/**
 * Script to scrape Baileys documentation from baileys.wiki
 * Creates a comprehensive markdown file for Claude Code reference
 */

import * as fs from 'fs'

const BASE_URL = 'https://baileys.wiki/docs/api'

interface DocSection {
  name: string
  path: string
  type: 'interface' | 'type' | 'function' | 'variable' | 'class' | 'enum'
}

// Relevant documentation sections to scrape
const SECTIONS: DocSection[] = [
  // Interfaces
  { name: 'Contact', path: '/interfaces/Contact', type: 'interface' },
  { name: 'GroupMetadata', path: '/interfaces/GroupMetadata', type: 'interface' },
  { name: 'GroupModificationResponse', path: '/interfaces/GroupModificationResponse', type: 'interface' },
  { name: 'PresenceData', path: '/interfaces/PresenceData', type: 'interface' },
  { name: 'BaileysEventEmitter', path: '/interfaces/BaileysEventEmitter', type: 'interface' },
  { name: 'WAUrlInfo', path: '/interfaces/WAUrlInfo', type: 'interface' },
  { name: 'RecentMessage', path: '/interfaces/RecentMessage', type: 'interface' },

  // Type Aliases - Chat
  { name: 'Chat', path: '/type-aliases/Chat', type: 'type' },
  { name: 'ChatModification', path: '/type-aliases/ChatModification', type: 'type' },
  { name: 'ChatUpdate', path: '/type-aliases/ChatUpdate', type: 'type' },
  { name: 'ChatMutation', path: '/type-aliases/ChatMutation', type: 'type' },

  // Type Aliases - Messages
  { name: 'WAMessage', path: '/type-aliases/WAMessage', type: 'type' },
  { name: 'WAMessageContent', path: '/type-aliases/WAMessageContent', type: 'type' },
  { name: 'WAMessageKey', path: '/type-aliases/WAMessageKey', type: 'type' },
  { name: 'WAMessageUpdate', path: '/type-aliases/WAMessageUpdate', type: 'type' },
  { name: 'MessageType', path: '/type-aliases/MessageType', type: 'type' },
  { name: 'MessageUpsertType', path: '/type-aliases/MessageUpsertType', type: 'type' },
  { name: 'AnyMessageContent', path: '/type-aliases/AnyMessageContent', type: 'type' },
  { name: 'AnyMediaMessageContent', path: '/type-aliases/AnyMediaMessageContent', type: 'type' },
  { name: 'AnyRegularMessageContent', path: '/type-aliases/AnyRegularMessageContent', type: 'type' },
  { name: 'MessageGenerationOptions', path: '/type-aliases/MessageGenerationOptions', type: 'type' },
  { name: 'MessageContentGenerationOptions', path: '/type-aliases/MessageContentGenerationOptions', type: 'type' },
  { name: 'MinimalMessage', path: '/type-aliases/MinimalMessage', type: 'type' },

  // Type Aliases - Media
  { name: 'MediaType', path: '/type-aliases/MediaType', type: 'type' },
  { name: 'MediaDownloadOptions', path: '/type-aliases/MediaDownloadOptions', type: 'type' },
  { name: 'WAMediaUpload', path: '/type-aliases/WAMediaUpload', type: 'type' },
  { name: 'DownloadableMessage', path: '/type-aliases/DownloadableMessage', type: 'type' },
  { name: 'MediaGenerationOptions', path: '/type-aliases/MediaGenerationOptions', type: 'type' },

  // Type Aliases - Groups
  { name: 'GroupParticipant', path: '/type-aliases/GroupParticipant', type: 'type' },
  { name: 'GroupInviteInfo', path: '/type-aliases/GroupInviteInfo', type: 'type' },
  { name: 'ParticipantAction', path: '/type-aliases/ParticipantAction', type: 'type' },

  // Type Aliases - Presence & Status
  { name: 'WAPresence', path: '/type-aliases/WAPresence', type: 'type' },
  { name: 'ConnectionState', path: '/type-aliases/ConnectionState', type: 'type' },

