ConusAI vendor v0.1.0-dev.20261027+9b1131fb

modules/documents/__tests__/support/unzip.ts

@@ -0,0 +1,45 @@
+// Reads a ZIP the way a spreadsheet application does: from the central directory at the end. It is
+// the test's independent reader of what `buildWorkbook` wrote, so the writer is not checked by itself.
+
+import { inflateRawSync } from "node:zlib"
+
+export function unzip(bytes: Uint8Array): Map<string, string> {
+  const view = new DataView(bytes.buffer, bytes.byteOffset, bytes.byteLength)
+  let eocd = bytes.length - 22
+  while (eocd >= 0 && view.getUint32(eocd, true) !== 0x06054b50) eocd -= 1
+  if (eocd < 0) throw new Error("no end-of-central-directory record")
+  const entries = view.getUint16(eocd + 10, true)
+  let at = view.getUint32(eocd + 16, true)
+  const files = new Map<string, string>()
+  for (let i = 0; i < entries; i += 1) {
+    if (view.getUint32(at, true) !== 0x02014b50) throw new Error("a bad central directory entry")
+    const method = view.getUint16(at + 10, true)
+    const size = view.getUint32(at + 20, true)
+    const nameLength = view.getUint16(at + 28, true)
+    const extraLength = view.getUint16(at + 30, true)
+    const commentLength = view.getUint16(at + 32, true)
+    const local = view.getUint32(at + 42, true)
+    const name = new TextDecoder().decode(bytes.subarray(at + 46, at + 46 + nameLength))
+    const dataAt = local + 30 + view.getUint16(local + 26, true) + view.getUint16(local + 28, true)
+    const raw = bytes.subarray(dataAt, dataAt + size)
+    files.set(name, new TextDecoder().decode(method === 8 ? inflateRawSync(raw) : raw))
+    at += 46 + nameLength + extraLength + commentLength
+  }
+  return files
+}
+
+/** The text and number cells of the first sheet, row by row. */
+export function sheetRows(bytes: Uint8Array): Array<Array<string | number>> {
+  const sheet = unzip(bytes).get("xl/worksheets/sheet1.xml") ?? ""
+  return [...sheet.matchAll(/<row [^>]*>(.*?)<\/row>/gs)].map(([, row]) =>
+    [...(row as string).matchAll(/<c ([^>]*)>(.*?)<\/c>/gs)].map(([, attributes, body]) => {
+      if (!(attributes as string).includes('t="inlineStr"')) return Number(/<v>(.*?)<\/v>/s.exec(body as string)?.[1])
+      const text = /<t[^>]*>(.*?)<\/t>/s.exec(body as string)?.[1] ?? ""
+      return text
+        .replace(/&quot;/g, '"')
+        .replace(/&gt;/g, ">")
+        .replace(/&lt;/g, "<")
+        .replace(/&amp;/g, "&")
+    }),
+  )
+}