stats B11: EP-Online energy labels (monthly total file → homes per label for NL, Gouda, Gouda PC4s; aggregates only), stats.ep_label_agg, monthly timer with key from ~/.config/stats/ep-online.env, charts on /stats/energy and /stats/gouda

2026-10-25 21:28:37 +00:00
parent 2a4466cb82
commit fa58e5c661
15 changed files with 33573 additions and 6 deletions
--- a/scripts/stats/eponline.ts
+++ b/scripts/stats/eponline.ts
@@ -0,0 +1,238 @@
+// B11 EP-Online (RVO energy labels): the monthly total file ("mutatiebestand", CSV in a zip,
+// ~1.6 GB unpacked, every currently registered label) aggregated to the number of homes per
+// energy label for the Netherlands, Gouda (postcodes 2800–2809) and each Gouda PC4. Only these
+// aggregates are stored; no address-level data leaves the run. Needs EP_ONLINE_API_KEY (loaded
+// by the runner from ~/.config/stats/ep-online.env; never logged). One snapshot per publication
+// date, kept over time, so the shares become a monthly series.
+import { mkdtemp, rm } from 'node:fs/promises';
+import { tmpdir } from 'node:os';
+import { join } from 'node:path';
+import {
+	http,
+	insertChunks,
+	log,
+	run,
+	sql,
+	writeSeries,
+	type ObsRow,
+	type SeriesRow
+} from './common';
+
+const API = 'https://public.ep-online.nl/api/v5';
+/** Best to worst (NTA 8800 scale since 2021). */
+export const LABELS = ['A++++', 'A+++', 'A++', 'A+', 'A', 'B', 'C', 'D', 'E', 'F', 'G'] as const;
+const GOUDA_PC4 = Array.from({ length: 10 }, (_, i) => String(2800 + i));
+const AREA_NAME: Record<string, string> = {
+	NL: 'Netherlands',
+	GOUDA: 'Gouda (postcodes 2800–2809)',
+	...Object.fromEntries(GOUDA_PC4.map((p) => [p, `Gouda ${p}`]))
+};
+const AREA_REGION: Record<string, string> = { NL: 'NLD', GOUDA: 'GM0513' };
+
+/** CSV split on ';' with optional "…" quoting ("" escapes). */
+function splitLine(line: string): string[] {
+	if (!line.includes('"')) return line.split(';');
+	const out: string[] = [];
+	let cur = '';
+	let q = false;
+	for (let i = 0; i < line.length; i++) {
+		const c = line[i];
+		if (q && c === '"' && line[i + 1] === '"') {
+			cur += '"';
+			i++;
+		} else if (c === '"') {
+			q = !q;
+		} else if (!q && c === ';') {
+			out.push(cur);
+			cur = '';
+		} else {
+			cur += c;
+		}
+	}
+	out.push(cur);
+	return out;
+}
+
+async function* lines(stream: ReadableStream<Uint8Array>) {
+	const decoder = new TextDecoder();
+	let buf = '';
+	for await (const chunk of stream) {
+		buf += decoder.decode(chunk, { stream: true });
+		let nl: number;
+		let start = 0;
+		while ((nl = buf.indexOf('\n', start)) >= 0) {
+			yield buf.slice(start, nl).replace(/\r$/, '');
+			start = nl + 1;
+		}
+		buf = buf.slice(start);
+	}
+	if (buf) yield buf;
+}
+
+/** Streams the CSV out of the zip and counts homes per label and area (one label per home). */
+export async function aggregate(zipPath: string) {
+	const proc = Bun.spawn(['unzip', '-p', zipPath], { stdout: 'pipe', stderr: 'ignore' });
+	const it = lines(proc.stdout as ReadableStream<Uint8Array>);
+	let published: string | null = null;
+	let ix: Record<string, number> | null = null;
+	// Newest registration per home: BAG verblijfsobject id, or the address when there is none.
