honeymoose
/
OpenSearch
mirror of https://github.com/honeymoose/OpenSearch.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

update ingest-user-agent regexes.yml (#25608) 

This new regexes are from:

3153c2f2ae/regexes.yaml
This commit is contained in:
Tal Levy 2017-07-10 08:43:11 -07:00 committed by GitHub
parent 7836bbf4d4
commit 8cf0528001
1 changed files with 201 additions and 55 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

256
plugins/ingest-user-agent/src/main/resources/regexes.yml
View File

@ -25,6 +25,13 @@ user_agent_parsers:
# Pingdom # Pingdom
- regex: '(Pingdom.com_bot_version_)(\d+)\.(\d+)' - regex: '(Pingdom.com_bot_version_)(\d+)\.(\d+)'
family_replacement: 'PingdomBot' family_replacement: 'PingdomBot'
# 'Mozilla/5.0 (Unknown; Linux x86_64) AppleWebKit/534.34 (KHTML, like Gecko) PingdomTMS/0.8.5 Safari/534.34'
- regex: '(PingdomTMS)/(\d+)\.(\d+)\.(\d+)'
family_replacement: 'PingdomBot'
#StatusCake
- regex: '(\(StatusCake\))'
family_replacement: 'StatusCakeBot'
# Facebook # Facebook
- regex: '(facebookexternalhit)/(\d+)\.(\d+)' - regex: '(facebookexternalhit)/(\d+)\.(\d+)'
@ -34,6 +41,10 @@ user_agent_parsers:
- regex: 'Google.*/\+/web/snippet' - regex: 'Google.*/\+/web/snippet'
family_replacement: 'GooglePlusBot' family_replacement: 'GooglePlusBot'
# Gmail
- regex: 'via ggpht.com GoogleImageProxy'
family_replacement: 'GmailImageProxy'
# Twitter # Twitter
- regex: '(Twitterbot)/(\d+)\.(\d+)' - regex: '(Twitterbot)/(\d+)\.(\d+)'
family_replacement: 'TwitterBot' family_replacement: 'TwitterBot'
@ -41,17 +52,17 @@ user_agent_parsers:
# Bots Pattern '/name-0.0' # Bots Pattern '/name-0.0'
- regex: '/((?:Ant-)?Nutch|[A-z]+[Bb]ot|[A-z]+[Ss]pider|Axtaris|fetchurl|Isara|ShopSalad|Tailsweep)[ \-](\d+)(?:\.(\d+)(?:\.(\d+))?)?' - regex: '/((?:Ant-)?Nutch|[A-z]+[Bb]ot|[A-z]+[Ss]pider|Axtaris|fetchurl|Isara|ShopSalad|Tailsweep)[ \-](\d+)(?:\.(\d+)(?:\.(\d+))?)?'
# Bots Pattern 'name/0.0' # Bots Pattern 'name/0.0'
- regex: '(008|Altresium|Argus|BaiduMobaider|BoardReader|DNSGroup|DataparkSearch|EDI|Goodzer|Grub|INGRID|Infohelfer|LinkedInBot|LOOQ|Nutch|PathDefender|Peew|PostPost|Steeler|Twitterbot|VSE|WebCrunch|WebZIP|Y!J-BR[A-Z]|YahooSeeker|envolk|sproose|wminer)/(\d+)(?:\.(\d+)(?:\.(\d+))?)?' - regex: '\b(008|Altresium|Argus|BaiduMobaider|BoardReader|DNSGroup|DataparkSearch|EDI|Goodzer|Grub|INGRID|Infohelfer|LinkedInBot|LOOQ|Nutch|PathDefender|Peew|PostPost|Steeler|Twitterbot|VSE|WebCrunch|WebZIP|Y!J-BR[A-Z]|YahooSeeker|envolk|sproose|wminer)/(\d+)(?:\.(\d+)(?:\.(\d+))?)?'
# MSIECrawler # MSIECrawler
- regex: '(MSIE) (\d+)\.(\d+)([a-z]\d?)?;.* MSIECrawler' - regex: '(MSIE) (\d+)\.(\d+)([a-z]\d?)?;.* MSIECrawler'
family_replacement: 'MSIECrawler' family_replacement: 'MSIECrawler'
# Downloader ... # Downloader ...
- regex: '(Google-HTTP-Java-Client|Apache-HttpClient|http%20client|Python-urllib|HttpMonitor|TLSProber|WinHTTP|JNLP)(?:[ /](\d+)(?:\.(\d+)(?:\.(\d+))?)?)?' - regex: '(Google-HTTP-Java-Client|Apache-HttpClient|http%20client|Python-urllib|HttpMonitor|TLSProber|WinHTTP|JNLP|okhttp)(?:[ /](\d+)(?:\.(\d+)(?:\.(\d+))?)?)?'
# Bots # Bots
- regex: '(1470\.net crawler|50\.nu|8bo Crawler Bot|Aboundex|Accoona-[A-z]+-Agent|AdsBot-Google(?:-[a-z]+)?|altavista|AppEngine-Google|archive.*?\.org_bot|archiver|Ask Jeeves|[Bb]ai[Dd]u[Ss]pider(?:-[A-Za-z]+)*|bingbot|BingPreview|blitzbot|BlogBridge|BoardReader(?: [A-Za-z]+)*|boitho.com-dc|BotSeer|\b\w*favicon\w*\b|\bYeti(?:-[a-z]+)?|Catchpoint bot|[Cc]harlotte|Checklinks|clumboot|Comodo HTTP\(S\) Crawler|Comodo-Webinspector-Crawler|ConveraCrawler|CRAWL-E|CrawlConvera|Daumoa(?:-feedfetcher)?|Feed Seeker Bot|findlinks|Flamingo_SearchEngine|FollowSite Bot|furlbot|Genieo|gigabot|GomezAgent|gonzo1|(?:[a-zA-Z]+-)?Googlebot(?:-[a-zA-Z]+)?|Google SketchUp|grub-client|gsa-crawler|heritrix|HiddenMarket|holmes|HooWWWer|htdig|ia_archiver|ICC-Crawler|Icarus6j|ichiro(?:/mobile)?|IconSurf|IlTrovatore(?:-Setaccio)?|InfuzApp|Innovazion Crawler|InternetArchive|IP2[a-z]+Bot|jbot\b|KaloogaBot|Kraken|Kurzor|larbin|LEIA|LesnikBot|Linguee Bot|LinkAider|LinkedInBot|Lite Bot|Llaut|lycos|Mail\.RU_Bot|masidani_bot|Mediapartners-Google|Microsoft .*? Bot|mogimogi|mozDex|MJ12bot|msnbot(?:-media *)?|msrbot|netresearch|Netvibes|NewsGator[^/]*|^NING|Nutch[^/]*|Nymesis|ObjectsSearch|Orbiter|OOZBOT|PagePeeker|PagesInventory|PaxleFramework|Peeplo Screenshot Bot|PlantyNet_WebRobot|Pompos|Read%20Later|Reaper|RedCarpet|Retreiver|Riddler|Rival IQ|scooter|Scrapy|Scrubby|searchsight|seekbot|semanticdiscovery|Simpy|SimplePie|SEOstats|SimpleRSS|SiteCon|Slurp|snappy|Speedy Spider|Squrl Java|TheUsefulbot|ThumbShotsBot|Thumbshots\.ru|TwitterBot|URL2PNG|Vagabondo|VoilaBot|^vortex|Votay bot|^voyager|WASALive.Bot|Web-sniffer|WebThumb|WeSEE:[A-z]+|WhatWeb|WIRE|WordPress|Wotbox|www\.almaden\.ibm\.com|Xenu(?:.s)? Link Sleuth|Xerka [A-z]+Bot|yacy(?:bot)?|Yahoo[a-z]*Seeker|Yahoo! Slurp|Yandex\w+|YodaoBot(?:-[A-z]+)?|YottaaMonitor|Yowedo|^Zao|^Zao-Crawler|ZeBot_www\.ze\.bz|ZooShot|ZyBorg)(?:[ /]v?(\d+)(?:\.(\d+)(?:\.(\d+))?)?)?' - regex: '(1470\.net crawler|50\.nu|8bo Crawler Bot|Aboundex|Accoona-[A-z]+-Agent|AdsBot-Google(?:-[a-z]+)?|altavista|AppEngine-Google|archive.*?\.org_bot|archiver|Ask Jeeves|[Bb]ai[Dd]u[Ss]pider(?:-[A-Za-z]+)*|bingbot|BingPreview|blitzbot|BlogBridge|Bloglovin|BoardReader(?: [A-Za-z]+)*|boitho.com-dc|BotSeer|\b\w*favicon\w*\b|\bYeti(?:-[a-z]+)?|Catchpoint(?: bot)?|[Cc]harlotte|Checklinks|clumboot|Comodo HTTP\(S\) Crawler|Comodo-Webinspector-Crawler|ConveraCrawler|CRAWL-E|CrawlConvera|Daumoa(?:-feedfetcher)?|Feed Seeker Bot|Feedbin|findlinks|Flamingo_SearchEngine|FollowSite Bot|furlbot|Genieo|gigabot|GomezAgent|gonzo1|(?:[a-zA-Z]+-)?Googlebot(?:-[a-zA-Z]+)?|Google SketchUp|grub-client|gsa-crawler|heritrix|HiddenMarket|holmes|HooWWWer|htdig|ia_archiver|ICC-Crawler|Icarus6j|ichiro(?:/mobile)?|IconSurf|IlTrovatore(?:-Setaccio)?|InfuzApp|Innovazion Crawler|InternetArchive|IP2[a-z]+Bot|jbot\b|KaloogaBot|Kraken|Kurzor|larbin|LEIA|LesnikBot|Linguee Bot|LinkAider|LinkedInBot|Lite Bot|Llaut|lycos|Mail\.RU_Bot|masscan|masidani_bot|Mediapartners-Google|Microsoft .*? Bot|mogimogi|mozDex|MJ12bot|msnbot(?:-media *)?|msrbot|Mtps Feed Aggregation System|netresearch|Netvibes|NewsGator[^/]*|^NING|Nutch[^/]*|Nymesis|ObjectsSearch|Orbiter|OOZBOT|PagePeeker|PagesInventory|PaxleFramework|Peeplo Screenshot Bot|PlantyNet_WebRobot|Pompos|Qwantify|Read%20Later|Reaper|RedCarpet|Retreiver|Riddler|Rival IQ|scooter|Scrapy|Scrubby|searchsight|seekbot|semanticdiscovery|Simpy|SimplePie|SEOstats|SimpleRSS|SiteCon|Slackbot-LinkExpanding|Slack-ImgProxy|Slurp|snappy|Speedy Spider|Squrl Java|Stringer|TheUsefulbot|ThumbShotsBot|Thumbshots\.ru|Tiny Tiny RSS|TwitterBot|WhatsApp|URL2PNG|Vagabondo|VoilaBot|^vortex|Votay bot|^voyager|WASALive.Bot|Web-sniffer|WebThumb|WeSEE:[A-z]+|WhatWeb|WIRE|WordPress|Wotbox|www\.almaden\.ibm\.com|Xenu(?:.s)? Link Sleuth|Xerka [A-z]+Bot|yacy(?:bot)?|Yahoo[a-z]*Seeker|Yahoo! Slurp|Yandex\w+|YodaoBot(?:-[A-z]+)?|YottaaMonitor|Yowedo|^Zao|^Zao-Crawler|ZeBot_www\.ze\.bz|ZooShot|ZyBorg)(?:[ /]v?(\d+)(?:\.(\d+)(?:\.(\d+))?)?)?'
# Bots General matcher 'name/0.0' # Bots General matcher 'name/0.0'
- regex: '(?:\/[A-Za-z0-9\.]+)? *([A-Za-z0-9 \-_\!\[\]:]*(?:[Aa]rchiver|[Ii]ndexer|[Ss]craper|[Bb]ot|[Ss]pider|[Cc]rawl[a-z]*))/(\d+)(?:\.(\d+)(?:\.(\d+))?)?' - regex: '(?:\/[A-Za-z0-9\.]+)? *([A-Za-z0-9 \-_\!\[\]:]*(?:[Aa]rchiver|[Ii]ndexer|[Ss]craper|[Bb]ot|[Ss]pider|[Cc]rawl[a-z]*))/(\d+)(?:\.(\d+)(?:\.(\d+))?)?'
@ -76,9 +87,11 @@ user_agent_parsers:
- regex: '\[(Pinterest)/[^\]]+\]' - regex: '\[(Pinterest)/[^\]]+\]'
- regex: '(Pinterest)(?: for Android(?: Tablet)?)?/(\d+)(?:\.(\d+)(?:\.(\d)+)?)?' - regex: '(Pinterest)(?: for Android(?: Tablet)?)?/(\d+)(?:\.(\d+)(?:\.(\d)+)?)?'
# Pale Moon
- regex: '(PaleMoon)/(\d+)\.(\d+)\.?(\d+)?'
family_replacement: 'Pale Moon'
# Firefox # Firefox
- regex: '(Pale[Mm]oon)/(\d+)\.(\d+)\.?(\d+)?'
family_replacement: 'Pale Moon (Firefox Variant)'
- regex: '(Fennec)/(\d+)\.(\d+)\.?([ab]?\d+[a-z]*)' - regex: '(Fennec)/(\d+)\.(\d+)\.?([ab]?\d+[a-z]*)'
family_replacement: 'Firefox Mobile' family_replacement: 'Firefox Mobile'
- regex: '(Fennec)/(\d+)\.(\d+)(pre)' - regex: '(Fennec)/(\d+)\.(\d+)(pre)'
@ -124,6 +137,11 @@ user_agent_parsers:
- regex: '(MyIBrow)/(\d+)\.(\d+)' - regex: '(MyIBrow)/(\d+)\.(\d+)'
family_replacement: 'My Internet Browser' family_replacement: 'My Internet Browser'
# UC Browser
# we need check it before opera. In other case case UC Browser detected look like Opera Mini
- regex: '(UC? ?Browser|UCWEB|U3)[ /]?(\d+)\.(\d+)\.(\d+)'
family_replacement: 'UC Browser'
# Opera will stop at 9.80 and hide the real version in the Version string. # Opera will stop at 9.80 and hide the real version in the Version string.
# see: http://dev.opera.com/articles/view/opera-ua-string-changes/ # see: http://dev.opera.com/articles/view/opera-ua-string-changes/
- regex: '(Opera Tablet).*Version/(\d+)\.(\d+)(?:\.(\d+))?' - regex: '(Opera Tablet).*Version/(\d+)\.(\d+)(?:\.(\d+))?'
@ -154,6 +172,10 @@ user_agent_parsers:
- regex: '(OPiOS)/(\d+).(\d+).(\d+)' - regex: '(OPiOS)/(\d+).(\d+).(\d+)'
family_replacement: 'Opera Mini' family_replacement: 'Opera Mini'
# Opera Neon
- regex: 'Chrome/.+( MMS)/(\d+).(\d+).(\d+)'
family_replacement: 'Opera Neon'
# Palm WebOS looks a lot like Safari. # Palm WebOS looks a lot like Safari.
- regex: '(hpw|web)OS/(\d+)\.(\d+)(?:\.(\d+))?' - regex: '(hpw|web)OS/(\d+)\.(\d+)(?:\.(\d+))?'
family_replacement: 'webOS Browser' family_replacement: 'webOS Browser'
@ -168,7 +190,7 @@ user_agent_parsers:
# Lightning (for Thunderbird) # Lightning (for Thunderbird)
# http://www.mozilla.org/projects/calendar/lightning/ # http://www.mozilla.org/projects/calendar/lightning/
- regex: '(Lightning)/(\d+)\.(\d+)\.?((?:[ab]?\d+[a-z]*)|(?:\d*))' - regex: 'Gecko/\d+ (Lightning)/(\d+)\.(\d+)\.?((?:[ab]?\d+[a-z]*)|(?:\d*))'
# Swiftfox # Swiftfox
- regex: '(Firefox)/(\d+)\.(\d+)\.(\d+(?:pre)?) \(Swiftfox\)' - regex: '(Firefox)/(\d+)\.(\d+)\.(\d+(?:pre)?) \(Swiftfox\)'
@ -198,8 +220,6 @@ user_agent_parsers:
- regex: '(Symphony) (\d+).(\d+)' - regex: '(Symphony) (\d+).(\d+)'
- regex: '(Minimo)'
- regex: 'PLAYSTATION 3.+WebKit' - regex: 'PLAYSTATION 3.+WebKit'
family_replacement: 'NetFront NX' family_replacement: 'NetFront NX'
- regex: 'PLAYSTATION 3' - regex: 'PLAYSTATION 3'
@ -218,7 +238,6 @@ user_agent_parsers:
- regex: '(Silk)/(\d+)\.(\d+)(?:\.([0-9\-]+))?' - regex: '(Silk)/(\d+)\.(\d+)(?:\.([0-9\-]+))?'
family_replacement: 'Amazon Silk' family_replacement: 'Amazon Silk'
# @ref: http://www.puffinbrowser.com # @ref: http://www.puffinbrowser.com
- regex: '(Puffin)/(\d+)\.(\d+)(?:\.(\d+))?' - regex: '(Puffin)/(\d+)\.(\d+)(?:\.(\d+))?'
@ -230,25 +249,43 @@ user_agent_parsers:
- regex: '(SamsungBrowser)/(\d+)\.(\d+)' - regex: '(SamsungBrowser)/(\d+)\.(\d+)'
family_replacement: 'Samsung Internet' family_replacement: 'Samsung Internet'
# Seznam.cz browser (based on WebKit)
- regex: '(SznProhlizec)/(\d+)\.(\d+)(?:\.(\d+))?'
family_replacement: 'Seznam.cz'
# Coc Coc browser, based on Chrome (used in Vietnam)
- regex: '(coc_coc_browser)/(\d+)\.(\d+)(?:\.(\d+))?'
family_replacement: 'Coc Coc'
# Baidu Browsers (desktop spoofs chrome & IE, explorer is mobile)
- regex: '(baidubrowser)[/\s](\d+)(?:\.(\d+)(?:\.(\d+))?)?'
family_replacement: 'Baidu Browser'
- regex: '(FlyFlow)/(\d+)\.(\d+)'
family_replacement: 'Baidu Explorer'
# MxBrowser is Maxthon. Must go before Mobile Chrome for Android
- regex: '(MxBrowser)/(\d+)\.(\d+)(?:\.(\d+))?'
family_replacement: 'Maxthon'
# Crosswalk must go before Mobile Chrome for Android
- regex: '(Crosswalk)/(\d+)\.(\d+)\.(\d+)\.(\d+)'
# Chrome Mobile # Chrome Mobile
- regex: '; wv\).+(Chrome)/(\d+)\.(\d+)\.(\d+)\.(\d+)'
family_replacement: 'Chrome Mobile WebView'
- regex: '(CrMo)/(\d+)\.(\d+)\.(\d+)\.(\d+)' - regex: '(CrMo)/(\d+)\.(\d+)\.(\d+)\.(\d+)'
family_replacement: 'Chrome Mobile' family_replacement: 'Chrome Mobile'
- regex: '(CriOS)/(\d+)\.(\d+)\.(\d+)\.(\d+)' - regex: '(CriOS)/(\d+)\.(\d+)\.(\d+)\.(\d+)'
family_replacement: 'Chrome Mobile iOS' family_replacement: 'Chrome Mobile iOS'
- regex: '(Chrome)/(\d+)\.(\d+)\.(\d+)\.(\d+) Mobile' - regex: '(Chrome)/(\d+)\.(\d+)\.(\d+)\.(\d+) Mobile(?:[ /]|$)'
family_replacement: 'Chrome Mobile'
- regex: ' Mobile .*(Chrome)/(\d+)\.(\d+)\.(\d+)\.(\d+)'
family_replacement: 'Chrome Mobile' family_replacement: 'Chrome Mobile'
# Chrome Frame must come before MSIE. # Chrome Frame must come before MSIE.
- regex: '(chromeframe)/(\d+)\.(\d+)\.(\d+)' - regex: '(chromeframe)/(\d+)\.(\d+)\.(\d+)'
family_replacement: 'Chrome Frame' family_replacement: 'Chrome Frame'
# UC Browser
- regex: '(UCBrowser)[ /](\d+)\.(\d+)\.(\d+)'
family_replacement: 'UC Browser'
- regex: '(UC Browser)[ /](\d+)\.(\d+)\.(\d+)'
- regex: '(UC Browser|UCBrowser|UCWEB)(\d+)\.(\d+)\.(\d+)'
family_replacement: 'UC Browser'
# Tizen Browser (second case included in browser/major.minor regex) # Tizen Browser (second case included in browser/major.minor regex)
- regex: '(SLP Browser)/(\d+)\.(\d+)' - regex: '(SLP Browser)/(\d+)\.(\d+)'
family_replacement: 'Tizen Browser' family_replacement: 'Tizen Browser'
@ -257,12 +294,6 @@ user_agent_parsers:
- regex: '(SE 2\.X) MetaSr (\d+)\.(\d+)' - regex: '(SE 2\.X) MetaSr (\d+)\.(\d+)'
family_replacement: 'Sogou Explorer' family_replacement: 'Sogou Explorer'
# Baidu Browsers (desktop spoofs chrome & IE, explorer is mobile)
- regex: '(baidubrowser)[/\s](\d+)'
family_replacement: 'Baidu Browser'
- regex: '(FlyFlow)/(\d+)\.(\d+)'
family_replacement: 'Baidu Explorer'
# QQ Browsers # QQ Browsers
- regex: '(MQQBrowser/Mini)(?:(\d+)(?:\.(\d+)(?:\.(\d+))?)?)?' - regex: '(MQQBrowser/Mini)(?:(\d+)(?:\.(\d+)(?:\.(\d+))?)?)?'
family_replacement: 'QQ Browser Mini' family_replacement: 'QQ Browser Mini'
@ -293,8 +324,18 @@ user_agent_parsers:
#### MAIN CASES - this catches > 50% of all browsers #### #### MAIN CASES - this catches > 50% of all browsers ####
# Slack desktop client (needs to be before Apple Mail, Electron, and Chrome as it gets wrongly detected on Mac OS otherwise)
- regex: '(Slack_SSB)/(\d+)\.(\d+)\.(\d+)'
family_replacement: 'Slack Desktop Client'
# HipChat provides a version on Mac, but not on Windows.
# Needs to be before Chrome on Windows, and AppleMail on Mac.
- regex: '(HipChat)/?(\d+)?'
family_replacement: 'HipChat Desktop Client'
# Browser/major_version.minor_version.beta_version # Browser/major_version.minor_version.beta_version
- regex: '(AdobeAIR|FireWeb|Jasmine|ANTGalio|Midori|Fresco|Lobo|PaleMoon|Maxthon|Lynx|OmniWeb|Dillo|Camino|Demeter|Fluid|Fennec|Epiphany|Shiira|Sunrise|Spotify|Flock|Netscape|Lunascape|WebPilot|NetFront|Netfront|Konqueror|SeaMonkey|Kazehakase|Vienna|Iceape|Iceweasel|IceWeasel|Iron|K-Meleon|Sleipnir|Galeon|GranParadiso|Opera Mini|iCab|NetNewsWire|ThunderBrowse|Iris|UP\.Browser|Bunjalloo|Google Earth|Raven for Mac|Openwave)/(\d+)\.(\d+)\.(\d+)' - regex: '\b(MobileIron|FireWeb|Jasmine|ANTGalio|Midori|Fresco|Lobo|PaleMoon|Maxthon|Lynx|OmniWeb|Dillo|Camino|Demeter|Fluid|Fennec|Epiphany|Shiira|Sunrise|Spotify|Flock|Netscape|Lunascape|WebPilot|NetFront|Netfront|Konqueror|SeaMonkey|Kazehakase|Vienna|Iceape|Iceweasel|IceWeasel|Iron|K-Meleon|Sleipnir|Galeon|GranParadiso|Opera Mini|iCab|NetNewsWire|ThunderBrowse|Iris|UP\.Browser|Bunjalloo|Google Earth|Raven for Mac|Openwave|MacOutlook|Electron)/(\d+)\.(\d+)\.(\d+)'
# Outlook 2007 # Outlook 2007
- regex: 'Microsoft Office Outlook 12\.\d+\.\d+|MSOffice 12' - regex: 'Microsoft Office Outlook 12\.\d+\.\d+|MSOffice 12'
@ -324,31 +365,50 @@ user_agent_parsers:
- regex: '(Airmail) (\d+)\.(\d+)(?:\.(\d+))?' - regex: '(Airmail) (\d+)\.(\d+)(?:\.(\d+))?'
# Thunderbird # Thunderbird
- regex: '(Thunderbird)/(\d+)\.(\d+)\.(\d+(?:pre)?)' - regex: '(Thunderbird)/(\d+)\.(\d+)(?:\.(\d+(?:pre)?))?'
family_replacement: 'Thunderbird' family_replacement: 'Thunderbird'
# Postbox
- regex: '(Postbox)/(\d+)\.(\d+)\.(\d+)'
family_replacement: 'Postbox'
# Barca
- regex: '(Barca(?:Pro)?)/(\d+)\.(\d+)(?:\.(\d+))?'
family_replacement: 'Barca'
# Lotus Notes
- regex: '(Lotus-Notes)/(\d+)\.(\d+)(?:\.(\d+))?'
family_replacement: 'Lotus Notes'
# Vivaldi uses "Vivaldi" # Vivaldi uses "Vivaldi"
- regex: '(Vivaldi)/(\d+)\.(\d+)\.(\d+)' - regex: '(Vivaldi)/(\d+)\.(\d+)\.(\d+)'
# Edge/major_version.minor_version # Edge/major_version.minor_version
- regex: '(Edge)/(\d+)\.(\d+)' - regex: '(Edge)/(\d+)(?:\.(\d+))?'
# Brave Browser https://brave.com/ # Brave Browser https://brave.com/
- regex: '(brave)/(\d+)\.(\d+)\.(\d+) Chrome' - regex: '(brave)/(\d+)\.(\d+)\.(\d+) Chrome'
family_replacement: 'Brave' family_replacement: 'Brave'
# Chrome/Chromium/major_version.minor_version.beta_version # Iron Browser ~since version 50
- regex: '(Chromium|Chrome)/(\d+)\.(\d+)\.(\d+)' - regex: '(Chrome)/(\d+)\.(\d+)\.(\d+)[\d.]* Iron[^/]'
family_replacement: 'Iron'
# Dolphin Browser # Dolphin Browser
# @ref: http://www.dolphin.com # @ref: http://www.dolphin.com
- regex: '\b(Dolphin)(?: |HDCN/|/INT\-)(\d+)\.(\d+)\.?(\d+)?' - regex: '\b(Dolphin)(?: |HDCN/|/INT\-)(\d+)\.(\d+)\.?(\d+)?'
