79 lines
2.6 KiB
TypeScript
79 lines
2.6 KiB
TypeScript
 
|
/**
|
||
|
|
* [Gitea #144] Word-level diff for the "(edited)" hover. Plain text only; the
|
||
|
|
* caller decides what to do with formatted bodies. LCS over word tokens, with
|
||
|
|
* whitespace preserved on the token that precedes it so the rendered result
|
||
|
|
* reads like the original message.
|
||
|
|
*/
|
||
|
|
|
||
|
|
export type DiffOp = 'same' | 'add' | 'del';
|
||
|
|
export type DiffToken = { op: DiffOp; text: string };
|
||
|
|
|
||
|
|
/** Above this many words (before + after) we give up: O(n·m) memory and an unreadable tooltip. */
|
||
|
|
export const WORD_DIFF_LIMIT = 400;
|
||
|
|
|
||
|
|
const tokenize = (text: string): string[] => text.match(/\S+\s*|\s+/g) ?? [];
|
||
|
|
const word = (token: string) => token.trim();
|
||
|
|
|
||
|
|
/**
|
||
|
|
* Diff `before` → `after` into runs. Returns `undefined` when the texts are
|
||
|
|
* identical or too long to diff meaningfully (caller should fall back to the
|
||
|
|
* full history viewer).
|
||
|
|
*/
|
||
|
|
export function wordDiff(before: string, after: string): DiffToken[] | undefined {
|
||
|
|
if (before === after) return undefined;
|
||
|
|
const a = tokenize(before);
|
||
|
|
const b = tokenize(after);
|
||
|
|
if (a.length + b.length > WORD_DIFF_LIMIT) return undefined;
|
||
|
|
|
||
|
|
// LCS table on trimmed words so whitespace-only changes don't count as edits.
|
||
|
|
const n = a.length;
|
||
|
|
const m = b.length;
|
||
|
|
const lcs: Uint16Array[] = Array.from({ length: n + 1 }, () => new Uint16Array(m + 1));
|
||
|
|
for (let i = n - 1; i >= 0; i -= 1) {
|
||
|
|
for (let j = m - 1; j >= 0; j -= 1) {
|
||
|
|
lcs[i][j] =
|
||
|
|
word(a[i]) === word(b[j]) ? lcs[i + 1][j + 1] + 1 : Math.max(lcs[i + 1][j], lcs[i][j + 1]);
|
||
|
|
}
|
||
|
|
}
|
||
|
|
|
||
|
|
const out: DiffToken[] = [];
|
||
|
|
const push = (op: DiffOp, text: string) => {
|
||
|
|
const last = out[out.length - 1];
|
||
|
|
if (last && last.op === op) last.text += text;
|
||
|
|
else out.push({ op, text });
|
||
|
|
};
|
||
|
|
let i = 0;
|
||
|
|
let j = 0;
|
||
|
|
while (i < n && j < m) {
|
||
|
|
if (word(a[i]) === word(b[j])) {
|
||
|
|
push('same', b[j]);
|
||
|
|
i += 1;
|
||
|
|
j += 1;
|
||
|
|
} else if (lcs[i + 1][j] >= lcs[i][j + 1]) {
|
||
|
|
push('del', a[i]);
|
||
|
|
i += 1;
|
||
|
|
} else {
|
||
|
|
push('add', b[j]);
|
||
|
|
j += 1;
|
||
|
|
}
|
||
|
|
}
|
||
|
|
while (i < n) push('del', a[i++]);
|
||
|
|
while (j < m) push('add', b[j++]);
|
||
|
|
|
||
|
|
// Only whitespace moved around: nothing worth showing.
|
||
|
|
if (out.every((t) => t.op === 'same' || t.text.trim() === '')) return undefined;
|
||
|
|
return out;
|
||
|
|
}
|
||
|
|
|
||
|
|
/** Number of changed words in a diff (for a compact summary). */
|
||
|
|
export const countChanges = (tokens: DiffToken[]): { added: number; removed: number } => {
|
||
|
|
let added = 0;
|
||
|
|
let removed = 0;
|
||
|
|
tokens.forEach((t) => {
|
||
|
|
const words = t.text.trim() === '' ? 0 : t.text.trim().split(/\s+/).length;
|
||
|
|
if (t.op === 'add') added += words;
|
||
|
|
if (t.op === 'del') removed += words;
|
||
|
|
});
|
||
|
|
return { added, removed };
|
||
|
|
};
|