+	// Value packs label index (4 bits) + Gouda PC4 slot (4 bits, 15 = elsewhere) + yyyymmdd.
+	const byVbo = new Map<number, number>();
+	const byAddr = new Map<string, number>();
+	let rows = 0;
+	let skipped = 0;
+	for await (const line of it) {
+		if (!ix) {
+			const m = line.match(/^PublicatieDatum;(\d\d)-(\d\d)-(\d{4})/);
+			if (m) published = `${m[3]}-${m[2]}-${m[1]}`;
+			if (line.startsWith('Registratiedatum;'))
+				ix = Object.fromEntries(line.split(';').map((h, i) => [h, i]));
+			continue;
+		}
+		if (!line) continue;
+		rows++;
+		const f = splitLine(line);
+		if (f[ix.Gebouwklasse] !== 'W') continue;
+		// Labels issued on a building-permit design are not a home's label yet.
+		if (f[ix.Status]?.startsWith('Vergunningsaanvraag')) {
+			skipped++;
+			continue;
+		}
+		const label = LABELS.indexOf(f[ix.Energieklasse] as (typeof LABELS)[number]);
+		if (label < 0) {
+			skipped++;
+			continue;
+		}
+		const pc = f[ix.Postcode] ?? '';
+		const slot = pc >= '2800' && pc < '2810' ? Number(pc[3]) : 15;
+		const date = Number(f[ix.Registratiedatum]) || 0;
+		const packed = date * 256 + slot * 16 + label;
+		const vbo = f[ix.BAGVerblijfsobjectID];
+		if (vbo && /^\d{16}$/.test(vbo)) {
+			const key = Number(vbo);
+			const prev = byVbo.get(key);
+			if (prev === undefined || prev < packed) byVbo.set(key, packed);
+		} else {
+			const key = [pc, f[ix.Huisnummer], f[ix.Huisletter], f[ix.Huisnummertoevoeging]].join('|');
+			const prev = byAddr.get(key);
+			if (prev === undefined || prev < packed) byAddr.set(key, packed);
+		}
+	}
+	const code = await proc.exited;
+	if (code !== 0) throw new Error(`unzip exited ${code}`);
+	if (!ix || !published) throw new Error('unexpected EP-Online file header');
+	const counts = new Map<string, number[]>();
+	const bump = (area: string, label: number) => {
+		let c = counts.get(area);
+		if (!c) counts.set(area, (c = LABELS.map(() => 0)));
+		c[label]++;
+	};
+	for (const v of [...byVbo.values(), ...byAddr.values()]) {
+		const label = v % 16;
+		const slot = Math.floor(v / 16) % 16;
+		bump('NL', label);
+		if (slot !== 15) {
+			bump('GOUDA', label);
+			bump(String(2800 + slot), label);
+		}
+	}
+	return { published, counts, rows, skipped, homes: byVbo.size + byAddr.size };
+}
+
+export async function eponline(opts: { force: boolean; job: string }) {
+	await run('eponline', opts.job, async (note) => {
+		const key = process.env.EP_ONLINE_API_KEY;
+		if (!key) throw new Error('EP_ONLINE_API_KEY is not set (see ~/.config/stats/ep-online.env)');
+		const info = (await (
+			await http(`${API}/Mutatiebestand/DownloadInfo?fileType=csv`, {
+				headers: { authorization: key, accept: 'application/json' }
+			})
+		).json()) as { bestandsnaam: string; downloadUrl: string };
+		const m = info.bestandsnaam?.match(/v(\d{4})(\d\d)(\d\d)/);
+		if (!m || !info.downloadUrl) throw new Error('DownloadInfo returned no file');
+		const snapshot = `${m[1]}-${m[2]}-${m[3]}`;
+		const [{ have }] = await sql`
+			select count(*)::int as have from stats.ep_label_agg where snapshot = ${snapshot}`;
+		if (have && !opts.force) {
+			log(`eponline: ${info.bestandsnaam} already loaded`);
+			await rebuildSeries();
+			return 0;
+		}
+		const dir = await mkdtemp(join(tmpdir(), 'eponline-'));
+		try {
+			const zip = join(dir, 'labels.zip');
+			// The download URL carries a token: never put it in logs or fetch_run errors.
