HIsheR
59 lines
3.5 KiB
TypeScript
59 lines
3.5 KiB
TypeScript
/*** 行级差异纯逻辑：切行 / 归一键 / 中间段 LCS 对齐；不碰 DOM，可被冒烟脚本直接跑 ***/
|
||
|
||
/*** 一条差异：same 原样、add 只在右边、del 只在左边 ***/
|
||
export type DiffOp = { type: 'same' | 'add' | 'del'; text: string };
|
||
/*** 比较用的一行：text 是显示原文，key 是比对键（忽略空白模式下把行内空白压成一个空格） ***/
|
||
export type DiffRow = { text: string; key: string };
|
||
// LCS 计算上限：中间段行数乘积超过这个值就退化为「整段替换」，避免大文件把主线程卡住
|
||
export const DP_LIMIT = 400_000;
|
||
|
||
/*** 作用：比对键归一——行内连续空白压成一个空格并去首尾空白；显示仍用原文，缩进不丢 ***/
|
||
export function normalizeLine(line: string): string {
|
||
return line.replace(/\s+/g, ' ').trim();
|
||
}
|
||
|
||
/*** 作用：切行并算比对键；末尾换行不产生一行空内容（否则每个文件都多一条假差异） ***/
|
||
export function toRows(text: string, ignore: boolean): DiffRow[] {
|
||
const lines = text.split(/\r?\n/);
|
||
if (lines.length > 1 && lines[lines.length - 1] === '') lines.pop(); // 尾部换行忽略
|
||
return lines.map((line) => ({ text: line, key: ignore ? normalizeLine(line) : line }));
|
||
}
|
||
|
||
/*** 作用：中间段做 LCS 对齐，输出 del / add / same 序列；超限或一侧为空时退化为整段替换 ***/
|
||
export function alignMiddle(a: DiffRow[], b: DiffRow[]): DiffOp[] {
|
||
if (!a.length || !b.length) return [...a.map((r) => ({ type: 'del' as const, text: r.text })), ...b.map((r) => ({ type: 'add' as const, text: r.text }))];
|
||
if (a.length * b.length > DP_LIMIT) return [...a.map((r) => ({ type: 'del' as const, text: r.text })), ...b.map((r) => ({ type: 'add' as const, text: r.text }))]; // 超限：语义仍正确，只是不再逐行对齐
|
||
const dp: number[][] = Array.from({ length: a.length + 1 }, () => new Array<number>(b.length + 1).fill(0)); // dp[i][j]：a 前 i 行与 b 前 j 行的 LCS 长度
|
||
for (let i = a.length - 1; i >= 0; i -= 1) {
|
||
for (let j = b.length - 1; j >= 0; j -= 1) {
|
||
dp[i][j] = a[i].key === b[j].key ? dp[i + 1][j + 1] + 1 : Math.max(dp[i + 1][j], dp[i][j + 1]);
|
||
}
|
||
}
|
||
const ops: DiffOp[] = [];
|
||
let i = 0;
|
||
let j = 0;
|
||
while (i < a.length && j < b.length) {
|
||
if (a[i].key === b[j].key) { ops.push({ type: 'same', text: a[i].text }); i += 1; j += 1; continue; }
|
||
if (dp[i + 1][j] >= dp[i][j + 1]) { ops.push({ type: 'del', text: a[i].text }); i += 1; continue; }
|
||
ops.push({ type: 'add', text: b[j].text }); j += 1;
|
||
}
|
||
for (; i < a.length; i += 1) ops.push({ type: 'del', text: a[i].text });
|
||
for (; j < b.length; j += 1) ops.push({ type: 'add', text: b[j].text });
|
||
return ops;
|
||
}
|
||
|
||
/*** 作用：逐行差异——剪公共前后缀后只对中间段做对齐，日常改动只占几行 ***/
|
||
export function diffLines(before: string, after: string, ignore: boolean): DiffOp[] {
|
||
const a = toRows(before, ignore);
|
||
const b = toRows(after, ignore);
|
||
let head = 0;
|
||
while (head < a.length && head < b.length && a[head].key === b[head].key) head += 1; // 公共前缀
|
||
let tail = 0;
|
||
while (tail < a.length - head && tail < b.length - head && a[a.length - 1 - tail].key === b[b.length - 1 - tail].key) tail += 1; // 公共后缀
|
||
return [
|
||
...a.slice(0, head).map((r) => ({ type: 'same' as const, text: r.text })),
|
||
...alignMiddle(a.slice(head, a.length - tail), b.slice(head, b.length - tail)),
|
||
...a.slice(a.length - tail).map((r) => ({ type: 'same' as const, text: r.text })),
|
||
];
|
||
}
|