Improve compact diff API #24

Merged
schmelczer merged 11 commits from asch/refactor-api into main 2025-11-16 15:43:19 +00:00
2 changed files with 42 additions and 39 deletions
Showing only changes of commit c9c2b775a8 - Show all commits

Don't unwrap to MAX number

Andras Schmelczer 2025-11-16 15:01:51 +00:00

View file

@ -168,13 +168,14 @@ where
let result = operation.merge_operations(&mut last_other_op); let result = operation.merge_operations(&mut last_other_op);
if let ref op @ (Operation::Insert { .. } | Operation::Equal { .. }) = result { if let ref op @ (Operation::Insert { .. } | Operation::Equal { .. }) = result {
let merged_length_signed = let merged_length_signed = isize::try_from(merged_length)
isize::try_from(merged_length).unwrap_or(isize::MAX); .expect("merged_length must fit in isize");
let seen_left_length_signed = let seen_left_length_signed = isize::try_from(seen_left_length)
isize::try_from(seen_left_length).unwrap_or(isize::MAX); .expect("seen_left_length must fit in isize");
let op_len_signed = isize::try_from(op.len()).unwrap_or(isize::MAX); let op_len_signed =
let original_length_signed = isize::try_from(op.len()).expect("op.len() must fit in isize");
isize::try_from(original_length).unwrap_or(isize::MAX); let original_length_signed = isize::try_from(original_length)
.expect("original_length must fit in isize");
let shift = merged_length_signed - seen_left_length_signed + op_len_signed let shift = merged_length_signed - seen_left_length_signed + op_len_signed
- original_length_signed; - original_length_signed;
@ -201,13 +202,14 @@ where
let result = operation.merge_operations(&mut last_other_op); let result = operation.merge_operations(&mut last_other_op);
if let ref op @ (Operation::Insert { .. } | Operation::Equal { .. }) = result { if let ref op @ (Operation::Insert { .. } | Operation::Equal { .. }) = result {
let merged_length_signed = let merged_length_signed = isize::try_from(merged_length)
isize::try_from(merged_length).unwrap_or(isize::MAX); .expect("merged_length must fit in isize");
let seen_right_length_signed = let seen_right_length_signed = isize::try_from(seen_right_length)
isize::try_from(seen_right_length).unwrap_or(isize::MAX); .expect("seen_right_length must fit in isize");
let op_len_signed = isize::try_from(op.len()).unwrap_or(isize::MAX); let op_len_signed =
let original_length_signed = isize::try_from(op.len()).expect("op.len() must fit in isize");
isize::try_from(original_length).unwrap_or(isize::MAX); let original_length_signed = isize::try_from(original_length)
.expect("original_length must fit in isize");
let shift = merged_length_signed - seen_right_length_signed + op_len_signed let shift = merged_length_signed - seen_right_length_signed + op_len_signed
- original_length_signed; - original_length_signed;
@ -355,7 +357,7 @@ where
/// Inserts are represented as strings, deletes as negative integers, /// Inserts are represented as strings, deletes as negative integers,
/// and equal spans as positive integers. /// and equal spans as positive integers.
#[must_use] #[must_use]
pub fn to_changes(&self) -> Vec<NumberOrString> { pub fn to_diff(&self) -> Vec<NumberOrString> {
let mut result: Vec<NumberOrString> = Vec::with_capacity(self.operations.len()); let mut result: Vec<NumberOrString> = Vec::with_capacity(self.operations.len());
let mut previous_equal: Option<usize> = None; let mut previous_equal: Option<usize> = None;
@ -372,7 +374,7 @@ where
Operation::Insert { text, .. } => { Operation::Insert { text, .. } => {
if let Some(prev_length) = previous_equal { if let Some(prev_length) = previous_equal {
result.push(NumberOrString::Number( result.push(NumberOrString::Number(
i64::try_from(prev_length).unwrap_or(i64::MAX), i64::try_from(prev_length).expect("prev_length must fit in i64"),
)); ));
previous_equal = None; previous_equal = None;