+			const res = await fetch(info.downloadUrl, {
+				headers: { 'user-agent': 'stanrunge.dev-stats/1.0' },
+				signal: AbortSignal.timeout(30 * 60_000)
+			}).catch((e) => {
+				throw new Error(`download failed (${e instanceof Error ? e.name : 'error'})`);
+			});
+			if (!res.ok || !res.body) throw new Error(`download failed: HTTP ${res.status}`);
+			await Bun.write(zip, res);
+			log(
+				`eponline: downloaded ${info.bestandsnaam} (${(Bun.file(zip).size / 1e6).toFixed(0)} MB)`
+			);
+			const agg = await aggregate(zip);
+			if (agg.published !== snapshot)
+				note(`file says published ${agg.published}, name says ${snapshot}`);
+			const nl = agg.counts.get('NL');
+			if (!nl || nl.reduce((a, b) => a + b, 0) < 1_000_000)
+				throw new Error(`implausible total: ${agg.homes} homes`);
+			const out: Record<string, unknown>[] = [];
+			for (const [area, c] of agg.counts)
+				c.forEach((homes, i) => out.push({ snapshot, area, label: LABELS[i], homes }));
+			await sql.begin(async (tx) => {
+				await tx`delete from stats.ep_label_agg where snapshot = ${snapshot}`;
+				await insertChunks(tx, 'ep_label_agg', out);
+			});
+			log(
+				`eponline: ${agg.rows} rows, ${agg.homes} homes with a label (${agg.skipped} permit/blank skipped)`
+			);
+		} finally {
+			await rm(dir, { recursive: true, force: true });
+		}
+		return rebuildSeries();
+	});
+}
+
+/** stats.obs: share of homes per label (%), per area, one point per monthly snapshot. */
+async function rebuildSeries(): Promise<number> {
+	const rows = await sql<{ snapshot: string; area: string; label: string; homes: number }[]>`
+		select snapshot::text, area, label, homes from stats.ep_label_agg`;
+	const totals = new Map<string, number>();
+	for (const r of rows)
+		totals.set(`${r.snapshot}|${r.area}`, (totals.get(`${r.snapshot}|${r.area}`) ?? 0) + r.homes);
+	const series: SeriesRow[] = [];
+	const obs: ObsRow[] = [];
+	const areas = [...new Set(rows.map((r) => r.area))];
+	for (const area of areas)
+		for (const label of [...LABELS, 'TOTAL']) {
+			const id = `eponline:${label === 'TOTAL' ? 'homes' : `share:${label}`}:${area}`;
+			series.push({
+				series_id: id,
+				source_id: 'eponline',
+				source_code: `${area}:${label}`,
+				title:
+					label === 'TOTAL'
+						? `Homes with a registered energy label: ${AREA_NAME[area] ?? area}`
+						: `Share of labelled homes with label ${label}: ${AREA_NAME[area] ?? area}`,
+				unit: label === 'TOTAL' ? 'homes' : '%',
+				freq: 'M',
+				region_code: AREA_REGION[area] ?? null,
+				theme: 'energy',
+				meta: { area, label }
+			});
+			for (const r of rows)
+				if (r.area === area && (label === 'TOTAL' ? r.label === LABELS[0] : r.label === label)) {
+					const total = totals.get(`${r.snapshot}|${area}`) ?? 0;
+					obs.push({
+						series_id: id,
+						period: `${r.snapshot.slice(0, 7)}-01`,
+						value: label === 'TOTAL' ? total : total ? (100 * r.homes) / total : null,
+						status: null
+					});
+				}
+		}
+	return writeSeries(series, obs);
+}