maestro/web/src/structuredDiff.ts

// Line-level diff for a `ToolDiff` row's `Diff` content block (#140). ACP
// hands over full `oldText`/`newText`, not a unified-diff patch, so the
// coloured hunks the design calls for have to be computed here.

export type DiffLineKind = "add" | "del" | "ctx" | "hunk" | "eof";

export interface DiffLine {
  kind: DiffLineKind;
  /** The line's text (a `hunk` line is a human summary, not source text). */
  text: string;
}

/** Interior unchanged runs longer than this collapse to a summary line
 *  flanked by this many lines of real context on each side. */
const CONTEXT = 3;

/** Bounds the O(n·m) LCS table: above this cell count, the changed middle is
 *  shown as one coarse del-then-add block instead of a fine-grained diff —
 *  still correct, just not minimal. */
const LCS_CELL_LIMIT = 250_000;

interface SourceLine {
  text: string;
  terminated: boolean;
}

function splitLines(text: string): SourceLine[] {
  if (text === "") return [];
  const parts = text.split("\n");
  const hasFinalNewline = text.endsWith("\n");
  if (hasFinalNewline) parts.pop();
  return parts.map((line, index) => ({
    text: line,
    terminated: index < parts.length - 1 || hasFinalNewline,
  }));
}

type Op = { kind: "add" | "del" | "ctx"; line: SourceLine };

function sameLine(a: SourceLine, b: SourceLine): boolean {
  return a.text === b.text && a.terminated === b.terminated;
}

/** Longest-common-subsequence diff of two line arrays, already trimmed of any
 *  shared prefix/suffix by the caller. */
function lcsDiff(a: SourceLine[], b: SourceLine[]): Op[] {
  const n = a.length;
  const m = b.length;
  if (n === 0) return b.map((line) => ({ kind: "add", line }));
  if (m === 0) return a.map((line) => ({ kind: "del", line }));
  if (n * m > LCS_CELL_LIMIT) {
    return [...a.map((line) => ({ kind: "del" as const, line })), ...b.map((line) => ({ kind: "add" as const, line }))];
  }
  // dp[i][j] = length of the LCS of a[i:] and b[j:].
  const columns = m + 1;
  const dp = new Uint32Array((n + 1) * columns);
  for (let i = n - 1; i >= 0; i--) {
    for (let j = m - 1; j >= 0; j--) {
      dp[i * columns + j] = sameLine(a[i], b[j])
        ? dp[(i + 1) * columns + j + 1] + 1
        : Math.max(dp[(i + 1) * columns + j], dp[i * columns + j + 1]);
    }
  }
  const ops: Op[] = [];
  let i = 0;
  let j = 0;
  while (i < n && j < m) {
    if (sameLine(a[i], b[j])) {
      ops.push({ kind: "ctx", line: a[i] });
      i++;
      j++;
    } else if (dp[(i + 1) * columns + j] >= dp[i * columns + j + 1]) {
      ops.push({ kind: "del", line: a[i] });
      i++;
    } else {
      ops.push({ kind: "add", line: b[j] });
      j++;
    }
  }
  while (i < n) ops.push({ kind: "del", line: a[i++] });
  while (j < m) ops.push({ kind: "add", line: b[j++] });
  return ops;
}

function diffLines(a: SourceLine[], b: SourceLine[]): Op[] {
  let pre = 0;
  while (pre < a.length && pre < b.length && sameLine(a[pre], b[pre])) pre++;
  let suf = 0;
  while (
    suf < a.length - pre &&
    suf < b.length - pre &&
    sameLine(a[a.length - 1 - suf], b[b.length - 1 - suf])
  ) {
    suf++;
  }
  const prefixCtx: Op[] = a.slice(0, pre).map((line) => ({ kind: "ctx", line }));
  const suffixCtx: Op[] = a.slice(a.length - suf).map((line) => ({ kind: "ctx", line }));
  const middle = lcsDiff(a.slice(pre, a.length - suf), b.slice(pre, b.length - suf));
  return [...prefixCtx, ...middle, ...suffixCtx];
}

/** Collapse a long run of unchanged lines to a summary, keeping up to
 *  {@link CONTEXT} real lines adjacent to each neighbouring change — same
 *  shape a unified diff gives a reader, without claiming line numbers this
 *  module never tracked. A run touching the start/end of the whole diff has
 *  no change to be adjacent to on that side, so it keeps no context there
 *  (unless it is the *only* run, i.e. nothing changed at all). */
function hunk(ops: Op[]): DiffLine[] {
  const out: DiffLine[] = [];
  const push = (op: Op) => {
    out.push({ kind: op.kind, text: op.line.text });
    if (!op.line.terminated) {
      out.push({ kind: "eof", text: "\\ No newline at end of file" });
    }
  };
  let i = 0;
  while (i < ops.length) {
    if (ops[i].kind !== "ctx") {
      push(ops[i]);
      i++;
      continue;
    }
    let j = i;
    while (j < ops.length && ops[j].kind === "ctx") j++;
    const run = ops.slice(i, j);
    const wholeDiff = i === 0 && j === ops.length;
    const keepHead = wholeDiff || i > 0 ? CONTEXT : 0;
    const keepTail = wholeDiff || j < ops.length ? CONTEXT : 0;
    if (run.length <= keepHead + keepTail) {
      run.forEach(push);
    } else {
      if (keepHead) run.slice(0, keepHead).forEach(push);
      const collapsed = run.slice(keepHead, run.length - keepTail);
      out.push({
        kind: "hunk",
        text: `⋯ ${collapsed.length} unchanged lines ⋯`,
      });
      if (collapsed.some((op) => !op.line.terminated)) {
        out.push({ kind: "eof", text: "\\ No newline at end of file" });
      }
      if (keepTail) run.slice(run.length - keepTail).forEach(push);
    }
    i = j;
  }
  return out;
}

/** Coloured, hunked diff lines for a `Diff` content block. `oldText` absent
 *  (a new file) renders every line of `newText` as added. */
export function computeDiffLines(oldText: string | null | undefined, newText: string): DiffLine[] {
  const newLines = splitLines(newText);
  if (oldText == null) return hunk(newLines.map((line) => ({ kind: "add", line })));
  const ops = diffLines(splitLines(oldText), newLines);
  return hunk(ops);
}

/** `+added -removed` counts straight off the diff — real derived data, not a
 *  fabricated duration or line count. */
export function diffStats(lines: DiffLine[]): { added: number; removed: number } {
  let added = 0;
  let removed = 0;
  for (const line of lines) {
    if (line.kind === "add") added++;
    else if (line.kind === "del") removed++;
  }
  return { added, removed };
}