Performance improvements for LCS and inline highlighting (#16)

This commit is contained in:
Armin Ronacher 2021-02-17 20:59:55 +01:00 committed by GitHub
parent bdb1f7a644
commit 45bcb3943c
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 34 additions and 13 deletions

View file

@ -2,6 +2,7 @@
//!
//! * time: `O((NM)D log (M)D)`
//! * space `O(MN)`
use std::collections::BTreeMap;
use std::ops::{Index, Range};
use crate::algorithms::DiffHook;
@ -66,7 +67,9 @@ where
d.equal(old_orig_idx, new_orig_idx, 1)?;
old_idx += 1;
new_idx += 1;
} else if table[new_idx][old_idx + 1] >= table[new_idx + 1][old_idx] {
} else if table.get(&(new_idx, old_idx + 1)).map_or(0, |&x| x)
>= table.get(&(new_idx + 1, old_idx)).map_or(0, |&x| x)
{
d.delete(old_orig_idx, 1, new_orig_idx)?;
old_idx += 1;
} else {
@ -117,7 +120,7 @@ fn make_table<Old, New>(
old_range: Range<usize>,
new: &New,
new_range: Range<usize>,
) -> Vec<Vec<u32>>
) -> BTreeMap<(usize, usize), u32>
where
Old: Index<usize> + ?Sized,
New: Index<usize> + ?Sized,
@ -125,17 +128,20 @@ where
{
let old_len = old_range.len();
let new_len = new_range.len();
let mut table = vec![vec![0; old_len + 1]; new_len + 1];
let mut table = BTreeMap::new();
for i in 0..new_len {
let i = new_len - i - 1;
table[i][old_len] = 0;
for j in 0..old_len {
let j = old_len - j - 1;
table[i][j] = if new[i] == old[j] {
table[i + 1][j + 1] + 1
for i in (0..new_len).rev() {
for j in (0..old_len).rev() {
let val = if new[i] == old[j] {
table.get(&(i + 1, j + 1)).map_or(0, |&x| x) + 1
} else {
table[i + 1][j].max(table[i][j + 1])
table
.get(&(i + 1, j))
.map_or(0, |&x| x)
.max(table.get(&(i, j + 1)).map_or(0, |&x| x))
};
if val > 0 {
table.insert((i, j), val);
}
}
}
@ -146,7 +152,13 @@ where
#[test]
fn test_table() {
let table = make_table(&vec![2, 3], 0..2, &vec![0, 1, 2], 0..3);
let expected = vec![vec![1, 0, 0], vec![1, 0, 0], vec![1, 0, 0], vec![0, 0, 0]];
let expected = {
let mut m = BTreeMap::new();
m.insert((1, 0), 1);
m.insert((0, 0), 1);
m.insert((2, 0), 1);
m
};
assert_eq!(table, expected);
}

View file

@ -8,6 +8,8 @@ use crate::{capture_diff, get_diff_ratio};
use std::ops::Index;
use super::utils::upper_seq_ratio;
struct MultiLookup<'bufs, 's, T: DiffableStr + ?Sized> {
strings: &'bufs [&'s T],
seqs: Vec<(&'s T, usize, usize)>,
@ -184,6 +186,8 @@ impl<'s, T: DiffableStr + ?Sized> fmt::Display for InlineChange<'s, T> {
}
}
const MIN_RATIO: f32 = 0.5;
pub(crate) fn iter_inline_changes<'x, 'diff, 'old, 'new, 'bufs, T>(
diff: &'diff TextDiff<'old, 'new, 'bufs, T>,
op: &DiffOp,
@ -204,6 +208,11 @@ where
let mut new_index = new_range.start;
let old_slices = &diff.old_slices()[old_range];
let new_slices = &diff.new_slices()[new_range];
if upper_seq_ratio(old_slices, new_slices) < MIN_RATIO {
return Box::new(diff.iter_changes(op).map(|x| x.into())) as Box<dyn Iterator<Item = _>>;
}
let old_lookup = MultiLookup::new(old_slices);
let new_lookup = MultiLookup::new(new_slices);
@ -215,7 +224,7 @@ where
0..new_lookup.len(),
);
if get_diff_ratio(&ops, old_lookup.len(), new_lookup.len()) < 0.5 {
if get_diff_ratio(&ops, old_lookup.len(), new_lookup.len()) < MIN_RATIO {
return Box::new(diff.iter_changes(op).map(|x| x.into())) as Box<dyn Iterator<Item = _>>;
}