feat: add RecipeRadar crawler

ua-parser · Sep 17, 2024 · cee7823 · cee7823
1 parent 1458dfa
commit cee7823
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 3 deletions.
diff --git a/regexes.yaml b/regexes.yaml
@@ -179,6 +179,10 @@ user_agent_parsers:
   - regex: '\[FB.{0,300};'
     family_replacement: 'Facebook'
 
+  # RecipeRadar crawler
+  - regex: '(RecipeRadar)/(\d+)\.(\d+)\.(\d+)'
+  - regex: '(RecipeRadar)/(\d+)\.(\d+)'
+
   # Bots General matcher 'name/0.0'
   - regex: '^.{0,200}?(?:\/[A-Za-z0-9\.]{0,50}|) {0,2}([A-Za-z0-9 \-_\!\[\]:]{0,50}(?:[Aa]rchiver|[Ii]ndexer|[Ss]craper|[Bb]ot|[Ss]pider|[Cc]rawl[a-z]{0,50}))[/ ](\d+)(?:\.(\d+)(?:\.(\d+)|)|)'
   # Bots containing bot(but not CUBOT)
@@ -5905,7 +5909,7 @@ device_parsers:
   ##########
   # Spiders (this is a hack...)
   ##########
-  - regex: '^.{0,100}(bot|BUbiNG|zao|borg|DBot|oegp|silk|Xenu|zeal|^NING|CCBot|crawl|htdig|lycos|slurp|teoma|voila|yahoo|Sogou|CiBra|Nutch|^Java/|^JNLP/|Daumoa|Daum|Genieo|ichiro|larbin|pompos|Scrapy|snappy|speedy|spider|msnbot|msrbot|vortex|^vortex|crawler|favicon|indexer|Riddler|scooter|scraper|scrubby|WhatWeb|WinHTTP|bingbot|BingPreview|openbot|gigabot|furlbot|polybot|seekbot|^voyager|archiver|Icarus6j|mogimogi|Netvibes|blitzbot|altavista|charlotte|findlinks|Retreiver|TLSProber|WordPress|SeznamBot|ProoXiBot|wsr\-agent|Squrl Java|EtaoSpider|PaperLiBot|SputnikBot|A6\-Indexer|netresearch|searchsight|baiduspider|YisouSpider|ICC\-Crawler|http%20client|Python-urllib|dataparksearch|converacrawler|Screaming Frog|AppEngine-Google|YahooCacheSystem|fast\-webcrawler|Sogou Pic Spider|semanticdiscovery|Innovazion Crawler|facebookexternalhit|Google.{0,200}/\+/web/snippet|Google-HTTP-Java-Client|BlogBridge|IlTrovatore-Setaccio|InternetArchive|GomezAgent|WebThumbnail|heritrix|NewsGator|PagePeeker|Reaper|ZooShot|holmes|NL-Crawler|Pingdom|StatusCake|WhatsApp|masscan|Google Web Preview|Qwantify|Yeti|OgScrper)'
+  - regex: '^.{0,100}(bot|BUbiNG|zao|borg|DBot|oegp|silk|Xenu|zeal|^NING|CCBot|crawl|htdig|lycos|slurp|teoma|voila|yahoo|Sogou|CiBra|Nutch|^Java/|^JNLP/|Daumoa|Daum|Genieo|ichiro|larbin|pompos|Scrapy|snappy|speedy|spider|msnbot|msrbot|vortex|^vortex|crawler|favicon|indexer|Riddler|scooter|scraper|scrubby|WhatWeb|WinHTTP|bingbot|BingPreview|openbot|gigabot|furlbot|polybot|seekbot|^voyager|archiver|Icarus6j|mogimogi|Netvibes|blitzbot|altavista|charlotte|findlinks|Retreiver|TLSProber|WordPress|SeznamBot|ProoXiBot|wsr\-agent|Squrl Java|EtaoSpider|PaperLiBot|SputnikBot|A6\-Indexer|netresearch|searchsight|baiduspider|YisouSpider|ICC\-Crawler|http%20client|Python-urllib|dataparksearch|converacrawler|Screaming Frog|AppEngine-Google|YahooCacheSystem|fast\-webcrawler|Sogou Pic Spider|semanticdiscovery|Innovazion Crawler|facebookexternalhit|Google.{0,200}/\+/web/snippet|Google-HTTP-Java-Client|BlogBridge|IlTrovatore-Setaccio|InternetArchive|GomezAgent|WebThumbnail|heritrix|NewsGator|PagePeeker|Reaper|ZooShot|holmes|NL-Crawler|Pingdom|StatusCake|WhatsApp|masscan|Google Web Preview|Qwantify|Yeti|OgScrper|RecipeRadar)'
     regex_flag: 'i'
     device_replacement: 'Spider'
     brand_replacement: 'Spider'

diff --git a/tests/test_device.yaml b/tests/test_device.yaml
@@ -80619,5 +80619,7 @@ test_cases:
     brand: 'Motorola'
     model: 'motorola moto g play (2021)'
 
-
-
+  - user_agent_string: 'Mozilla/5.0 (compatible; Linux x86_64; python-requests/2.32.3; RecipeRadar/0.1; +https://www.reciperadar.com)'
+    family: 'Spider'
+    brand: 'Spider'
+    model: 'Desktop'
diff --git a/tests/test_ua.yaml b/tests/test_ua.yaml
@@ -8889,3 +8889,9 @@ test_cases:
     minor: '3'
     patch: '0'
     patch_minor: '0'
+
+  - user_agent_string: 'Mozilla/5.0 (compatible; Linux x86_64; python-requests/2.32.3; RecipeRadar/0.1; +https://www.reciperadar.com)'
+    family: 'RecipeRadar'
+    major: '0'
+    minor: '1'
+    patch: