stats §8: CBS Curaçao collector (topic-page scrape, sha256 manifest, CPI since 1970 from PDF, GDP by industry, government finance, population since 1900, LFS, tourism) into cw_cbs_* + stats.obs region CUW; cw_cbcs_bop for manual uploads; weekly timer

scripts/stats/xlsx.ts

@@ -0,0 +1,78 @@
+// Minimal xlsx reader for the stats collector (no dependency): unzips the workbook with the
+// host's `unzip` and reads cell values from the sheet XML. Enough for hand-formatted statistical
+// tables: returns each sheet as a grid of strings/numbers (merged cells are not expanded).
+export type Cell = string | number | null;
+export type Sheet = { name: string; rows: Cell[][] };
+
+async function unzipEntry(path: string, entry: string): Promise<string | null> {
+	const proc = Bun.spawn(['unzip', '-p', path, entry], { stdout: 'pipe', stderr: 'ignore' });
+	const text = await new Response(proc.stdout).text();
+	return (await proc.exited) === 0 ? text : null;
+}
+
+const decode = (s: string) =>
+	s
+		.replace(/&lt;/g, '<')
+		.replace(/&gt;/g, '>')
+		.replace(/&quot;/g, '"')
+		.replace(/&apos;/g, "'")
+		.replace(/&#(\d+);/g, (_, n) => String.fromCodePoint(Number(n)))
+		.replace(/&#x([0-9a-f]+);/gi, (_, n) => String.fromCodePoint(parseInt(n, 16)))
+		.replace(/&amp;/g, '&');
+
+/** Text of all <t> runs inside a fragment (shared strings and inline strings). */
+const textOf = (xml: string) =>
+	decode([...xml.matchAll(/<t(?:\s[^>]*)?>([\s\S]*?)<\/t>/g)].map((m) => m[1]).join(''));
+
+function colIndex(ref: string) {
+	let n = 0;
+	for (const ch of ref.replace(/\d+$/, '')) n = n * 26 + (ch.charCodeAt(0) - 64);
+	return n - 1;
+}
+
+export async function readXlsx(path: string): Promise<Sheet[]> {
+	const workbook = await unzipEntry(path, 'xl/workbook.xml');
+	const rels = await unzipEntry(path, 'xl/_rels/workbook.xml.rels');
+	if (!workbook || !rels) throw new Error('not an xlsx workbook');
+	const shared = await unzipEntry(path, 'xl/sharedStrings.xml');
+	const strings = shared
+		? [...shared.matchAll(/<si>([\s\S]*?)<\/si>/g)].map((m) => textOf(m[1]))
+		: [];
+	const target = new Map(
+		[...rels.matchAll(/<Relationship\b[^>]*>/g)].map((m) => [
+			m[0].match(/Id="([^"]+)"/)?.[1],
+			m[0].match(/Target="([^"]+)"/)?.[1]
+		])
+	);
+	const sheets: Sheet[] = [];
+	for (const m of workbook.matchAll(/<sheet\b[^>]*>/g)) {
+		const name = decode(m[0].match(/name="([^"]*)"/)?.[1] ?? '');
+		const rid = m[0].match(/r:id="([^"]+)"/)?.[1];
+		let file = target.get(rid) ?? '';
+		file = file.startsWith('/') ? file.slice(1) : `xl/${file}`;
+		const xml = await unzipEntry(path, file);
+		if (!xml) continue;
+		const rows: Cell[][] = [];
+		for (const r of xml.matchAll(/<row\b[^>]*?r="(\d+)"[^>]*?(?:\/>|>([\s\S]*?)<\/row>)/g)) {
+			const row: Cell[] = [];
+			for (const c of (r[2] ?? '').matchAll(/<c\b([^>]*?)(?:\/>|>([\s\S]*?)<\/c>)/g)) {
+				const ref = c[1].match(/r="([A-Z]+\d+)"/)?.[1];
+				if (!ref) continue;
+				const type = c[1].match(/t="(\w+)"/)?.[1];
+				const body = c[2] ?? '';
+				const v = body.match(/<v>([\s\S]*?)<\/v>/)?.[1];
+				let value: Cell = null;
+				if (type === 's' && v != null) value = strings[Number(v)] ?? null;
+				else if (type === 'inlineStr') value = textOf(body);
+				else if (type === 'str' || type === 'e') value = v != null ? decode(v) : null;
+				else if (type === 'b') value = v === '1' ? 'TRUE' : 'FALSE';
+				else if (v != null && v !== '') value = Number(v);
+				if (typeof value === 'string') value = value.trim() || null;
+				row[colIndex(ref)] = value;
+			}
+			rows[Number(r[1]) - 1] = Array.from(row, (x) => x ?? null);
+		}
+		sheets.push({ name, rows: Array.from(rows, (x) => x ?? []) });
+	}
+	return sheets;
+}