# Headless Chrome
# https://chromium.googlesource.com/chromium/src/+/lkgr/headless/README.md
# Currently only available on Linux
- regex: 'HeadlessChrome'
family_replacement: 'HeadlessChrome'
# Browser/major_version.minor_version # Browser/major_version.minor_version
- regex: '(bingbot|Bolt|Jasmine|IceCat|Skyfire|Midori|Maxthon|Lynx|Arora|IBrowse|Dillo|Camino|Shiira|Fennec|Phoenix|Chrome|Flock|Netscape|Lunascape|Epiphany|WebPilot|Opera Mini|Opera|NetFront|Netfront|Konqueror|Googlebot|SeaMonkey|Kazehakase|Vienna|Iceape|Iceweasel|IceWeasel|Iron|K-Meleon|Sleipnir|Galeon|GranParadiso|iCab|iTunes|MacAppStore|NetNewsWire|Space Bison|Stainless|Orca|Dolfin|BOLT|Minimo|Tizen Browser|Polaris|Abrowser|Planetweb|ICE Browser|mDolphin|qutebrowser|Otter|QupZilla)/(\d+)\.(\d+)\.?(\d+)?' - regex: '(bingbot|Bolt|AdobeAIR|Jasmine|IceCat|Skyfire|Midori|Maxthon|Lynx|Arora|IBrowse|Dillo|Camino|Shiira|Fennec|Phoenix|Flock|Netscape|Lunascape|Epiphany|WebPilot|Opera Mini|Opera|NetFront|Netfront|Konqueror|Googlebot|SeaMonkey|Kazehakase|Vienna|Iceape|Iceweasel|IceWeasel|Iron|K-Meleon|Sleipnir|Galeon|GranParadiso|iCab|iTunes|MacAppStore|NetNewsWire|Space Bison|Stainless|Orca|Dolfin|BOLT|Minimo|Tizen Browser|Polaris|Abrowser|Planetweb|ICE Browser|mDolphin|qutebrowser|Otter|QupZilla|MailBar|kmail2|YahooMobileMail|ExchangeWebServices|ExchangeServicesClient|Dragon|Outlook-iOS-Android)/(\d+)\.(\d+)(?:\.(\d+))?'
# Chrome/Chromium/major_version.minor_version # Chrome/Chromium/major_version.minor_version
- regex: '(Chromium|Chrome)/(\d+)\.(\d+)' - regex: '(Chromium|Chrome)/(\d+)\.(\d+)(?:\.(\d+))?'
########## ##########
# IE Mobile needs to happen before Android to catch cases such as: # IE Mobile needs to happen before Android to catch cases such as:
@ -362,10 +422,13 @@ user_agent_parsers:
- regex: '(IEMobile)[ /](\d+)\.(\d+)' - regex: '(IEMobile)[ /](\d+)\.(\d+)'
family_replacement: 'IE Mobile' family_replacement: 'IE Mobile'
# Baca Berita App News Reader
- regex: '(BacaBerita App)\/(\d+)\.(\d+)\.(\d+)'
# Browser major_version.minor_version.beta_version (space instead of slash) # Browser major_version.minor_version.beta_version (space instead of slash)
- regex: '(iRider|Crazy Browser|SkipStone|iCab|Lunascape|Sleipnir|Maemo Browser) (\d+)\.(\d+)\.(\d+)' - regex: '(iRider|Crazy Browser|SkipStone|iCab|Lunascape|Sleipnir|Maemo Browser) (\d+)\.(\d+)\.(\d+)'
# Browser major_version.minor_version (space instead of slash) # Browser major_version.minor_version (space instead of slash)
- regex: '(iCab|Lunascape|Opera|Android|Jasmine|Polaris) (\d+)\.(\d+)\.?(\d+)?' - regex: '(iCab|Lunascape|Opera|Android|Jasmine|Polaris|Microsoft SkyDriveSync|The Bat!) (\d+)\.(\d+)\.?(\d+)?'
# Kindle WebKit # Kindle WebKit
- regex: '(Kindle)/(\d+)\.(\d+)' - regex: '(Kindle)/(\d+)\.(\d+)'
@ -395,6 +458,7 @@ user_agent_parsers:
- regex: '(MSIE) (\d+)\.(\d+).*XBLWP7' - regex: '(MSIE) (\d+)\.(\d+).*XBLWP7'
family_replacement: 'IE Large Screen' family_replacement: 'IE Large Screen'
#### END MAIN CASES #### #### END MAIN CASES ####
#### SPECIAL CASES #### #### SPECIAL CASES ####
@ -423,13 +487,13 @@ user_agent_parsers:
family_replacement: 'Bon Echo' family_replacement: 'Bon Echo'
# @note: iOS / OSX Applications # @note: iOS / OSX Applications
- regex: '(iPod|iPhone|iPad).+Version/(\d+)\.(\d+)(?:\.(\d+))?.* Safari' - regex: '(iPod|iPhone|iPad).+Version/(\d+)\.(\d+)(?:\.(\d+))?.*[ +]Safari'
family_replacement: 'Mobile Safari' family_replacement: 'Mobile Safari'
- regex: '(iPod|iPhone|iPad).+Version/(\d+)\.(\d+)(?:\.(\d+))?' - regex: '(iPod|iPhone|iPad).+Version/(\d+)\.(\d+)(?:\.(\d+))?'
family_replacement: 'Mobile Safari UI/WKWebView' family_replacement: 'Mobile Safari UI/WKWebView'
- regex: '(iPod|iPhone|iPad);.*CPU.*OS (\d+)_(\d+)(?:_(\d+))?.*Mobile.* Safari' - regex: '(iPod|iPod touch|iPhone|iPad);.*CPU.*OS[ +](\d+)_(\d+)(?:_(\d+))?.*Mobile.*[ +]Safari'
family_replacement: 'Mobile Safari' family_replacement: 'Mobile Safari'
- regex: '(iPod|iPhone|iPad);.*CPU.*OS (\d+)_(\d+)(?:_(\d+))?.*Mobile' - regex: '(iPod|iPod touch|iPhone|iPad);.*CPU.*OS[ +](\d+)_(\d+)(?:_(\d+))?.*Mobile'
family_replacement: 'Mobile Safari UI/WKWebView' family_replacement: 'Mobile Safari UI/WKWebView'
- regex: '(iPod|iPhone|iPad).* Safari' - regex: '(iPod|iPhone|iPad).* Safari'
family_replacement: 'Mobile Safari' family_replacement: 'Mobile Safari'
@ -550,11 +614,19 @@ user_agent_parsers:
- regex: '(python-requests)/(\d+)\.(\d+)' - regex: '(python-requests)/(\d+)\.(\d+)'
family_replacement: 'Python Requests' family_replacement: 'Python Requests'
# headless user-agents
- regex: '\b(Windows-Update-Agent|Microsoft-CryptoAPI|SophosUpdateManager|SophosAgent|Debian APT-HTTP|Ubuntu APT-HTTP|libcurl-agent|libwww-perl|urlgrabber|curl|Wget|OpenBSD ftp|jupdate)(?:[ /](\d+)(?:\.(\d+)(?:\.(\d+))?)?)?'
- regex: '(Java)[/ ]{0,1}\d+\.(\d+)\.(\d+)[_-]*([a-zA-Z0-9]+)*' - regex: '(Java)[/ ]{0,1}\d+\.(\d+)\.(\d+)[_-]*([a-zA-Z0-9]+)*'
# Roku Digital-Video-Players https://www.roku.com/ # Roku Digital-Video-Players https://www.roku.com/
- regex: '^(Roku)/DVP-(\d+)\.(\d+)' - regex: '^(Roku)/DVP-(\d+)\.(\d+)'
# Kurio App News Reader https://kurio.co.id/
- regex: '(Kurio)\/(\d+)\.(\d+)\.(\d+)'
family_replacement: 'Kurio App'
os_parsers: os_parsers:
########## ##########
# HbbTV vendors # HbbTV vendors
@ -624,6 +696,10 @@ os_parsers:
- regex: '(Windows Phone) (?:OS[ /])?(\d+)\.(\d+)' - regex: '(Windows Phone) (?:OS[ /])?(\d+)\.(\d+)'
# Again a MS-special one: iPhone.*Outlook-iOS-Android/x.x is erroneously detected as Android
- regex: '(CPU[ +]OS|iPhone[ +]OS|CPU[ +]iPhone)[ +]+(\d+)[_\.](\d+)(?:[_\.](\d+))?.*Outlook-iOS-Android'
os_replacement: 'iOS'
########## ##########
# Android # Android
# can actually detect rooted android os. do we care? # can actually detect rooted android os. do we care?
@ -652,7 +728,7 @@ os_parsers:
# UCWEB # UCWEB
- regex: '^UCWEB.*; (Adr) (\d+)\.(\d+)(?:[.\-]([a-z0-9]+))?;' - regex: '^UCWEB.*; (Adr) (\d+)\.(\d+)(?:[.\-]([a-z0-9]+))?;'
os_replacement: 'Android' os_replacement: 'Android'
- regex: '^UCWEB.*; (iPad OS|iPh OS) (\d+)_(\d+)(?:_(\d+))?;' - regex: '^UCWEB.*; (iPad|iPh|iPd) OS (\d+)_(\d+)(?:_(\d+))?;'
os_replacement: 'iOS' os_replacement: 'iOS'
- regex: '^UCWEB.*; (wds) (\d+)\.(\d+)(?:\.(\d+))?;' - regex: '^UCWEB.*; (wds) (\d+)\.(\d+)(?:\.(\d+))?;'
os_replacement: 'Windows Phone' os_replacement: 'Windows Phone'
@ -735,14 +811,14 @@ os_parsers:
# Tizen OS from Samsung # Tizen OS from Samsung
# spoofs Android so pushing it above # spoofs Android so pushing it above
########## ##########
- regex: '(Tizen)/(\d+)\.(\d+)' - regex: '(Tizen)[/ ](\d+)\.(\d+)'
########## ##########
# Mac OS # Mac OS
# @ref: http://en.wikipedia.org/wiki/Mac_OS_X#Versions # @ref: http://en.wikipedia.org/wiki/Mac_OS_X#Versions
# @ref: http://www.puredarwin.org/curious/versions # @ref: http://www.puredarwin.org/curious/versions
########## ##########
- regex: '((?:Mac ?|; )OS X)[\s/](?:(\d+)[_.](\d+)(?:[_.](\d+))?|Mach-O)' - regex: '((?:Mac[ +]?|; )OS[ +]X)[\s+/](?:(\d+)[_.](\d+)(?:[_.](\d+))?|Mach-O)'
os_replacement: 'Mac OS X' os_replacement: 'Mac OS X'
# Leopard # Leopard
- regex: ' (Dar)(win)/(9).(\d+).*\((?:i386|x86_64|Power Macintosh)\)' - regex: ' (Dar)(win)/(9).(\d+).*\((?:i386|x86_64|Power Macintosh)\)'
@ -789,7 +865,7 @@ os_parsers:
- regex: '(Apple\s?TV)(?:/(\d+)\.(\d+))?' - regex: '(Apple\s?TV)(?:/(\d+)\.(\d+))?'
os_replacement: 'ATV OS X' os_replacement: 'ATV OS X'
- regex: '(CPU OS|iPhone OS|CPU iPhone) +(\d+)[_\.](\d+)(?:[_\.](\d+))?' - regex: '(CPU[ +]OS|iPhone[ +]OS|CPU[ +]iPhone|CPU IPhone OS)[ +]+(\d+)[_\.](\d+)(?:[_\.](\d+))?'
os_replacement: 'iOS' os_replacement: 'iOS'
# remaining cases are mostly only opera uas, so catch opera as to not catch iphone spoofs # remaining cases are mostly only opera uas, so catch opera as to not catch iphone spoofs
@ -836,11 +912,30 @@ os_parsers:
os_replacement: 'Mac OS X' os_replacement: 'Mac OS X'
os_v1_replacement: '10' os_v1_replacement: '10'
os_v2_replacement: '10' os_v2_replacement: '10'
- regex: '(CF)(Network)/(760)\.(\d)'
os_replacement: 'Mac OS X'
os_v1_replacement: '10'
os_v2_replacement: '11'
- regex: '(CF)(Network)/758\.(\d)' - regex: '(CF)(Network)/758\.(\d)'
os_replacement: 'iOS' os_replacement: 'iOS'
os_v1_replacement: '9' os_v1_replacement: '9'
- regex: '(CF)(Network)/808\.(\d)'
os_replacement: 'iOS'
os_v1_replacement: '10'
########## ##########
# CFNetwork macOS Apps (must be before CFNetwork iOS Apps
# @ref: https://en.wikipedia.org/wiki/Darwin_(operating_system)#Release_history
##########
- regex: 'CFNetwork/.* Darwin/16\.\d+.*\(x86_64\)'
os_replacement: 'Mac OS X'
os_v1_replacement: '10'
os_v2_replacement: '12'
- regex: 'CFNetwork/8.* Darwin/15\.\d+.*\(x86_64\)'
os_replacement: 'Mac OS X'
os_v1_replacement: '10'
os_v2_replacement: '11'
##########
# CFNetwork iOS Apps # CFNetwork iOS Apps
# @ref: https://en.wikipedia.org/wiki/Darwin_(operating_system)#Release_history # @ref: https://en.wikipedia.org/wiki/Darwin_(operating_system)#Release_history
########## ##########
@ -867,9 +962,13 @@ os_parsers:
os_replacement: 'iOS' os_replacement: 'iOS'
os_v1_replacement: '9' os_v1_replacement: '9'
os_v2_replacement: '0' os_v2_replacement: '0'
# iOS Apps - regex: 'CFNetwork/8.* Darwin/(16)\.\d+'
- regex: '\b(iOS[ /]|iPhone(?:/| v|[ _]OS[/,]|; | OS : |\d,\d/|\d,\d; )|iPad/)(\d{1,2})[_\.](\d{1,2})(?:[_\.](\d+))?'
os_replacement: 'iOS' os_replacement: 'iOS'
os_v1_replacement: '10'
# iOS Apps
- regex: '\b(iOS[ /]|iOS; |iPhone(?:/| v|[ _]OS[/,]|; | OS : |\d,\d/|\d,\d; )|iPad/)(\d{1,2})[_\.](\d{1,2})(?:[_\.](\d+))?'
