1 + mag as u32 - n_significant_d as u32);
(((expect as f64) / factor as f64).floor() as u64) * factor
}
};
let res = parse_large_numstr::<u64>(string, lang).expect(string);
// TODO: add support for zero values
let res = parse_large_numstr::<u64>(string, lang).unwrap_or_default();
assert_eq!(res, rounded, "{string} (lang: {lang}, exact: {expect})");
}
}