char-slop/berry

git clone https://git.t4t.associates/char-slop/berry

Charlotte Somdiff: add line counts to diff index3c216312026-10-09

// A file's change as the view draws it: numbered lines, unchanged stretches folded, and what changed within lines.

import type { FileDiff, Sent } from "./types.ts";

/** One line of a change, numbered on the sides it is on. */
export type Line = {
  mark: " " | "-" | "+";
  /** Without its mark; a carriage return stays, for the view to show. */
  text: string;
  old: number | null;
  new: number | null;
  /** Where it is among the lines of each side the change holds, which are what is highlighted. */
  at: { old: number | null; new: number | null };
  /** `false` for the last line of a file that ends without one. */
  newline: boolean;
};

export type Piece =
  | { kind: "lines"; lines: Line[] }
  /** Unchanged lines, folded until they are opened. */
  | { kind: "fold"; lines: Line[] }
  /** Lines the patch left out: how many there are, when that is known. */
  | { kind: "gap"; count: number };

/** A row of the side-by-side view. */
export type Pair = { old: Line | null; new: Line | null };

/** Character ranges, `[start, end)`. */
export type Range = readonly [number, number];

/** Added and removed lines, excluding context and patch metadata. */
export function lineCounts(file: FileDiff) {
  let additions = 0;
  let removals = 0;
  for (const hunk of file.hunks) {
    for (const line of hunk.lines) {
      if (line[0] === "+") additions++;
      else if (line[0] === "-") removals++;
    }
  }
  return { additions, removals };
}

/** Every line `file`'s hunks hold, in order, and each side's text, for highlighting. */
export function linesOf(file: FileDiff) {
  const sides = { old: [] as string[], new: [] as string[] };
  const hunks = file.hunks.map((hunk) => {
    let old = hunk.old;
    let neu = hunk.new;
    const lines: Line[] = [];
    for (const raw of hunk.lines) {
      const mark = raw[0];
      if (mark === "\\") {
        const last = lines.at(-1);
        if (last) last.newline = false;
        continue;
      }
      if (mark !== " " && mark !== "-" && mark !== "+") continue;
      const text = raw.slice(1);
      const onOld = mark !== "+";
      const onNew = mark !== "-";
      lines.push({
        mark,
        text,
        old: onOld ? old++ : null,
        new: onNew ? neu++ : null,
        at: {
          old: onOld ? sides.old.push(text.replace(/\r$/, "")) - 1 : null,
          new: onNew ? sides.new.push(text.replace(/\r$/, "")) - 1 : null,
        },
        newline: true,
      });
    }
    return { hunk, lines, end: { old, new: neu } };
  });
  return { hunks, sides };
}

/**
 * `file`'s lines with each stretch of more than a few unchanged lines folded, but for `context` lines next to each
 * change; and gaps where the patch leaves lines out.
 */
export function pieces(file: FileDiff, context = 3): Piece[] {
  const out: Piece[] = [];
  const shown = (lines: Line[]) => {
    const last = out.at(-1);
    if (lines.length === 0) return;
    if (last?.kind === "lines") last.lines.push(...lines);
    else out.push({ kind: "lines", lines });
  };
  let previous: { old: number; new: number } = { old: 1, new: 1 };
  for (const { hunk, lines, end } of linesOf(file).hunks) {
    const missing = Math.max(hunk.old - previous.old, hunk.new - previous.new);
    if (missing > 0) out.push({ kind: "gap", count: missing });
    previous = end;
    let i = 0;
    while (i < lines.length) {
      let j = i;
      while (j < lines.length && lines[j].mark === " ") j++;
      if (j === i) {
        while (j < lines.length && lines[j].mark !== " ") j++;
        shown(lines.slice(i, j));
        i = j;
        continue;
      }
      const before = i === 0 ? 0 : context;
      const after = j === lines.length ? 0 : context;
      // folding away less than this saves nothing
      if (j - i - before - after < 4) shown(lines.slice(i, j));
      else {
        shown(lines.slice(i, i + before));
        out.push({ kind: "fold", lines: lines.slice(i + before, j - after) });
        shown(lines.slice(j - after, j));
      }
      i = j;
    }
  }
  return out;
}