os_replacement: 'iOS'
- regex: '\((iOS);'
########## ##########
# Apple TV # Apple TV
@ -994,6 +1093,12 @@ os_parsers:
- regex: '(WebTV)/(\d+).(\d+)' - regex: '(WebTV)/(\d+).(\d+)'
##########
# Chromecast
##########
- regex: '(CrKey)(?:[/](\d+)\.(\d+)(?:\.(\d+))?)?'
os_replacement: 'Chromecast'
########## ##########
# Misc mobile # Misc mobile
########## ##########
@ -1015,7 +1120,7 @@ os_parsers:
- regex: '\((Bada);' - regex: '\((Bada);'
# just os # just os
- regex: '(Windows|Android|WeTab|Maemo)' - regex: '(Windows|Android|WeTab|Maemo|Web0S)'
- regex: '(Ubuntu|Kubuntu|Arch Linux|CentOS|Slackware|Gentoo|openSUSE|SUSE|Red Hat|Fedora|PCLinuxOS|Mageia|(?:Free|Open|Net|\b)BSD)' - regex: '(Ubuntu|Kubuntu|Arch Linux|CentOS|Slackware|Gentoo|openSUSE|SUSE|Red Hat|Fedora|PCLinuxOS|Mageia|(?:Free|Open|Net|\b)BSD)'
# Linux + Kernel Version # Linux + Kernel Version
- regex: '(Linux)(?:[ /](\d+)\.(\d+)(?:\.(\d+))?)?' - regex: '(Linux)(?:[ /](\d+)\.(\d+)(?:\.(\d+))?)?'
@ -1031,7 +1136,7 @@ device_parsers:
# Mobile Spiders # Mobile Spiders
# Catch the mobile crawler before checking for iPhones / Androids. # Catch the mobile crawler before checking for iPhones / Androids.
######### #########
- regex: '(?:(?:iPhone|Windows CE|Android).*(?:(?:Bot|Yeti)-Mobile|YRSpider|bots?/\d|(?:bot|spider)\.html)|AdsBot-Google-Mobile.*iPhone)' - regex: '(?:(?:iPhone|Windows CE|Windows Phone|Android).*(?:(?:Bot|Yeti)-Mobile|YRSpider|BingPreview|bots?/\d|(?:bot|spider)\.html)|AdsBot-Google-Mobile.*iPhone)'
regex_flag: 'i' regex_flag: 'i'
device_replacement: 'Spider' device_replacement: 'Spider'
brand_replacement: 'Spider' brand_replacement: 'Spider'
@ -1925,6 +2030,10 @@ device_parsers:
device_replacement: '$1' device_replacement: '$1'
brand_replacement: 'Google' brand_replacement: 'Google'
model_replacement: '$1' model_replacement: '$1'
- regex: '; *(Pixel \w+) Build'
device_replacement: '$1'
brand_replacement: 'Google'
model_replacement: '$1'
######### #########
# Gigabyte # Gigabyte
@ -2077,7 +2186,7 @@ device_parsers:
device_replacement: '$1' device_replacement: '$1'
brand_replacement: 'Huawei' brand_replacement: 'Huawei'
model_replacement: '$1' model_replacement: '$1'
- regex: '; *([^;]+) Build/Huawei' - regex: '; *([^;]+) Build/(?:Huawei|HUAWEI)'
device_replacement: '$1' device_replacement: '$1'
brand_replacement: 'Huawei' brand_replacement: 'Huawei'
model_replacement: '$1' model_replacement: '$1'
@ -2093,7 +2202,7 @@ device_parsers:
device_replacement: 'Huawei Ideos$1' device_replacement: 'Huawei Ideos$1'
brand_replacement: 'Huawei' brand_replacement: 'Huawei'
model_replacement: 'Ideos$1' model_replacement: 'Ideos$1'
- regex: '; *(Orange Daytona|Pulse|Pulse Mini|Vodafone 858|C8500|C8600|C8650|C8660|Nexus 6P) Build' - regex: '; *(Orange Daytona|Pulse|Pulse Mini|Vodafone 858|C8500|C8600|C8650|C8660|Nexus 6P|ATH-.+?) Build[/ ]'
device_replacement: 'Huawei $1' device_replacement: 'Huawei $1'
brand_replacement: 'Huawei' brand_replacement: 'Huawei'
model_replacement: '$1' model_replacement: '$1'
@ -2635,6 +2744,10 @@ device_parsers:
device_replacement: '$1 $2' device_replacement: '$1 $2'
brand_replacement: 'LG' brand_replacement: 'LG'
model_replacement: '$2' model_replacement: '$2'
- regex: '(Web0S); Linux/(SmartTV)'
device_replacement: 'LG $1 $2'
brand_replacement: 'LG'
model_replacement: '$1 $2'
######### #########
# Malata # Malata
@ -3007,6 +3120,19 @@ device_parsers:
brand_replacement: 'Odys' brand_replacement: 'Odys'
model_replacement: '$1' model_replacement: '$1'
#########
# OnePlus
# @ref https://oneplus.net/
#########
- regex: '; (ONE [a-zA-Z]\d+) Build/'
device_replacement: 'OnePlus $1'
brand_replacement: 'OnePlus'
model_replacement: '$1'
- regex: '; (ONEPLUS [a-zA-Z]\d+) Build/'
device_replacement: 'OnePlus $1'
brand_replacement: 'OnePlus'
model_replacement: '$1'
######### #########
# Orion # Orion
# @ref: http://www.orion.ua/en/products/computer-products/tablet-pcs.html # @ref: http://www.orion.ua/en/products/computer-products/tablet-pcs.html
@ -3247,6 +3373,15 @@ device_parsers:
brand_replacement: 'Quanta' brand_replacement: 'Quanta'
model_replacement: '$1' model_replacement: '$1'
#########
# RCA
# @ref: http://rcamobilephone.com/
#########
- regex: '; (RCT\w+) Build/'
device_replacement: '$1'
brand_replacement: 'RCA'
model_replacement: '$1'
######### #########
# Rockchip # Rockchip
# @ref: http://www.rock-chips.com/a/cn/product/index.html # @ref: http://www.rock-chips.com/a/cn/product/index.html
@ -4152,7 +4287,7 @@ device_parsers:
######### #########
# Noka Windows Phones # Noka Windows Phones
######### #########
- regex: 'Windows Phone [^;]+; .*?IEMobile/[^;\)]+[;\)] ?(?:ARM; ?Touch; ?|Touch; ?)?(?:NOKIA|Nokia)[^;]*; *(?:NOKIA ?|Nokia ?|LUMIA ?|[Ll]umia ?)*(\d{3,}[^;\)]*)' - regex: 'Windows Phone [^;]+; .*?IEMobile/[^;\)]+[;\)] ?(?:ARM; ?Touch; ?|Touch; ?)?(?:rv:11; )?(?:NOKIA|Nokia)[^;]*; *(?:NOKIA ?|Nokia ?|LUMIA ?|[Ll]umia ?)*(\d{3,}[^;\)]*)'
device_replacement: 'Lumia $1' device_replacement: 'Lumia $1'
brand_replacement: 'Nokia' brand_replacement: 'Nokia'
model_replacement: 'Lumia $1' model_replacement: 'Lumia $1'
@ -4330,10 +4465,10 @@ device_parsers:
# cannot determine specific device type from ua string. (3g, 3gs, 4, etc) # cannot determine specific device type from ua string. (3g, 3gs, 4, etc)
########## ##########
# @note: on some ua the device can be identified e.g. iPhone5,1 # @note: on some ua the device can be identified e.g. iPhone5,1
- regex: '((?:iPhone|iPad|iPod)\d+,\d+)' - regex: '(iPhone|iPad|iPod)(\d+,\d+)'
device_replacement: '$1' device_replacement: '$1'
brand_replacement: 'Apple' brand_replacement: 'Apple'
model_replacement: '$1' model_replacement: '$1$2'
# @note: iPad needs to be before iPhone # @note: iPad needs to be before iPhone
- regex: '(iPad)(?:;| Simulator;)' - regex: '(iPad)(?:;| Simulator;)'
device_replacement: '$1' device_replacement: '$1'
@ -4352,6 +4487,12 @@ device_parsers:
device_replacement: '$1$2,$3' device_replacement: '$1$2,$3'
brand_replacement: 'Apple' brand_replacement: 'Apple'
model_replacement: '$1$2,$3' model_replacement: '$1$2,$3'
# @note: newer desktop applications don't show device info
# This is here so as to not have them recorded as iOS-Device
- regex: 'CFNetwork/.* Darwin/\d+\.\d+\.\d+ \(x86_64\)'
device_replacement: 'Mac'
brand_replacement: 'Apple'
model_replacement: 'Mac'
# @note: iOS applications do not show device info # @note: iOS applications do not show device info
- regex: 'CFNetwork/.* Darwin/\d' - regex: 'CFNetwork/.* Darwin/\d'
device_replacement: 'iOS-Device' device_replacement: 'iOS-Device'
@ -4617,6 +4758,11 @@ device_parsers:
########## ##########
# Samsung # Samsung
########## ##########
# Samsung Smart-TV
- regex: '(SMART-TV); .* Tizen '
device_replacement: 'Samsung $1'
brand_replacement: 'Samsung'
model_replacement: '$1'
# Samsung Symbian Devices # Samsung Symbian Devices
- regex: 'SymbianOS/9\.\d.* Samsung[/\-]([A-Za-z0-9 \-]+)' - regex: 'SymbianOS/9\.\d.* Samsung[/\-]([A-Za-z0-9 \-]+)'
device_replacement: 'Samsung $1' device_replacement: 'Samsung $1'
@ -4707,27 +4853,27 @@ device_parsers:
######### #########
# Android General Device Matching (far from perfect) # Android General Device Matching (far from perfect)
######### #########
- regex: 'Android[\- ][\d]+\.[\d]+; [A-Za-z]{2}\-[A-Za-z]{0,2}; WOWMobile (.+) Build' - regex: 'Android[\- ][\d]+\.[\d]+; [A-Za-z]{2}\-[A-Za-z]{0,2}; WOWMobile (.+) Build[/ ]'
brand_replacement: 'Generic_Android' brand_replacement: 'Generic_Android'
model_replacement: '$1' model_replacement: '$1'
- regex: 'Android[\- ][\d]+\.[\d]+\-update1; [A-Za-z]{2}\-[A-Za-z]{0,2} *; *(.+?) Build' - regex: 'Android[\- ][\d]+\.[\d]+\-update1; [A-Za-z]{2}\-[A-Za-z]{0,2} *; *(.+?) Build[/ ]'
brand_replacement: 'Generic_Android' brand_replacement: 'Generic_Android'
model_replacement: '$1' model_replacement: '$1'
- regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}; *[A-Za-z]{2}[_\-][A-Za-z]{0,2}\-? *; *(.+?) Build' - regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}; *[A-Za-z]{2}[_\-][A-Za-z]{0,2}\-? *; *(.+?) Build[/ ]'
brand_replacement: 'Generic_Android' brand_replacement: 'Generic_Android'
model_replacement: '$1' model_replacement: '$1'
- regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}; *[A-Za-z]{0,2}\- *; *(.+?) Build' - regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}; *[A-Za-z]{0,2}\- *; *(.+?) Build[/ ]'
brand_replacement: 'Generic_Android' brand_replacement: 'Generic_Android'
model_replacement: '$1' model_replacement: '$1'
# No build info at all - "Build" follows locale immediately # No build info at all - "Build" follows locale immediately
- regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}; *[a-z]{0,2}[_\-]?[A-Za-z]{0,2};? Build' - regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}; *[a-z]{0,2}[_\-]?[A-Za-z]{0,2};? Build[/ ]'
device_replacement: 'Generic Smartphone' device_replacement: 'Generic Smartphone'
brand_replacement: 'Generic' brand_replacement: 'Generic'
model_replacement: 'Smartphone' model_replacement: 'Smartphone'
- regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}; *\-?[A-Za-z]{2}; *(.+?) Build' - regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}; *\-?[A-Za-z]{2}; *(.+?) Build[/ ]'
brand_replacement: 'Generic_Android' brand_replacement: 'Generic_Android'
model_replacement: '$1' model_replacement: '$1'
- regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}(?:;.*)?; *(.+?) Build' - regex: 'Android[\- ][\d]+(?:\.[\d]+){1,2}(?:;.*)?; *(.+?) Build[/ ]'
brand_replacement: 'Generic_Android' brand_replacement: 'Generic_Android'
model_replacement: '$1' model_replacement: '$1'
@ -4774,7 +4920,7 @@ device_parsers:
########## ##########
# Spiders (this is hack...) # Spiders (this is hack...)
########## ##########
- regex: '(bot|zao|borg|DBot|oegp|silk|Xenu|zeal|^NING|CCBot|crawl|htdig|lycos|slurp|teoma|voila|yahoo|Sogou|CiBra|Nutch|^Java/|^JNLP/|Daumoa|Genieo|ichiro|larbin|pompos|Scrapy|snappy|speedy|spider|msnbot|msrbot|vortex|^vortex|crawler|favicon|indexer|Riddler|scooter|scraper|scrubby|WhatWeb|WinHTTP|bingbot|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.*/\+/web/snippet|Google-HTTP-Java-Client|BlogBridge|IlTrovatore-Setaccio|InternetArchive|GomezAgent|WebThumbnail|heritrix|NewsGator|PagePeeker|Reaper|ZooShot|holmes)' - regex: '(bot|zao|borg|DBot|oegp|silk|Xenu|zeal|^NING|CCBot|crawl|htdig|lycos|slurp|teoma|voila|yahoo|Sogou|CiBra|Nutch|^Java/|^JNLP/|Daumoa|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.*/\+/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)'
regex_flag: 'i' regex_flag: 'i'
device_replacement: 'Spider' device_replacement: 'Spider'
brand_replacement: 'Spider' brand_replacement: 'Spider'