Advertisement
Guest User

CustomScraper

a guest
Aug 13th, 2013
88
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
XML 19.24 KB | None | 0 0
  1. <?xml version="1.0" encoding="UTF-8"?>
  2. <!-- should be self-explanatory -->
  3. <scraper framework="1.1" date="2013-04-04">
  4.  
  5.     <!-- input: $1=query string -->
  6.     <!-- returns:   the url we should use to do the search -->
  7.     <CreateSearchUrl dest="3">
  8.         <RegExp input="$$1" output="&lt;url cache=&quot;cache-\1$$4.xml&quot;&gt;http://documentalis.com:9999/api/documentary/find?docname=\1$$4&lt;/url&gt;" dest="3">
  9.             <RegExp input="$$2" output="%20(\1)" dest="4">
  10.                 <expression clear="yes">(.+)</expression>
  11.             </RegExp>
  12.             <expression/>
  13.         </RegExp>
  14.     </CreateSearchUrl>
  15.  
  16.     <!-- input: $1=html $2=search query -->
  17.     <!-- returns:   results in xml format <results><movie><title>*</title><url>*</url>*#urls<extra>*</extra></movie>*</results> -->
  18.     <GetSearchResults dest="1">
  19.         <RegExp input="$$4" output="&lt;?xml version=&quot;1.0&quot; encoding=&quot;utf-8&quot; standalone=&quot;yes&quot;?&gt;&lt;results&gt;\1&lt;/results&gt;" dest="1">
  20.             <RegExp input="$$1" output="&lt;entity&gt;&lt;title&gt;\3&lt;/title&gt;&lt;language&gt;\2&lt;/language&gt;&lt;url cache=&quot;\1-\2.xml&quot;&gt;http://documentalis.com:9999/api/documentary/\1/all/\2.zip&lt;/url&gt;&lt;id&gt;\1&lt;/id&gt;&lt;/entity&gt;" dest="4">
  21.                 <expression repeat="yes">&lt;seriesid&gt;([0-9]*)&lt;/seriesid&gt;[^&lt;]*&lt;language&gt;([^&lt;]*)&lt;/language&gt;[^&lt;]*&lt;SeriesName&gt;([^&lt;]*)&lt;/SeriesName&gt;</expression>
  22.             </RegExp>
  23.             <expression noclean="1"/>
  24.         </RegExp>
  25.     </GetSearchResults>
  26.  
  27.     <!-- input: $1..#urls=html -->
  28.     <!-- input: $(#urls+1): extra !-->
  29.     <!-- returns:   results in xml format <details><plot>*</plot><director>*</director><premiered>*</premiered><episodeguide>*</episodeguide></details> -->
  30.     <GetDetails dest="7">
  31.         <RegExp input="$$4" output="&lt;?xml version=&quot;1.0&quot; encoding=&quot;utf-8&quot; standalone=&quot;yes&quot;?&gt;&lt;details&gt;\1&lt;/details&gt;" dest="7">
  32.             <RegExp input="$$1" output="\1" dest="5">
  33.                 <expression noclean="1">&lt;Series&gt;.*?&lt;id&gt;$$2&lt;/id&gt;(.*)</expression>
  34.             </RegExp>
  35.             <RegExp input="$$5" output="&lt;plot&gt;\1&lt;/plot&gt;" dest="4">
  36.                 <expression noclean="1">&lt;Overview&gt;([^&lt;]*)&lt;/Overview&gt;</expression>
  37.             </RegExp>
  38.             <RegExp input="$$2" output="&lt;id&gt;\1&lt;/id&gt;" dest="4+">
  39.                 <expression/>
  40.             </RegExp>
  41.             <RegExp input="$$1" output="&lt;mpaa&gt;\1&lt;/mpaa&gt;" dest="4+">
  42.                 <expression>&lt;ContentRating&gt;([^&lt;]*)&lt;/ContentRating&gt;</expression>
  43.             </RegExp>
  44.             <RegExp input="$$5" output="&lt;premiered&gt;\1&lt;/premiered&gt;" dest="4+">
  45.                 <expression>&lt;FirstAired&gt;([^&lt;]*)&lt;/FirstAired&gt;</expression>
  46.             </RegExp>
  47.             <RegExp input="$$5" output="&lt;rating&gt;\1&lt;/rating&gt;" dest="4+">
  48.                 <expression>&lt;Rating&gt;([^&lt;]*)&lt;/Rating&gt;</expression>
  49.             </RegExp>
  50.             <RegExp input="$$5" output="&lt;studio&gt;\1&lt;/studio&gt;" dest="4+">
  51.                 <expression>&lt;Network&gt;([^&lt;]*)&lt;/Network&gt;</expression>
  52.             </RegExp>
  53.             <RegExp input="$$5" output="&lt;title&gt;\1&lt;/title&gt;" dest="4+">
  54.                 <expression>&lt;SeriesName&gt;([^&lt;]*)&lt;/SeriesName&gt;</expression>
  55.             </RegExp>
  56.             <RegExp input="$$5" output="\1" dest="10">