/** Rows side by side: unchanged lines beside themselves, and each run of removals beside the additions that follow. */
export function pairs(lines: readonly Line[]): Pair[] {
  const out: Pair[] = [];
  let i = 0;
  while (i < lines.length) {
    if (lines[i].mark === " ") {
      out.push({ old: lines[i], new: lines[i] });
      i++;
      continue;
    }
    const removed: Line[] = [];
    const added: Line[] = [];
    while (i < lines.length && lines[i].mark !== " ") {
      (lines[i].mark === "-" ? removed : added).push(lines[i]);
      i++;
    }
    for (let k = 0; k < Math.max(removed.length, added.length); k++) {
      out.push({ old: removed[k] ?? null, new: added[k] ?? null });
    }
  }
  return out;
}

const TOKEN = /[\p{L}\p{N}_]+|\s+|[^\p{L}\p{N}_\s]/gu;

/**
 * What changed between two versions of a line, word by word, as ranges in each; `null` when they share too little for
 * that to say more than the whole line changing.
 */
export function wordChanges(a: string, b: string): { old: Range[]; new: Range[] } | null {
  const x = [...a.matchAll(TOKEN)];
  const y = [...b.matchAll(TOKEN)];
  if (x.length * y.length > 250_000) return null;
  // longest common subsequence of tokens, from the ends back
  const width = y.length + 1;
  const table = new Uint16Array((x.length + 1) * width);
  for (let i = x.length - 1; i >= 0; i--) {
    for (let j = y.length - 1; j >= 0; j--) {
      table[i * width + j] = x[i][0] === y[j][0]
        ? table[(i + 1) * width + j + 1] + 1
        : Math.max(table[(i + 1) * width + j], table[i * width + j + 1]);
    }
  }
  const kept = { old: new Set<number>(), new: new Set<number>() };
  let shared = 0;
  for (let i = 0, j = 0; i < x.length && j < y.length;) {
    if (x[i][0] === y[j][0]) {
      kept.old.add(i++);
      kept.new.add(j++);
      shared += x[i - 1][0].trim() ? x[i - 1][0].length : 0;
    } else if (table[(i + 1) * width + j] >= table[i * width + j + 1]) i++;
    else j++;
  }
  const visible = Math.max(a.replace(/\s/g, "").length, b.replace(/\s/g, "").length);
  if (visible > 0 && shared / visible < 0.4) return null;
  const ranges = (tokens: RegExpExecArray[], kept: Set<number>) => {
    const out: [number, number][] = [];
    for (const [i, token] of tokens.entries()) {
      if (kept.has(i)) continue;
      const last = out.at(-1);
      if (last && last[1] === token.index) last[1] += token[0].length;
      else out.push([token.index, token.index + token[0].length]);
    }
    return out;
  };
  return { old: ranges(x, kept.old), new: ranges(y, kept.new) };
}

/** Whitespace at the end of a line, before any carriage return. */
export function trailing(text: string): Range | null {
  const match = text.match(/[ \t]+(?=\r?$)/);
  return match ? [match.index!, match.index! + match[0].length] : null;
}

/** A short, stable fingerprint of a text. */
function fingerprint(text: string) {
  let h1 = 0xdeadbeef;
  let h2 = 0x41c6ce57;
  for (let i = 0; i < text.length; i++) {
    const c = text.charCodeAt(i);
    h1 = Math.imul(h1 ^ c, 2654435761);
    h2 = Math.imul(h2 ^ c, 1597334677);
  }
  h1 = Math.imul(h1 ^ (h1 >>> 16), 2246822507) ^ Math.imul(h2 ^ (h2 >>> 13), 3266489909);
  h2 = Math.imul(h2 ^ (h2 >>> 16), 2246822507) ^ Math.imul(h1 ^ (h1 >>> 13), 3266489909);
  return (4294967296 * (2097151 & h2) + (h1 >>> 0)).toString(36);
}

/** A file's change by its path and a fingerprint of all of it, so marks of having viewed it lapse once it changes again. */
export const fileKey = (file: FileDiff) => `${file.path}\0${fingerprint(JSON.stringify(file))}`;

/** `files` for a view that has the changes `known` names already, which so cost it only their keys. */
export function sending(files: readonly FileDiff[], known: readonly string[]): Sent[] {
  const has = new Set(known);
  return files.map((file) => {
    const key = fileKey(file);
    return { key, diff: has.has(key) ? null : file };
  });
}