Initial catalog maker baseline

scripts/read-sql-sheet.mjs

@@ -0,0 +1,45 @@
+import fs from "node:fs";
+import { createRequire } from "node:module";
+
+const require = createRequire(import.meta.url);
+const JSZip = require("jszip");
+
+const workbookPath = process.argv[2] || "xxx_endpoint_bo_2026-08-04_v1.xlsm";
+const fromRow = Number.parseInt(process.argv[3] || "1", 10);
+const toRow = Number.parseInt(process.argv[4] || String(fromRow), 10);
+
+const zip = await JSZip.loadAsync(fs.readFileSync(workbookPath));
+const sharedStringsXml = await zip.file("xl/sharedStrings.xml").async("string");
+const sharedStrings = [...sharedStringsXml.matchAll(/<si[\s\S]*?<\/si>/g)].map((match) =>
+  decodeXml([...match[0].matchAll(/<t[^>]*>([\s\S]*?)<\/t>/g)].map((part) => part[1]).join("")),
+);
+
+const sqlSheetXml = await zip.file("xl/worksheets/sheet5.xml").async("string");
+
+for (const rowMatch of sqlSheetXml.matchAll(/<row[^>]* r="(\d+)"[\s\S]*?<\/row>/g)) {
+  const rowNumber = Number(rowMatch[1]);
+  if (rowNumber < fromRow || rowNumber > toRow) continue;
+
+  const cells = {};
+  for (const cellMatch of rowMatch[0].matchAll(
+    /<c[^>]* r="([A-Z]+)\d+"([^>]*)>([\s\S]*?)<\/c>/g,
+  )) {
+    const [, column, attributes, body] = cellMatch;
+    const inlineString = body.match(/<is>[\s\S]*?<t[^>]*>([\s\S]*?)<\/t>[\s\S]*?<\/is>/);
+    const rawValue = body.match(/<v>([\s\S]*?)<\/v>/)?.[1] ?? inlineString?.[1] ?? "";
+    cells[column] = attributes.includes('t="s"')
+      ? sharedStrings[Number(rawValue)] || ""
+      : decodeXml(rawValue);
+  }
+
+  console.log(JSON.stringify({ row: rowNumber, cells }, null, 2));
+}
+
+function decodeXml(value) {
+  return String(value ?? "")
+    .replaceAll("&lt;", "<")
+    .replaceAll("&gt;", ">")
+    .replaceAll("&amp;", "&")
+    .replaceAll("&quot;", '"')
+    .replaceAll("&apos;", "'");
+}