style: cargo fmt --workspace (whitespace/wrapping only, no semantic change)
Whole-workspace rustfmt pass picked up while iterating on Mamba GPU backward work. Verified formatting-only via diff sampling; no logic changed. Co-Authored-By: Claude Sonnet 5 <[email protected]>
This commit is contained in:
@@ -76,9 +76,26 @@ impl TextNormalizer {
|
||||
abbreviation_map.insert("e.g.".to_string(), "for example".to_string());
|
||||
|
||||
let number_words = vec![
|
||||
"zero", "one", "two", "three", "four", "five", "six", "seven", "eight", "nine",
|
||||
"ten", "eleven", "twelve", "thirteen", "fourteen", "fifteen", "sixteen",
|
||||
"seventeen", "eighteen", "nineteen",
|
||||
"zero",
|
||||
"one",
|
||||
"two",
|
||||
"three",
|
||||
"four",
|
||||
"five",
|
||||
"six",
|
||||
"seven",
|
||||
"eight",
|
||||
"nine",
|
||||
"ten",
|
||||
"eleven",
|
||||
"twelve",
|
||||
"thirteen",
|
||||
"fourteen",
|
||||
"fifteen",
|
||||
"sixteen",
|
||||
"seventeen",
|
||||
"eighteen",
|
||||
"nineteen",
|
||||
];
|
||||
|
||||
let tens_words = vec![
|
||||
@@ -164,14 +181,16 @@ impl TextNormalizer {
|
||||
};
|
||||
|
||||
// Preserve the trailing space if it exists
|
||||
result = re.replace_all(&result, |caps: ®ex::Captures| {
|
||||
let matched = &caps[0];
|
||||
if matched.ends_with(' ') {
|
||||
format!("{} ", expansion)
|
||||
} else {
|
||||
expansion.to_string()
|
||||
}
|
||||
}).to_string();
|
||||
result = re
|
||||
.replace_all(&result, |caps: ®ex::Captures| {
|
||||
let matched = &caps[0];
|
||||
if matched.ends_with(' ') {
|
||||
format!("{} ", expansion)
|
||||
} else {
|
||||
expansion.to_string()
|
||||
}
|
||||
})
|
||||
.to_string();
|
||||
}
|
||||
|
||||
Ok(result)
|
||||
@@ -210,7 +229,10 @@ impl TextNormalizer {
|
||||
if ones == 0 {
|
||||
return self.tens_words[tens as usize].to_string();
|
||||
}
|
||||
return format!("{} {}", self.tens_words[tens as usize], self.number_words[ones as usize]);
|
||||
return format!(
|
||||
"{} {}",
|
||||
self.tens_words[tens as usize], self.number_words[ones as usize]
|
||||
);
|
||||
}
|
||||
|
||||
if num < 1000 {
|
||||
@@ -219,7 +241,11 @@ impl TextNormalizer {
|
||||
if remainder == 0 {
|
||||
return format!("{} hundred", self.number_words[hundreds as usize]);
|
||||
}
|
||||
return format!("{} hundred {}", self.number_words[hundreds as usize], self.number_to_words(remainder));
|
||||
return format!(
|
||||
"{} hundred {}",
|
||||
self.number_words[hundreds as usize],
|
||||
self.number_to_words(remainder)
|
||||
);
|
||||
}
|
||||
|
||||
if num < 1_000_000 {
|
||||
@@ -228,7 +254,11 @@ impl TextNormalizer {
|
||||
if remainder == 0 {
|
||||
return format!("{} thousand", self.number_to_words(thousands));
|
||||
}
|
||||
return format!("{} thousand {}", self.number_to_words(thousands), self.number_to_words(remainder));
|
||||
return format!(
|
||||
"{} thousand {}",
|
||||
self.number_to_words(thousands),
|
||||
self.number_to_words(remainder)
|
||||
);
|
||||
}
|
||||
|
||||
// For larger numbers, just return the digit string
|
||||
@@ -259,9 +289,7 @@ impl TextNormalizer {
|
||||
|
||||
/// Remove punctuation marks
|
||||
fn remove_punctuation(&self, text: &str) -> String {
|
||||
text.chars()
|
||||
.filter(|c| !c.is_ascii_punctuation())
|
||||
.collect()
|
||||
text.chars().filter(|c| !c.is_ascii_punctuation()).collect()
|
||||
}
|
||||
|
||||
/// Normalize whitespace (collapse multiple spaces into one)
|
||||
@@ -335,7 +363,10 @@ mod tests {
|
||||
assert_eq!(normalizer.number_to_words(100), "one hundred");
|
||||
assert_eq!(normalizer.number_to_words(256), "two hundred fifty six");
|
||||
assert_eq!(normalizer.number_to_words(1000), "one thousand");
|
||||
assert_eq!(normalizer.number_to_words(1234), "one thousand two hundred thirty four");
|
||||
assert_eq!(
|
||||
normalizer.number_to_words(1234),
|
||||
"one thousand two hundred thirty four"
|
||||
);
|
||||
assert_eq!(normalizer.number_to_words(-5), "minus five");
|
||||
}
|
||||
|
||||
@@ -350,7 +381,9 @@ mod tests {
|
||||
};
|
||||
let normalizer = TextNormalizer::new(config);
|
||||
|
||||
let result = normalizer.normalize("I have 3 apples and 42 oranges.").unwrap();
|
||||
let result = normalizer
|
||||
.normalize("I have 3 apples and 42 oranges.")
|
||||
.unwrap();
|
||||
assert_eq!(result, "I have three apples and forty two oranges.");
|
||||
|
||||
let result = normalizer.normalize("The year 2024 was great.").unwrap();
|
||||
@@ -446,8 +479,13 @@ mod tests {
|
||||
};
|
||||
let normalizer = TextNormalizer::new(config);
|
||||
|
||||
let result = normalizer.normalize("Dr. Smith said I have 3 apples for 5 dollars").unwrap();
|
||||
assert_eq!(result, "doctor smith said i have three apples for five dollars");
|
||||
let result = normalizer
|
||||
.normalize("Dr. Smith said I have 3 apples for 5 dollars")
|
||||
.unwrap();
|
||||
assert_eq!(
|
||||
result,
|
||||
"doctor smith said i have three apples for five dollars"
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
@@ -470,7 +508,9 @@ mod tests {
|
||||
};
|
||||
let normalizer = TextNormalizer::new(config);
|
||||
|
||||
let result = normalizer.normalize("Mr. and Mrs. Smith live on Oak St.").unwrap();
|
||||
let result = normalizer
|
||||
.normalize("Mr. and Mrs. Smith live on Oak St.")
|
||||
.unwrap();
|
||||
assert_eq!(result, "Mister and Missus Smith live on Oak Street");
|
||||
}
|
||||
|
||||
@@ -485,7 +525,9 @@ mod tests {
|
||||
};
|
||||
let normalizer = TextNormalizer::new(config);
|
||||
|
||||
let result = normalizer.normalize("Dr. Smith has 42 patients.").unwrap();
|
||||
let result = normalizer
|
||||
.normalize("Dr. Smith has 42 patients.")
|
||||
.unwrap();
|
||||
assert_eq!(result, "doctor smith has forty two patients.");
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user