implemented a lot of features

lib/spotnet/xml.ts

@@ -0,0 +1,84 @@
+import { inflateRawSync } from 'node:zlib'
+
+export interface SpotXml {
+  title: string
+  category: number | null
+  subcats: string[]
+  size: number
+  website: string
+  /** Message ids (without <>) of the compressed NZB in alt.binaries.ftd. */
+  nzbSegments: string[]
+  created: number | null
+}
+
+const ENTITIES: Record<string, string> = { amp: '&', lt: '<', gt: '>', quot: '"', apos: "'" }
+
+function text(value: string | undefined): string {
+  if (value === undefined) return ''
+  const cdata = /^\s*<!\[CDATA\[([\s\S]*)\]\]>\s*$/.exec(value)
+  if (cdata) return cdata[1]
+  return value.replace(/&(#x[0-9a-f]+|#\d+|[a-z]+);/gi, (match, entity: string) => {
+    if (entity[0] === '#') {
+      const code = entity[1].toLowerCase() === 'x' ? Number.parseInt(entity.slice(2), 16) : Number.parseInt(entity.slice(1), 10)
+      return Number.isFinite(code) && code > 0 && code <= 0x10ffff ? String.fromCodePoint(code) : match
+    }
+    return ENTITIES[entity.toLowerCase()] ?? match
+  })
+}
+
+const element = (xml: string, name: string) => new RegExp(`<${name}(?:\\s[^>]*)?>([\\s\\S]*?)</${name}>`, 'i').exec(xml)?.[1]
+
+/** The spot XML from the X-XML header lines (a tolerant scanner, as spots are often invalid XML). */
+export function parseSpotXml(xml: string): SpotXml | null {
+  const posting = element(xml, 'Posting')
+  if (!posting) return null
+  const categoryXml = element(posting, 'Category') ?? ''
+  const key = Number.parseInt(element(posting, 'Key') ?? '', 10)
+  const rawCategory = Number.parseInt(categoryXml.replace(/<[\s\S]*$/, ''), 10)
+  const nzb = element(posting, 'NZB') ?? ''
+  return {
+    title: text(element(posting, 'Title')).trim(),
+    // Newer keys number categories from 1.
+    category: Number.isNaN(rawCategory) ? null : key !== 1 ? rawCategory - 1 : rawCategory,
+    subcats: [...categoryXml.matchAll(/<(?:Sub|SubCat)>([^<]*)<\/(?:Sub|SubCat)>/gi)].map((m) => m[1].trim()),
+    size: Number.parseInt(element(posting, 'Size') ?? '', 10) || 0,
+    website: text(element(posting, 'Website')).trim(),
+    nzbSegments: [...nzb.matchAll(/<Segment>([^<>]*)<\/Segment>/gi)].map((m) => m[1].trim()).filter(Boolean),
+    created: Number.parseInt(element(posting, 'Created') ?? '', 10) || null,
+  }
+}
+
+/** Header lines of a raw HEAD response; X-XML (split over many lines) joined. */
+export function parseSpotArticleHeaders(raw: string): { headers: Record<string, string>; xml: string } {
+  const headers: Record<string, string> = {}
+  let xml = ''
+  for (const line of raw.split(/\r?\n/)) {
+    const colon = line.indexOf(':')
+    if (colon <= 0) continue
+    const name = line.slice(0, colon).toLowerCase()
+    const value = line.slice(colon + 1).replace(/^ /, '')
+    if (name === 'x-xml') xml += value
+    else headers[name] ??= value
+  }
+  return { headers, xml }
+}
+
+/**
+ * Spot binaries (the NZB, images): article bodies joined without line
+ * breaks, Spotnet's escapes undone (=A NUL, =B CR, =C LF, =D =), then raw
+ * deflate (PHP gzinflate).
+ */
+export function decodeSpotBinary(bodies: Buffer[], compressed = true): Buffer {
+  const joined = bodies
+    .map((body) =>
+      body
+        .toString('latin1')
+        .split('\r\n')
+        // NNTP dot-stuffing.
+        .map((line) => (line.startsWith('..') ? line.slice(1) : line))
+        .join('')
+    )
+    .join('')
+  const unescaped = Buffer.from(joined.replace(/=([ABCD])/g, (_, c: string) => ({ A: '\0', B: '\r', C: '\n', D: '=' })[c]!), 'latin1')
+  return compressed ? inflateRawSync(unescaped) : unescaped
+}