# ============================================
# ROBOTS.TXT - InDreams Phuket Real Estate
# SPRINT 4.2: SEO 2025 Optimized Configuration
# Last Updated: December 2025
# ============================================

# Default rules for all bots
User-agent: *

# Allow main content
Allow: /
Allow: /property/
Allow: /project/
Allow: /location/
Allow: /articleabout/
Allow: /newsabout/
Allow: /serviceabout/

# Allow only first 4 pages of pagination
Allow: /*?page=0
Allow: /*?page=1
Allow: /*?page=2
Allow: /*?page=3
Allow: /*?page=4

# ============================================
# SMART FILTER STRATEGY (Allow filters, use canonical)
# ============================================

# Allow all search/filter pages - SEO is built on them!
# Canonical tags handle deduplication, not robots.txt
Allow: /s45/search/
Allow: /search/

# Block only truly useless parameters
Disallow: /*?*sort=*
Disallow: /*?*view=*
Disallow: /*?*ajax*
Disallow: /*?*destination=*
Disallow: /*?*_=*

# Block deep pagination (page 10+)
Disallow: /*?page=1?*
Disallow: /*?page=2?*
Disallow: /*?page=3?*
Disallow: /*?page=4?*
Disallow: /*?page=5?*

# Block UTM and tracking parameters
Disallow: /*?*utm_*
Disallow: /*?*gclid*
Disallow: /*?*fbclid*
Disallow: /*?*yclid*

# Block admin and system paths
Disallow: /admin/
Disallow: /user/
Disallow: /batch/
Disallow: /cron.php
Disallow: /install.php
Disallow: /update.php
Disallow: /xmlrpc.php
Disallow: /CHANGELOG.txt
Disallow: /INSTALL.txt
Disallow: /LICENSE.txt
Disallow: /MAINTAINERS.txt
Disallow: /UPGRADE.txt
Disallow: /README.txt
Disallow: /INSTALL.mysql.txt
Disallow: /INSTALL.pgsql.txt
Disallow: /INSTALL.sqlite.txt
Disallow: /authorize.php
Disallow: /sites/default/files/private/

# Block module/theme paths
Disallow: /includes/
Disallow: /misc/
Disallow: /modules/
Disallow: /profiles/
Disallow: /scripts/
Disallow: /themes/
Disallow: /sites/all/modules/
Disallow: /sites/all/themes/

# (Removed duplicate - already blocked above)

# Block PDF generation URLs
Disallow: /PhuketPdf/

# Block internal API endpoints
Disallow: /s45PhuketQuery
Disallow: /s45PhuketFormAPI
Disallow: /s45PhuketFormAPI2
Disallow: /s45phuket-exec
Disallow: /s45phuket-perenos
Disallow: /s45phuket-perebor

# Block old/legacy paths
Disallow: /node/
Disallow: /taxonomy/

# ============================================
# Sitemaps - All Language Versions
# ============================================

# English (main)
Sitemap: https://indreamsphuket.com/sitemap.xml
Sitemap: https://indreamsphuket.com/sitemap_en.xml

# Russian
Sitemap: https://indreamsphuket.ru/sitemap.xml
Sitemap: https://indreamsphuket.ru/sitemap_ru.xml

# Thai
Sitemap: https://th.indreamsphuket.com/sitemap.xml
Sitemap: https://th.indreamsphuket.com/sitemap_th.xml

# Chinese
Sitemap: https://ch.indreamsphuket.com/sitemap.xml
Sitemap: https://ch.indreamsphuket.com/sitemap_zh.xml

# ============================================
# Bot-specific rules
# ============================================

# Google specific
User-agent: Googlebot
Allow: /
Crawl-delay: 1

# Google Images
User-agent: Googlebot-Image
Allow: /sites/default/files/
Allow: /sites/all/modules/__s45/s45_phuket/res/

# Yandex specific
User-agent: Yandex
Allow: /
Crawl-delay: 2
Host: https://indreamsphuket.ru

# Bing
User-agent: Bingbot
Allow: /
Crawl-delay: 1

# Block bad bots
User-agent: AhrefsBot
Disallow: /

User-agent: SemrushBot
Disallow: /

User-agent: MJ12bot
Disallow: /

User-agent: DotBot
Disallow: /

User-agent: BLEXBot
Disallow: /

# ============================================
# Host directive (for Yandex)
# ============================================
Host: https://indreamsphuket.com