  57.                 <expression noclean="1">&lt;Genre&gt;([^&lt;]*)&lt;/Genre&gt;</expression>
  58.             </RegExp>
  59.             <RegExp input="$$10" output="&lt;genre&gt;\1&lt;/genre&gt;" dest="4+">
  60.                 <expression repeat="yes">([^\|]*)\|</expression>
  61.             </RegExp>
  62.             <RegExp input="$$10" output="&lt;genre&gt;\1&lt;/genre&gt;" dest="4+">
  63.                 <expression repeat="yes">([^,]*),</expression>
  64.             </RegExp>
  65.             <RegExp input="$$10" output="&lt;genre&gt;\1&lt;/genre&gt;" dest="4+">
  66.                 <expression>([^\|,]+)$</expression>
  67.             </RegExp>
  68.             <!-- actors with thumbs -->
  69.             <RegExp input="$$5" output="&lt;actor&gt;&lt;name&gt;\2&lt;/name&gt;&lt;role&gt;\3&lt;/role&gt;&lt;thumb&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;&lt;/actor&gt;" dest="4+">
  70.                 <expression repeat="yes" noclean="1,2,3">&lt;Actor&gt;.*?&lt;Image&gt;([^&lt;]+)&lt;/Image&gt;.*?&lt;Name&gt;([^&lt;]*)&lt;/Name&gt;.*?&lt;Role&gt;([^&lt;]*)</expression>
  71.             </RegExp>
  72.             <!-- actors without thumbs -->
  73.             <RegExp input="$$5" output="&lt;actor&gt;&lt;name&gt;\2&lt;/name&gt;&lt;role&gt;\3&lt;/role&gt;&lt;/actor&gt;" dest="4+">
  74.                 <expression repeat="yes" noclean="1,2,3">&lt;Actor&gt;.*?&lt;Image&gt;([^&lt;]*)&lt;/Image&gt;.*?&lt;Name&gt;([^&lt;]*)&lt;/Name&gt;.*?&lt;Role&gt;([^&lt;]*)</expression>
  75.             </RegExp>
  76.             <!-- documentary banner -->
  77.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;banner&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  78.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;series&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;graphical&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;$INFO[language]&lt;/Language&gt;</expression>
  79.             </RegExp>
  80.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;banner&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  81.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;series&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;graphical&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;((?!$INFO[language])[a-z])*&lt;/Language&gt;</expression>
  82.             </RegExp>
  83.             <!-- documentary banner text -->
  84.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;banner&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  85.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;series&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;text&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;$INFO[language]&lt;/Language&gt;</expression>
  86.             </RegExp>
  87.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;banner&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  88.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;series&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;text&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;((?!$INFO[language])[a-z])*&lt;/Language&gt;</expression>
  89.             </RegExp>
  90.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;banner&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  91.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;series&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;blank&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;&lt;/Language&gt;</expression>
  92.             </RegExp>
  93.             <!-- documentary season covers -->
  94.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;poster&quot; type=&quot;season&quot; season=&quot;\2&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  95.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;season&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;season&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;$INFO[language]&lt;/Language&gt;[^&lt;]*[^S]*Season&gt;([0-9]+)&lt;/Season&gt;</expression>
