lms-sb/src/app/(student)/wrapped/_client/parse.worker.ts

// Web Worker: парсит .md-тексты в агрегаты (off-main, без фриза UI).
// Получает RawNote[], возвращает Stats. Тексты заметок дальше воркера не идут.

import type { RawNote, Stats, Period, Archetype, GraphData } from "./types";

const WIKILINK = /\[\[([^\]|#]+)/g;
const TAG = /(^|\s)#([\p{L}\d_/-]+)/gu;
const DAY = 864e5;

function filterByPeriod(notes: RawNote[], period: Period): RawNote[] {
  if (period === "all") return notes;
  const cutoff = Date.now() - (period === "year" ? 365 * DAY : 90 * DAY);
  return notes.filter((n) => n.mtime >= cutoff);
}

function buildGraph(degree: Map<string, number>, edges: { source: string; target: string }[]): GraphData {
  const TOP = 150;
  const top = [...degree.entries()].sort((a, b) => b[1] - a[1]).slice(0, TOP);
  const nodes = top.map(([id, deg]) => ({ id, deg }));
  const ids = new Set(nodes.map((n) => n.id));
  const links = edges.filter((e) => ids.has(e.source) && ids.has(e.target)).slice(0, 600);
  return { nodes, links };
}

function buildStats(
  notes: RawNote[],
  degree: Map<string, number>,
  tags: Map<string, number>,
  linkCount: number,
  period: Period,
  edges: { source: string; target: string }[],
): Stats {
  const noteCount = notes.length;
  const topTags = [...tags.entries()].sort((a, b) => b[1] - a[1]).slice(0, 5).map(([tag, n]) => ({ tag, n }));
  const topLinked = [...degree.entries()].sort((a, b) => b[1] - a[1]).slice(0, 5).map(([name, d]) => ({ name, degree: d }));

  // Сирота — заметка, на которую никто не ссылается (грубая оценка по имени).
  const linkedTargets = new Set(degree.keys());
  const orphans = notes.filter((n) => !linkedTargets.has(n.name.replace(/\.md$/i, ""))).length;

  const months = new Map<string, number>();
  for (const n of notes) {
    const d = new Date(n.mtime);
    const k = `${d.getFullYear()}-${String(d.getMonth() + 1).padStart(2, "0")}`;
    months.set(k, (months.get(k) ?? 0) + 1);
  }
  const busiestMonth = [...months.entries()].sort((a, b) => b[1] - a[1])[0]?.[0] ?? "";

  const linkDensity = noteCount ? linkCount / noteCount : 0;
  const tagDensity = noteCount ? tags.size / noteCount : 0;
  const orphanRatio = noteCount ? orphans / noteCount : 1;

  let archetype: Archetype;
  if (linkDensity >= 2) archetype = "architect";
  else if (tagDensity >= 0.5) archetype = "librarian";
  else if (orphanRatio >= 0.4) archetype = "gardener";
  else archetype = "pragmatist";

  const score = Math.max(0, Math.min(100, Math.round(
    Math.min(linkDensity / 3, 1) * 45 + Math.min(tagDensity, 1) * 25 + Math.min(noteCount / 300, 1) * 30,
  )));

  const achievements: string[] = [];
  if (noteCount >= 100) achievements.push(`${noteCount} заметок`);
  if (linkCount >= 200) achievements.push(`${linkCount} связей`);
  if (tags.size >= 30) achievements.push(`${tags.size} тегов`);
  if (linkDensity >= 3) achievements.push("Плотная сеть знаний");
  if (orphanRatio < 0.1 && noteCount >= 50) achievements.push("Почти нет сирот");

  return { noteCount, linkCount, tagCount: tags.size, topTags, topLinked, orphans, busiestMonth, achievements, archetype, score, period, graph: buildGraph(degree, edges) };
}

self.onmessage = (e: MessageEvent<{ notes: RawNote[]; period: Period }>) => {
  const { notes, period } = e.data;
  const filtered = filterByPeriod(notes, period);
  const degree = new Map<string, number>();
  const tags = new Map<string, number>();
  let linkCount = 0;
  const edges: { source: string; target: string }[] = [];
  for (const n of filtered) {
    const src = n.name.replace(/\.md$/i, "");
    for (const m of n.text.matchAll(WIKILINK)) {
      linkCount++;
      const t = m[1].trim();
      if (t) {
        degree.set(t, (degree.get(t) ?? 0) + 1);
        edges.push({ source: src, target: t });
      }
    }
    for (const m of n.text.matchAll(TAG)) {
      const t = m[2];
      tags.set(t, (tags.get(t) ?? 0) + 1);
    }
  }
  self.postMessage(buildStats(filtered, degree, tags, linkCount, period, edges));
};