  // Type Aliases - Events
  { name: 'BaileysEventMap', path: '/type-aliases/BaileysEventMap', type: 'type' },
  { name: 'BaileysEvent', path: '/type-aliases/BaileysEvent', type: 'type' },

  // Type Aliases - Auth
  { name: 'AuthenticationCreds', path: '/type-aliases/AuthenticationCreds', type: 'type' },
  { name: 'AuthenticationState', path: '/type-aliases/AuthenticationState', type: 'type' },

  // Type Aliases - Socket
  { name: 'WASocket', path: '/type-aliases/WASocket', type: 'type' },
  { name: 'SocketConfig', path: '/type-aliases/SocketConfig', type: 'type' },
  { name: 'UserFacingSocketConfig', path: '/type-aliases/UserFacingSocketConfig', type: 'type' },

  // Functions - Messages
  { name: 'generateWAMessage', path: '/functions/generateWAMessage', type: 'function' },
  { name: 'generateWAMessageContent', path: '/functions/generateWAMessageContent', type: 'function' },
  { name: 'generateWAMessageFromContent', path: '/functions/generateWAMessageFromContent', type: 'function' },
  { name: 'extractMessageContent', path: '/functions/extractMessageContent', type: 'function' },
  { name: 'getContentType', path: '/functions/getContentType', type: 'function' },
  { name: 'normalizeMessageContent', path: '/functions/normalizeMessageContent', type: 'function' },

  // Functions - Media
  { name: 'downloadMediaMessage', path: '/functions/downloadMediaMessage', type: 'function' },
  { name: 'downloadContentFromMessage', path: '/functions/downloadContentFromMessage', type: 'function' },
  { name: 'prepareWAMessageMedia', path: '/functions/prepareWAMessageMedia', type: 'function' },
  { name: 'extractImageThumb', path: '/functions/extractImageThumb', type: 'function' },
  { name: 'generateThumbnail', path: '/functions/generateThumbnail', type: 'function' },
  { name: 'generateProfilePicture', path: '/functions/generateProfilePicture', type: 'function' },
  { name: 'getMediaKeys', path: '/functions/getMediaKeys', type: 'function' },
  { name: 'encryptedStream', path: '/functions/encryptedStream', type: 'function' },

  // Functions - JID
  { name: 'jidDecode', path: '/functions/jidDecode', type: 'function' },
  { name: 'jidEncode', path: '/functions/jidEncode', type: 'function' },
  { name: 'jidNormalizedUser', path: '/functions/jidNormalizedUser', type: 'function' },
  { name: 'isJidGroup', path: '/functions/isJidGroup', type: 'function' },
  { name: 'isJidBroadcast', path: '/functions/isJidBroadcast', type: 'function' },
  { name: 'isJidStatusBroadcast', path: '/functions/isJidStatusBroadcast', type: 'function' },
  { name: 'areJidsSameUser', path: '/functions/areJidsSameUser', type: 'function' },
  { name: 'isLidUser', path: '/functions/isLidUser', type: 'function' },

  // Functions - History
  { name: 'downloadAndProcessHistorySyncNotification', path: '/functions/downloadAndProcessHistorySyncNotification', type: 'function' },
  { name: 'processHistoryMessage', path: '/functions/processHistoryMessage', type: 'function' },
  { name: 'getHistoryMsg', path: '/functions/getHistoryMsg', type: 'function' },

  // Functions - Socket
  { name: 'makeWASocket', path: '/functions/makeWASocket', type: 'function' },
  { name: 'makeCacheableSignalKeyStore', path: '/functions/makeCacheableSignalKeyStore', type: 'function' },
  { name: 'useMultiFileAuthState', path: '/functions/useMultiFileAuthState', type: 'function' },

  // Functions - Utility
  { name: 'delay', path: '/functions/delay', type: 'function' },
  { name: 'toNumber', path: '/functions/toNumber', type: 'function' },
  { name: 'toBuffer', path: '/functions/toBuffer', type: 'function' },
  { name: 'unixTimestampSeconds', path: '/functions/unixTimestampSeconds', type: 'function' },

  // Variables
  { name: 'WAMessageStatus', path: '/variables/WAMessageStatus', type: 'variable' },
  { name: 'WAMessageStubType', path: '/variables/WAMessageStubType', type: 'variable' },
  { name: 'S_WHATSAPP_NET', path: '/variables/S_WHATSAPP_NET', type: 'variable' },
  { name: 'Browsers', path: '/variables/Browsers', type: 'variable' },
  { name: 'MEDIA_KEYS', path: '/variables/MEDIA_KEYS', type: 'variable' },
  { name: 'DEFAULT_CONNECTION_CONFIG', path: '/variables/DEFAULT_CONNECTION_CONFIG', type: 'variable' },

  // Enums
  { name: 'DisconnectReason', path: '/enumerations/DisconnectReason', type: 'enum' },
]

async function fetchPage(url: string): Promise<string> {
  const response = await fetch(url)
  if (!response.ok) {
    throw new Error(`Failed to fetch ${url}: ${response.status}`)
  }
  return response.text()
}

function extractContent(html: string, name: string, type: string): string {
  // Remove script tags and style tags
  html = html.replace(/<script[^>]*>[\s\S]*?<\/script>/gi, '')
  html = html.replace(/<style[^>]*>[\s\S]*?<\/style>/gi, '')
  html = html.replace(/<nav[^>]*>[\s\S]*?<\/nav>/gi, '')
  html = html.replace(/<footer[^>]*>[\s\S]*?<\/footer>/gi, '')
  html = html.replace(/<header[^>]*>[\s\S]*?<\/header>/gi, '')
  html = html.replace(/<aside[^>]*>[\s\S]*?<\/aside>/gi, '')

  // Extract main/article content
  let mainMatch = html.match(/<main[^>]*>([\s\S]*?)<\/main>/i)
  if (!mainMatch) {
    mainMatch = html.match(/<article[^>]*>([\s\S]*?)<\/article>/i)
  }
  const content = mainMatch ? mainMatch[1] : html

  // Convert HTML to markdown-like text
  let text = content
    // Headers
    .replace(/<h1[^>]*>([\s\S]*?)<\/h1>/gi, '\n# $1\n')
    .replace(/<h2[^>]*>([\s\S]*?)<\/h2>/gi, '\n## $1\n')
    .replace(/<h3[^>]*>([\s\S]*?)<\/h3>/gi, '\n### $1\n')
    .replace(/<h4[^>]*>([\s\S]*?)<\/h4>/gi, '\n#### $1\n')
    // Code blocks
    .replace(/<pre[^>]*><code[^>]*>([\s\S]*?)<\/code><\/pre>/gi, '\n```typescript\n$1\n```\n')
    .replace(/<code[^>]*>([\s\S]*?)<\/code>/gi, '`$1`')
    // Lists
    .replace(/<li[^>]*>([\s\S]*?)<\/li>/gi, '- $1\n')
    .replace(/<ul[^>]*>/gi, '\n')
    .replace(/<\/ul>/gi, '\n')
    // Paragraphs
    .replace(/<p[^>]*>([\s\S]*?)<\/p>/gi, '\n$1\n')
    // Links - keep only text for cleaner output
    .replace(/<a[^>]*>([\s\S]*?)<\/a>/gi, '$1')
    // Bold/Italic
    .replace(/<strong[^>]*>([\s\S]*?)<\/strong>/gi, '**$1**')
    .replace(/<em[^>]*>([\s\S]*?)<\/em>/gi, '*$1*')
    // Line breaks
    .replace(/<br\s*\/?>/gi, '\n')
    // Divs and spans
    .replace(/<div[^>]*>/gi, '\n')
    .replace(/<\/div>/gi, '\n')
    .replace(/<span[^>]*>/gi, '')
    .replace(/<\/span>/gi, '')
    // Tables (simplified)
    .replace(/<table[^>]*>/gi, '\n')
    .replace(/<\/table>/gi, '\n')
    .replace(/<tr[^>]*>/gi, '')
    .replace(/<\/tr>/gi, '\n')
    .replace(/<td[^>]*>([\s\S]*?)<\/td>/gi, '| $1 ')
    .replace(/<th[^>]*>([\s\S]*?)<\/th>/gi, '| **$1** ')
    // Remove remaining tags
    .replace(/<[^>]+>/g, '')
    // Decode HTML entities
    .replace(/&lt;/g, '<')
    .replace(/&gt;/g, '>')
    .replace(/&amp;/g, '&')
    .replace(/&quot;/g, '"')
    .replace(/&#39;/g, "'")
    .replace(/&nbsp;/g, ' ')
    // Clean up whitespace
    .replace(/\n\s*\n\s*\n/g, '\n\n')
    .replace(/^\s+|\s+$/g, '')

  return text
}

async function scrapeSection(section: DocSection): Promise<string> {
  const url = `${BASE_URL}${section.path}`
  console.log(`Fetching ${section.name}...`)

  try {
    const html = await fetchPage(url)
    const content = extractContent(html, section.name, section.type)

    return `
---

## ${section.type.charAt(0).toUpperCase() + section.type.slice(1)}: ${section.name}

**Source:** ${url}

${content}
`
  } catch (error) {
    console.error(`Error fetching ${section.name}:`, (error as Error).message)
    return `
---

## ${section.type.charAt(0).toUpperCase() + section.type.slice(1)}: ${section.name}

**Source:** ${url}

*Error: Could not fetch documentation*
`
  }
}

async function main() {
  console.log('Starting Baileys documentation scrape...\n')

  const markdown: string[] = [
    `# Baileys API Documentation Reference

> Auto-generated documentation for WhatsApp Nucleo development
> Source: https://baileys.wiki
> Generated: ${new Date().toISOString()}

This document contains the relevant Baileys API documentation for developing the WhatsApp Nucleo messaging features.

## Table of Contents

### Interfaces
${SECTIONS.filter(s => s.type === 'interface').map(s => `- [${s.name}](#interface-${s.name.toLowerCase()})`).join('\n')}

### Type Aliases
${SECTIONS.filter(s => s.type === 'type').map(s => `- [${s.name}](#type-${s.name.toLowerCase()})`).join('\n')}

### Functions
${SECTIONS.filter(s => s.type === 'function').map(s => `- [${s.name}](#function-${s.name.toLowerCase()})`).join('\n')}

### Variables
${SECTIONS.filter(s => s.type === 'variable').map(s => `- [${s.name}](#variable-${s.name.toLowerCase()})`).join('\n')}

### Enumerations
${SECTIONS.filter(s => s.type === 'enum').map(s => `- [${s.name}](#enum-${s.name.toLowerCase()})`).join('\n')}

`
  ]

  // Group sections by type
  const groupedSections = {
    interface: SECTIONS.filter(s => s.type === 'interface'),
    type: SECTIONS.filter(s => s.type === 'type'),
    function: SECTIONS.filter(s => s.type === 'function'),
    variable: SECTIONS.filter(s => s.type === 'variable'),
    enum: SECTIONS.filter(s => s.type === 'enum'),
  }

  // Process each group
  for (const [type, sections] of Object.entries(groupedSections)) {
    markdown.push(`\n# ${type.charAt(0).toUpperCase() + type.slice(1)}s\n`)

    for (const section of sections) {
      const content = await scrapeSection(section)
      markdown.push(content)

      // Small delay to be nice to the server
      await new Promise(r => setTimeout(r, 300))
    }
  }

  // Write to file
  const outputPath = './docs/baileys-api-reference.md'
  fs.writeFileSync(outputPath, markdown.join('\n'))

  console.log(`\nDocumentation saved to ${outputPath}`)
  console.log(`Total sections: ${SECTIONS.length}`)
}

main().catch(console.error)