  96.             </RegExp>
  97.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;poster&quot; type=&quot;season&quot; season=&quot;\3&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  98.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;season&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;season&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;((?!$INFO[language])[a-z])*&lt;/Language&gt;[^&lt;]*[^S]*Season&gt;([0-9]+)&lt;/Season&gt;</expression>
  99.             </RegExp>
  100.             <!-- documentary season wide -->
  101.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;banner&quot; type=&quot;season&quot; season=&quot;\2&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  102.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;season&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;seasonwide&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;$INFO[language]&lt;/Language&gt;[^&lt;]*[^S]*Season&gt;([0-9]+)&lt;/Season&gt;</expression>
  103.             </RegExp>
  104.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;banner&quot; type=&quot;season&quot; season=&quot;\3&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  105.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;season&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;seasonwide&lt;/BannerType2&gt;[^&lt;]*&lt;Language&gt;((?!$INFO[language])[a-z])*&lt;/Language&gt;[^&lt;]*[^S]*Season&gt;([0-9]+)&lt;/Season&gt;</expression>
  106.             </RegExp>
  107.             <!-- documentary posters -->
  108.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;poster&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  109.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;poster&lt;/BannerType&gt;</expression>
  110.             </RegExp>
  111.             <RegExp input="$$5" output="&lt;thumb aspect=&quot;poster&quot; type=&quot;season&quot; season=&quot;-1&quot;&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  112.                 <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;poster&lt;/BannerType&gt;</expression>
  113.             </RegExp>
  114.             <!-- documentary fanarts -->
  115.             <RegExp conditional="fanart" input="$$7" output="&lt;fanart url=&quot;http://documentalis.com:9999/api/assets/&quot;&gt;\1&lt;/fanart&gt;" dest="4+">
  116.                 <RegExp input="$$5" output="&lt;thumb dim=&quot;\2&quot; colors=&quot;\3&quot; preview=&quot;_cache/\1&quot;&gt;\1&lt;/thumb&gt;" dest="7+">
  117.                     <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;fanart&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;([^&lt;]*)&lt;/BannerType2&gt;[^&lt;]*&lt;Colors&gt;([^&lt;]*)&lt;/Colors&gt;[^&lt;]*&lt;Language&gt;$INFO[language]&lt;/Language&gt;</expression>
  118.                 </RegExp>
  119.                 <RegExp input="$$5" output="&lt;thumb dim=&quot;\2&quot; colors=&quot;\3&quot; preview=&quot;_cache/\1&quot;&gt;\1&lt;/thumb&gt;" dest="7+">
  120.                     <expression repeat="yes">&lt;BannerPath&gt;([^&lt;]*)&lt;/BannerPath&gt;[^&lt;]*&lt;BannerType&gt;fanart&lt;/BannerType&gt;[^&lt;]*&lt;BannerType2&gt;([^&lt;]*)&lt;/BannerType2&gt;[^&lt;]*&lt;Colors&gt;([^&lt;]*)&lt;/Colors&gt;[^&lt;]*&lt;Language&gt;((?!$INFO[language])[a-z])*&lt;/Language&gt;</expression>
  121.                 </RegExp>
  122.                 <expression noclean="1"/>
  123.             </RegExp>
  124.             <RegExp input="$$3" output="\1" dest="6">
  125.                 <expression>.*/(.*).zip</expression>
  126.             </RegExp>
  127.             <RegExp input="$$3" output="&lt;episodeguide&gt;&lt;url cache=&quot;$$2-$$6.xml&quot;&gt;\1&lt;/url&gt;&lt;/episodeguide&gt;" dest="4+">
  128.                 <expression/>
  129.             </RegExp>
  130.             <expression noclean="1"/>
  131.         </RegExp>
  132.     </GetDetails>
  133.  
