From 45c0e98106e2473606fbdc6c3d7644c5578f37ea Mon Sep 17 00:00:00 2001
From: Liviu-Mihail Concioiu <liviu.concioiu@gmail.com>
Date: Thu, 19 Sep 2024 17:57:37 +0200
Subject: [PATCH] Adds detection for Oh Dear and improves detection for
 vuhuvBot and Bravebot (#7837)

* Change category for Bravebot
* Improves detection for vuhuvBot
* Adds detection for Oh Dear
ref #7836
---
 Tests/fixtures/bots.yml | 41 +++++++++++++++++++++++++++++++++++++----
 regexes/bots.yml        | 18 +++++++++++++-----
 2 files changed, 50 insertions(+), 9 deletions(-)

diff --git a/Tests/fixtures/bots.yml b/Tests/fixtures/bots.yml
index f4cbb482e4..5200b9e578 100644
--- a/Tests/fixtures/bots.yml
+++ b/Tests/fixtures/bots.yml
@@ -5812,9 +5812,15 @@
 -
   user_agent: Mozilla/5.0 (compatible; vuhuvBot/1.0; +http://vuhuv.com/bot.html)
   bot:
-    name: Vuhuv Bot
-    category: Crawler
-    url: http://vuhuv.com/bot.html
+    name: vuhuvBot
+    category: Search bot
+    url: https://vuhuv.com/bot.html
+-
+  user_agent: Mozilla/5.0 (compatible; vuhuvRBT/2.0; +https://vuhuv.com/rbt.html)
+  bot:
+    name: vuhuvBot
+    category: Search bot
+    url: https://vuhuv.com/bot.html
 -
   user_agent: Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.1; Trident/6.0) SiteCheck-sitecrawl by Siteimprove.com
   bot:
@@ -8238,7 +8244,7 @@
   user_agent: Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Bravebot/1.0; +https://search.brave.com/help/brave-search-crawler) Chrome/W.X.Y.Z Safari/537.36
   bot:
     name: Bravebot
-    category: Crawler
+    category: Search bot
     url: https://search.brave.com/help/brave-search-crawler
     producer:
       name: Brave Software, Inc.
@@ -8267,3 +8273,30 @@
     producer:
       name: Valve Corporation
       url: https://www.valvesoftware.com/
+-
+  user_agent: Oh Dear Sitemap check https://ohdear.app
+  bot:
+    name: Oh Dear
+    category: Site Monitor
+    url: https://ohdear.app/docs/faq/what-is-the-oh-dear-crawler-doing-in-my-logs
+    producer:
+      name: Immutable, SNC
+      url: https://ohdear.app/
+-
+  user_agent: Mozilla/5.0 (compatible; OhDear/1.1; +https://ohdear.app/checker; brokenLinks)
+  bot:
+    name: Oh Dear
+    category: Site Monitor
+    url: https://ohdear.app/docs/faq/what-is-the-oh-dear-crawler-doing-in-my-logs
+    producer:
+      name: Immutable, SNC
+      url: https://ohdear.app/
+-
+  user_agent: Mozilla/5.0 (compatible; OhDear/1.1; +https://ohdear.app/checker; mixedContent)
+  bot:
+    name: Oh Dear
+    category: Site Monitor
+    url: https://ohdear.app/docs/faq/what-is-the-oh-dear-crawler-doing-in-my-logs
+    producer:
+      name: Immutable, SNC
+      url: https://ohdear.app/
diff --git a/regexes/bots.yml b/regexes/bots.yml
index 757f7ad777..e0a1efd41f 100644
--- a/regexes/bots.yml
+++ b/regexes/bots.yml
@@ -876,10 +876,10 @@
     name: 'HubSpot Inc.'
     url: 'https://www.hubspot.com'
 
-- regex: 'vuhuvBot'
-  name: 'Vuhuv Bot'
-  category: 'Crawler'
-  url: 'http://vuhuv.com/bot.html'
+- regex: 'vuhuv(?:Bot|RBT)'
+  name: 'vuhuvBot'
+  category: 'Search bot'
+  url: 'https://vuhuv.com/bot.html'
 
 - regex: 'HTTPMon/[\d.]+'
   name: 'HTTPMon'
@@ -4787,7 +4787,7 @@
 
 - regex: 'Bravebot'
   name: 'Bravebot'
-  category: 'Crawler'
+  category: 'Search bot'
   url: 'https://search.brave.com/help/brave-search-crawler'
   producer:
     name: 'Brave Software, Inc.'
@@ -4806,6 +4806,14 @@
     name: 'Valve Corporation'
     url: 'https://www.valvesoftware.com/'
 
+- regex: 'ohdear\.app'
+  name: 'Oh Dear'
+  category: 'Site Monitor'
+  url: 'https://ohdear.app/docs/faq/what-is-the-oh-dear-crawler-doing-in-my-logs'
+  producer:
+    name: 'Immutable, SNC'
+    url: 'https://ohdear.app/'
+
 # Generic bots
 - regex: 'nuhk|grub-client|Download Demon|SearchExpress|Microsoft URL Control|borg|altavista|dataminr\.com|teoma|oegp|http%20client|htdig|mogimogi|larbin|scrubby|searchsight|semanticdiscovery|snappy|vortex(?!(?: Build|Plus| CM62| HD65))|zeal(?!ot)|dataparksearch|findlinks|BrowserMob|URL2PNG|ZooShot|GomezA|Google SketchUp|Read%20Later|7Siters|centuryb\.o\.t9|InterNaetBoten|EasyBib AutoCite|Bidtellect|tomnomnom/meg|cortex|Re-re Studio|adreview|AHC/|NameOfAgent|Request-Promise|ALittle Client|Hello,? world|wp_is_mobile|0xAbyssalDoesntExist|Anarchy99|^revolt|nvd0rz|xfa1|Hakai|gbrmss|fuck-your-hp|IDBTE4M CODE87|Antoine|Insomania|Hells-Net|b3astmode|Linux Gnu \(cow\)|Test Certificate Info|iplabel|Magellan|TheSafex?Internetx?Search|Searcherx?web|kirkland-signature|LinkChain|survey-security-dot-txt|infrawatch|Time/|r00ts3c-owned-you|nvdorz|Root Slut|NiggaBalls|BotPoke|GlobalWebSearch|xx032_bo9vs83_2a|sslshed|geckotrail|Wordup|^xenu|^(?:chrome|firefox|Abcd|Dark|KvshClient|Node.js|Report Runner|url|Zeus|ZmEu)$'
   name: 'Generic Bot'