} }
@ -390,12 +392,13 @@ where
} => { } => {
if let Some(prev_length) = previous_equal { if let Some(prev_length) = previous_equal {
result.push(NumberOrString::Number( result.push(NumberOrString::Number(
i64::try_from(prev_length).unwrap_or(i64::MAX), i64::try_from(prev_length).expect("prev_length must fit in i64"),
)); ));
previous_equal = None; previous_equal = None;
} }
let count = i64::try_from(*deleted_character_count).unwrap_or(i64::MAX); let count = i64::try_from(*deleted_character_count)
.expect("deleted_character_count must fit in i64");
result.push(NumberOrString::Number(-count)); result.push(NumberOrString::Number(-count));
} }
} }
@ -403,7 +406,7 @@ where
if let Some(prev_length) = previous_equal { if let Some(prev_length) = previous_equal {
result.push(NumberOrString::Number( result.push(NumberOrString::Number(
i64::try_from(prev_length).unwrap_or(i64::MAX), i64::try_from(prev_length).expect("prev_length must fit in i64"),
)); ));
} }
@ -412,19 +415,19 @@ where
/// Deserialize an `EditedText` from a change list and the original text. /// Deserialize an `EditedText` from a change list and the original text.
#[must_use] #[must_use]
pub fn from_changes( pub fn from_diff(
original_text: &'a str, original_text: &'a str,
simple_operations: Vec<NumberOrString>, diff: Vec<NumberOrString>,
tokenizer: &Tokenizer<T>, tokenizer: &Tokenizer<T>,
) -> EditedText<'a, T> { ) -> EditedText<'a, T> {
let mut operations: Vec<Operation<T>> = Vec::with_capacity(simple_operations.len()); let mut operations: Vec<Operation<T>> = Vec::with_capacity(diff.len());
let mut order = 0; let mut order = 0;
for simple_operation in simple_operations { for item in diff {
match simple_operation { match item {
NumberOrString::Number(length) => { NumberOrString::Number(length) => {
if length >= 0 { if length >= 0 {
let length = usize::try_from(length).unwrap_or(usize::MAX); let length = usize::try_from(length).expect("length must fit in usize");
let original_characters: String = let original_characters: String =
original_text.chars().skip(order).take(length).collect(); original_text.chars().skip(order).take(length).collect();
@ -435,7 +438,8 @@ where
order += token.get_original_length(); order += token.get_original_length();
} }
} else { } else {
let length = usize::try_from(-length).unwrap_or(usize::MAX); let length =
usize::try_from(-length).expect("negative length must fit in usize");
operations.push(Operation::create_delete(order, length)); operations.push(Operation::create_delete(order, length));
order += length; order += length;
} }
@ -509,7 +513,7 @@ mod tests {
let original = "Merging text is hard!"; let original = "Merging text is hard!";
let changes = "Merging text is easy with reconcile!"; let changes = "Merging text is easy with reconcile!";
let result = EditedText::from_strings(original, &changes.into()); let result = EditedText::from_strings(original, &changes.into());
let serialized = serde_yaml::to_string(&result.to_changes()).unwrap(); let serialized = serde_yaml::to_string(&result.to_diff()).unwrap();
let expected = concat!("- 15\n", "- -6\n", "- ' easy with reconcile!'\n",); let expected = concat!("- 15\n", "- -6\n", "- ' easy with reconcile!'\n",);
assert_eq!(serialized, expected); assert_eq!(serialized, expected);
@ -523,9 +527,9 @@ mod tests {
let edited_text = EditedText::from_strings(original, &updated.into()); let edited_text = EditedText::from_strings(original, &updated.into());
let changes = edited_text.to_changes(); let changes = edited_text.to_diff();
let deserialized_edited_text = let deserialized_edited_text =
EditedText::from_changes(original, changes, &*BuiltinTokenizer::Word); EditedText::from_diff(original, changes, &*BuiltinTokenizer::Word);
assert_eq!(deserialized_edited_text.apply().text(), updated); assert_eq!(deserialized_edited_text.apply().text(), updated);
} }

View file

@ -87,7 +87,7 @@ struct V {
impl V { impl V {
fn new(max_d: usize) -> Self { fn new(max_d: usize) -> Self {
// max_d should fit in isize for the algorithm to work correctly // max_d should fit in isize for the algorithm to work correctly
let offset = isize::try_from(max_d).unwrap_or(isize::MAX); let offset = isize::try_from(max_d).expect("max_d must fit in isize");
Self { Self {
offset, offset,
v: vec![0; 2 * max_d], v: vec![0; 2 * max_d],
@ -101,16 +101,15 @@ impl Index<isize> for V {
type Output = usize; type Output = usize;
fn index(&self, index: isize) -> &Self::Output { fn index(&self, index: isize) -> &Self::Output {
let idx = usize::try_from(index + self.offset).unwrap_or(usize::MAX); let idx = usize::try_from(index + self.offset).expect("index + offset must fit in usize");
&self.v[idx.min(self.v.len().saturating_sub(1))] &self.v[idx]
} }
} }
impl IndexMut<isize> for V { impl IndexMut<isize> for V {
fn index_mut(&mut self, index: isize) -> &mut Self::Output { fn index_mut(&mut self, index: isize) -> &mut Self::Output {
let idx = usize::try_from(index + self.offset).unwrap_or(usize::MAX); let idx = usize::try_from(index + self.offset).expect("index + offset must fit in usize");
let len = self.v.len(); &mut self.v[idx]
&mut self.v[idx.min(len.saturating_sub(1))]
} }
} }
@ -145,7 +144,7 @@ where
// By Lemma 1 in the paper, the optimal edit script length is odd or even as // By Lemma 1 in the paper, the optimal edit script length is odd or even as
// `delta` is odd or even. // `delta` is odd or even.
let delta = isize::try_from(n).unwrap_or(isize::MAX) - isize::try_from(m).unwrap_or(isize::MAX); let delta = isize::try_from(n).expect("n must fit in isize") - isize::try_from(m).expect("m must fit in isize");
let odd = delta & 1 == 1; let odd = delta & 1 == 1;
// The initial point at (0, -1) // The initial point at (0, -1)
@ -157,7 +156,7 @@ where
assert!(vf.len() >= d_max); assert!(vf.len() >= d_max);
assert!(vb.len() >= d_max); assert!(vb.len() >= d_max);
let d_max_isize = isize::try_from(d_max).unwrap_or(isize::MAX); let d_max_isize = isize::try_from(d_max).expect("d_max must fit in isize");
for d in 0..d_max_isize { for d in 0..d_max_isize {
// Forward path // Forward path
for k in (-d..=d).rev().step_by(2) { for k in (-d..=d).rev().step_by(2) {
@ -166,7 +165,7 @@ where
} else { } else {
vf[k - 1] + 1 vf[k - 1] + 1
}; };
let y = usize::try_from(isize::try_from(x).unwrap_or(isize::MAX) - k).unwrap_or(0); let y = usize::try_from(isize::try_from(x).expect("x must fit in isize") - k).expect("x - k must be non-negative and fit in usize");
// The coordinate of the start of a snake // The coordinate of the start of a snake
let (x0, y0) = (x, y); let (x0, y0) = (x, y);
@ -204,7 +203,7 @@ where
} else { } else {
vb[k - 1] + 1 vb[k - 1] + 1
}; };
let mut y = usize::try_from(isize::try_from(x).unwrap_or(isize::MAX) - k).unwrap_or(0); let mut y = usize::try_from(isize::try_from(x).expect("x must fit in isize") - k).expect("x - k must be non-negative and fit in usize");
// The coordinate of the start of a snake // The coordinate of the start of a snake
if x < n && y < m { if x < n && y < m {