  134.     <!-- input: $1=html !-->
  135.     <!-- input: $2=series url !-->
  136.     <!-- returns:   results in xml format <episodeguide><episode><title>*</title><url>*</url><season>*</season><epnum>*</epnum><thumb>*</thumb><id>*</id><aired>*</aired></episode>*</episodeguide> !-->
  137.     <GetEpisodeList dest="3">
  138.         <RegExp input="$$4" output="&lt;episodeguide&gt;\1&lt;/episodeguide&gt;" dest="3">
  139.             <RegExp input="$$2" output="\2-\3" dest="10">
  140.                 <expression>http://documentalis.com:9999/api/documentary/([0-9]*)/all/(.+).zip</expression>
  141.             </RegExp>
  142.             <RegExp conditional="!dvdorder">
  143.                 <!-- Regular episodes (Absolute order) -->
  144.                 <RegExp conditional="absolutenumber" input="$$1" output="&lt;episode&gt;&lt;id&gt;\1&lt;/id&gt;&lt;title&gt;\2&lt;/title&gt;&lt;aired&gt;\3&lt;/aired&gt;&lt;epnum&gt;\4&lt;/epnum&gt;&lt;season&gt;1&lt;/season&gt;&lt;url cache=&quot;$$10.xml&quot;&gt;$$2&lt;/url&gt;&lt;/episode&gt;" dest="4+">
  145.                     <expression repeat="yes">&lt;Episode&gt;.*?&lt;id&gt;([0-9]*).*?&lt;EpisodeName&gt;([^&lt;]*).*?&lt;FirstAired&gt;([^&lt;]*).*?&lt;absolute_number&gt;([0-9]*).*?&lt;/Episode&gt;</expression>
  146.                 </RegExp>
  147.                 <!-- Specials (Absolute order) -->
  148.                 <RegExp conditional="absolutenumber" input="$$1" output="&lt;episode&gt;&lt;id&gt;\1&lt;/id&gt;&lt;title&gt;\2&lt;/title&gt;&lt;aired&gt;\4&lt;/aired&gt;&lt;epnum&gt;\3&lt;/epnum&gt;&lt;season&gt;0&lt;/season&gt;&lt;url cache=&quot;$$10.xml&quot;&gt;$$2&lt;/url&gt;&lt;/episode&gt;" dest="4+">
  149.                     <expression repeat="yes">&lt;Episode&gt;.*?&lt;id&gt;([0-9]*).*?&lt;EpisodeName&gt;([^&lt;]*).*?&lt;EpisodeNumber&gt;([0-9]*).*?&lt;FirstAired&gt;([^&lt;]*).*?&lt;SeasonNumber&gt;0&lt;/SeasonNumber&gt;.*?&lt;/Episode&gt;</expression>
  150.                 </RegExp>
  151.                 <!-- Normal episodes -->
  152.                 <RegExp conditional="!absolutenumber" input="$$1" output="&lt;episode&gt;&lt;title&gt;\2&lt;/title&gt;&lt;url cache=&quot;$$10.xml&quot;&gt;$$2&lt;/url&gt;&lt;epnum&gt;\3&lt;/epnum&gt;&lt;season&gt;\5&lt;/season&gt;&lt;id&gt;\1&lt;/id&gt;&lt;aired&gt;\4&lt;/aired&gt;&lt;/episode&gt;" dest="4">
  153.                     <expression repeat="yes">&lt;Episode&gt;.*?&lt;id&gt;([0-9]+).*?&lt;EpisodeName&gt;([^&lt;]*).*?&lt;EpisodeNumber&gt;([0-9]+)[^&lt;]*.*?&lt;FirstAired&gt;([^&lt;]*)&lt;/FirstAired&gt;.*?&lt;SeasonNumber&gt;([0-9]+)[^&lt;]*.*?&lt;/Episode&gt;</expression>
  154.                 </RegExp>
  155.             </RegExp>
  156.             <!-- DVD order -->
  157.             <RegExp conditional="dvdorder" input="$$1" output="&lt;episode&gt;&lt;title&gt;\4&lt;/title&gt;&lt;url cache=&quot;$$10.xml&quot;&gt;$$2&lt;/url&gt;&lt;epnum&gt;\2&lt;/epnum&gt;&lt;season&gt;\3&lt;/season&gt;&lt;id&gt;\1&lt;/id&gt;&lt;aired&gt;\5&lt;/aired&gt;&lt;/episode&gt;" dest="4">
  158.                 <expression repeat="yes">&lt;Episode&gt;.*?&lt;id&gt;([0-9]+).*?&lt;Combined_episodenumber&gt;([^&lt;]*).*?&lt;Combined_season&gt;([^&lt;]*).*?&lt;EpisodeName&gt;([^&lt;]*).*?&lt;FirstAired&gt;([^&lt;]*)&lt;/FirstAired&gt;.*?&lt;/Episode&gt;</expression>
  159.             </RegExp>
  160.             <expression noclean="1"/>
  161.         </RegExp>
  162.     </GetEpisodeList>
  163.  
  164.     <!-- input: $1=html -->
  165.     <!-- returns:   results in xml format <details><writer>*</writer><director>*</director><cast>*</cast><rating>*</rating><rank>*</rank><plot>*</plot> -->
  166.     <GetEpisodeDetails dest="3">
  167.         <RegExp input="$$4" output="&lt;?xml version=&quot;1.0&quot; encoding=&quot;utf-8&quot; standalone=&quot;yes&quot;?&gt;&lt;details&gt;\1&lt;/details&gt;" dest="3">
  168.             <RegExp input="$$1" output="\1" dest="8">
  169.                 <expression noclean="1">&lt;Episode&gt;.*?&lt;id&gt;$$2&lt;/id&gt;(.*?)&lt;/Episode&gt;</expression>
  170.             </RegExp>
  171.             <RegExp input="$$1" output="&lt;uniqueid&gt;$$2&lt;/uniqueid&gt;" dest="4">
  172.                 <expression/>
  173.             </RegExp>
  174.             <RegExp input="$$8" output="&lt;plot&gt;\1&lt;/plot&gt;" dest="4+">
  175.                 <expression>&lt;Overview&gt;([^&lt;]*)&lt;/Overview&gt;</expression>
  176.             </RegExp>
  177.             <RegExp input="$$8" output="\1" dest="6">
  178.                 <expression noclean="1">&lt;Writer&gt;([^&lt;]*)&lt;/Writer&gt;</expression>
  179.             </RegExp>
  180.             <RegExp input="$$6" output="&lt;credits&gt;\1&lt;/credits&gt;" dest="4+">
  181.                 <expression repeat="yes">([^\|]*)\|</expression>
  182.             </RegExp>
  183.             <RegExp input="$$6" output="&lt;credits&gt;\1&lt;/credits&gt;" dest="4+">
  184.                 <expression repeat="yes">([^,]*),</expression>
  185.             </RegExp>
  186.             <RegExp input="$$6" output="&lt;credits&gt;\1&lt;/credits&gt;" dest="4+">
  187.                 <expression>([^\|,]+)$</expression>
  188.             </RegExp>
  189.             <RegExp input="$$8" output="\1" dest="6">
  190.                 <expression noclean="1">&lt;Director&gt;([^&lt;]*)&lt;/Director&gt;</expression>
  191.             </RegExp>
  192.             <RegExp input="$$6" output="&lt;director&gt;\1&lt;/director&gt;" dest="4+">
  193.                 <expression repeat="yes">([^\|]*)\|</expression>
  194.             </RegExp>
  195.             <RegExp input="$$6" output="&lt;director&gt;\1&lt;/director&gt;" dest="4+">
  196.                 <expression repeat="yes">([^,]*),</expression>
  197.             </RegExp>
  198.             <RegExp input="$$6" output="&lt;director&gt;\1&lt;/director&gt;" dest="4+">
  199.                 <expression>([^\|,]+)$</expression>
  200.             </RegExp>
  201.             <RegExp input="$$8" output="\1" dest="6">
  202.                 <expression noclean="1">&lt;GuestStars&gt;([^&lt;]*)&lt;/GuestStars&gt;</expression>
  203.             </RegExp>
  204.             <RegExp input="$$6" output="&lt;actor&gt;&lt;name&gt;\1&lt;/name&gt;&lt;/actor&gt;" dest="4+">
  205.                 <expression repeat="yes">([^\|]*)\|</expression>
  206.             </RegExp>
  207.             <RegExp input="$$6" output="&lt;actor&gt;&lt;name&gt;\1&lt;/name&gt;&lt;/actor&gt;" dest="4+">
  208.                 <expression repeat="yes">([^,]*),</expression>
  209.             </RegExp>
  210.             <RegExp input="$$6" output="&lt;actor&gt;&lt;name&gt;\1&lt;/name&gt;&lt;/actor&gt;" dest="4+">
  211.                 <expression>([^\|,]+)$</expression>
  212.             </RegExp>
  213.             <RegExp input="$$8" output="&lt;title&gt;\1&lt;/title&gt;" dest="4+">
  214.                 <expression>&lt;EpisodeName&gt;([^&lt;]*)&lt;/EpisodeName&gt;</expression>
  215.             </RegExp>
  216.             <!-- Regular Episodes - Absolute Order -->
  217.             <RegExp conditional="absolutenumber" input="$$8" output="&lt;season&gt;1&lt;/season&gt;&lt;episode&gt;\2&lt;/episode&gt;" dest="4+">
