Webklar-Kundenbereich/analytics/src/stats.js

// Statistik für das Portal: Vergangenheit aus `daily`, noch nicht verdichtete Tage live aus den Rohdaten.
import { dayOf, aggregateRows } from './jobs.js';

const TOP = 20;

// Kalendertag (YYYY-MM-DD) um `n` Tage verschoben, rein kalendarisch (UTC-Datumsrechnung, keine Zeitumstellung).
function shiftDay(day, n) {
  const [y, m, d] = day.split('-').map(Number);
  return new Date(Date.UTC(y, m - 1, d + n)).toISOString().slice(0, 10);
}

// Liste der Kalendertage (YYYY-MM-DD), oldest first, endend am Tag `endDay`.
function dayList(endDay, range) {
  const days = [];
  for (let i = range - 1; i >= 0; i--) days.push(shiftDay(endDay, -i));
  return days;
}

const top = (map) => [...map].map(([key, count]) => ({ key, count })).filter((e) => e.count > 0)
  .sort((a, b) => b.count - a.count || (a.key < b.key ? -1 : 1)).slice(0, TOP);

// Sammelt die Zähler für eine zusammenhängende Tagesliste: verdichtete Tage aus `daily`,
// Tage ohne agg_state (heute, ggf. gestern vor 02:30) live aus den Rohdaten. Kein Tag zählt doppelt.
function collect(db, site, days, now) {
  const first = days[0];
  const last = days[days.length - 1];
  const inRange = new Set(days);
  const perDay = new Map(days.map((d) => [d, { pv: 0, visitors: 0, contacts: 0, en: 0, sec: 0, scroll: 0 }]));
  const dims = {};
  const apply = (day, dim, key, n) => {
    const t = perDay.get(day);
    if (dim === 'pv_total') t.pv += n;
    else if (dim === 'visitors') t.visitors += n;
    else if (dim === 'ct_kind') t.contacts += n;
    else if (dim === 'en_count') t.en += n;
    else if (dim === 'en_sec_sum') t.sec += n;
    else if (dim === 'scroll_sum') t.scroll += n;
    if (!dims[dim]) dims[dim] = new Map();
    dims[dim].set(key, (dims[dim].get(key) || 0) + n);
  };

  const done = new Set(db.prepare('SELECT day FROM agg_state WHERE site = ? AND day >= ? AND day <= ?').all(site, first, last).map((r) => r.day));
  for (const r of db.prepare('SELECT day, dim, key, count FROM daily WHERE site = ? AND day >= ? AND day <= ?').all(site, first, last)) {
    if (inRange.has(r.day)) apply(r.day, r.dim, r.key, r.count);
  }
  // Ein Berliner Tag beginnt 22:00/23:00 UTC des Vortags: zwei Tage Rand reichen auf beiden Seiten, danach filtert dayOf
  const [fy, fm, fd] = first.split('-').map(Number);
  const [ly, lm, ld] = last.split('-').map(Number);
  const lo = Date.UTC(fy, fm - 1, fd - 2);
  const hi = Math.min(now, Date.UTC(ly, lm - 1, ld + 2));
  const live = new Map();
  for (const r of db.prepare('SELECT ts, type, path, vp, ref, country, device, visitor, sec, scroll, kind FROM events_raw WHERE site = ? AND ts >= ? AND ts <= ? ORDER BY ts')
    .all(site, lo, hi)) {
    const d = dayOf(r.ts);
    if (!inRange.has(d) || done.has(d)) continue;
    if (!live.has(d)) live.set(d, []);
    live.get(d).push(r);
  }
  for (const [day, rows] of live) for (const [dim, key, n] of aggregateRows(rows)) apply(day, dim, key, n);
  return { days, perDay, dims };
}

const avg = (sum, n) => (n ? Math.round(sum / n) : null);

// Formt gesammelte Zähler zu banner + totals + series (gleiche Form für aktuellen Zeitraum und Vorzeitraum).
function shape({ days, perDay, dims }) {
  const sum = (dim) => [...(dims[dim] || new Map()).values()].reduce((a, b) => a + b, 0);
  const banner = { shown: sum('bn_shown'), accept: sum('bn_accept'), reject: sum('bn_reject') };
  const decided = banner.accept + banner.reject;
  const en = sum('en_count');
  return {
    banner,
    totals: {
      pv: sum('pv_total'),
      // Summe der Tages-Besucher (kein tagesübergreifender Abgleich möglich: Salz wechselt täglich)
      visitors: sum('visitors'),
      contacts: sum('ct_kind'),
      consentRate: decided ? banner.accept / decided : 0,
      engaged: en,
      avgSec: avg(sum('en_sec_sum'), en),
      avgScroll: avg(sum('scroll_sum'), en),
    },
    series: days.map((day) => {
      const t = perDay.get(day);
      return { day, pv: t.pv, visitors: t.visitors, contacts: t.contacts, avgSec: avg(t.sec, t.en) };
    }),
  };
}

// Frühester Berliner Tag mit Daten dieser Website ('' = keine): kleinster Tag aus daily und events_raw.
function firstDayOf(db, site) {
  const fromDaily = db.prepare('SELECT MIN(day) AS d FROM daily WHERE site = ?').get(site)?.d || '';
  const ts = db.prepare('SELECT MIN(ts) AS t FROM events_raw WHERE site = ?').get(site)?.t;
  const fromRaw = ts == null ? '' : dayOf(ts);
  if (!fromDaily) return fromRaw;
  if (!fromRaw) return fromDaily;
  return fromDaily < fromRaw ? fromDaily : fromRaw;
}

// Der Zeitraum sind die `range` abgeschlossenen Kalendertage, die gestern enden (wie GA4 "Letzte 28 Tage"): Der
// angebrochene heutige Tag würde gegen einen vollen Vorzeitraum als Einbruch erscheinen. Heute steht getrennt in `today`.
export function computeStats(db, site, range, now = Date.now()) {
  const todayDay = dayOf(now);
  const days = dayList(shiftDay(todayDay, -1), range);
  const cur = collect(db, site, days, now);
  const { banner, totals, series } = shape(cur);
  // Heute bis jetzt: live aus den Rohdaten (der Tag ist noch nicht verdichtet)
  const t = shape(collect(db, site, [todayDay], now)).series[0];
  // Vorzeitraum: die `range` Kalendertage direkt vor dem aktuellen Zeitraum (kalendarisch, nie now - range * 24 h)
  const prevDays = dayList(shiftDay(days[0], -1), range);
  const prevCollected = collect(db, site, prevDays, now);
  const prev = shape(prevCollected);
  const firstDay = firstDayOf(db, site);
  return {
    range,
    from: days[0],
    to: days[days.length - 1],
    firstDay,
    totals,
    series,
    today: { day: todayDay, pv: t.pv, visitors: t.visitors, contacts: t.contacts },
    pages: top(cur.dims.pv_path || new Map()),
    refs: top(cur.dims.pv_ref || new Map()),
    countries: top(cur.dims.pv_country || new Map()),
    devices: top(cur.dims.pv_device || new Map()),
    contacts: top(cur.dims.ct_kind || new Map()),
    banner,
    previous: {
      from: prevDays[0],
      to: prevDays[prevDays.length - 1],
      complete: firstDay !== '' && firstDay <= prevDays[0],
      totals: prev.totals,
      series: prev.series,
      contacts: top(prevCollected.dims.ct_kind || new Map()),
    },
  };
}