  218.                 <expression>&lt;SeasonNumber&gt;([1-9]*)&lt;/SeasonNumber&gt;.*?&lt;absolute_number&gt;([0-9]*)&lt;/absolute_number&gt;</expression>
  219.             </RegExp>
  220.             <!-- Specials - Absolute Order -->
  221.             <RegExp conditional="absolutenumber" input="$$8" output="&lt;season&gt;0&lt;/season&gt;&lt;episode&gt;\1&lt;/episode&gt;" dest="4+">
  222.                 <expression>&lt;EpisodeNumber&gt;([0-9]*)&lt;/EpisodeNumber&gt;.*?&lt;SeasonNumber&gt;0&lt;/SeasonNumber&gt;</expression>
  223.             </RegExp>
  224.             <!-- Normal Season/Episodes -->
  225.             <RegExp conditional="!absolutenumber" input="$$9" output="&lt;season&gt;\1&lt;/season&gt;" dest="4+">
  226.                 <RegExp conditional="!dvdorder" input="$$8" output="\1" dest="9">
  227.                     <expression clear="yes">&lt;SeasonNumber&gt;([^&lt;]*)&lt;/SeasonNumber&gt;</expression>
  228.                 </RegExp>
  229.                 <RegExp conditional="dvdorder" input="$$8" output="\1" dest="9">
  230.                     <expression>&lt;Combined_season&gt;([^&lt;]*)&lt;/Combined_season&gt;</expression>
  231.                 </RegExp>
  232.                 <expression/>
  233.             </RegExp>
  234.             <RegExp conditional="!absolutenumber" input="$$9" output="&lt;episode&gt;\1&lt;/episode&gt;" dest="4+">
  235.                 <RegExp conditional="!dvdorder" input="$$8" output="\1" dest="9">
  236.                     <expression clear="yes">&lt;EpisodeNumber&gt;([^&lt;]*)&lt;/EpisodeNumber&gt;</expression>
  237.                 </RegExp>
  238.                 <RegExp conditional="dvdorder" input="$$8" output="\1" dest="9">
  239.                     <expression>&lt;Combined_episodenumber&gt;([^&lt;]*)&lt;/Combined_episodenumber&gt;</expression>
  240.                 </RegExp>
  241.                 <expression/>
  242.             </RegExp>
  243.             <!-- episode filenames -->
  244.             <RegExp input="$$8" output="&lt;thumb&gt;http://documentalis.com:9999/api/assets/\1&lt;/thumb&gt;" dest="4+">
  245.                 <expression>&lt;filename&gt;([^&lt;]+)&lt;/filename&gt;</expression>
  246.             </RegExp>
  247.             <RegExp input="$$8" output="&lt;aired&gt;\1&lt;/aired&gt;" dest="4+">
  248.                 <expression>&lt;FirstAired&gt;([^&lt;]+)&lt;/FirstAired&gt;</expression>
  249.             </RegExp>
  250.             <RegExp input="$$8" output="&lt;displayseason&gt;\1&lt;/displayseason&gt;" dest="4+">
  251.                 <expression>&lt;airsbefore_season&gt;([^&lt;]+)&lt;/airsbefore_season&gt;</expression>
  252.             </RegExp>
  253.             <RegExp input="$$8" output="&lt;displayepisode&gt;\1&lt;/displayepisode&gt;" dest="4+">
  254.                 <expression>&lt;airsbefore_episode&gt;([^&lt;]+)&lt;/airsbefore_episode&gt;</expression>
  255.             </RegExp>
  256.             <RegExp input="$$8" output="&lt;displayafterseason&gt;\1&lt;/displayafterseason&gt;" dest="4+">
  257.                 <expression>&lt;airsafter_season&gt;([^&lt;]+)&lt;/airsafter_season&gt;</expression>
  258.             </RegExp>
  259.             <RegExp input="$$8" output="&lt;rating&gt;\1&lt;/rating&gt;" dest="4+">
  260.                 <expression>&lt;Rating&gt;([^&lt;]+)&lt;/Rating&gt;</expression>
  261.             </RegExp>
  262.             <RegExp input="$$1" output="&lt;runtime&gt;\1&lt;/runtime&gt;" dest="4+">
  263.                 <expression>&lt;Runtime&gt;([^&lt;]+)&lt;/Runtime&gt;</expression>
  264.             </RegExp>
  265.             <expression noclean="1"/>
  266.         </RegExp>
  267.     </GetEpisodeDetails>
  268. </scraper>
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement