aboutsummaryrefslogtreecommitdiff
path: root/system/scrapers/video
diff options
context:
space:
mode:
Diffstat (limited to 'system/scrapers/video')
-rw-r--r--system/scrapers/video/daum.xml282
1 files changed, 185 insertions, 97 deletions
diff --git a/system/scrapers/video/daum.xml b/system/scrapers/video/daum.xml
index 896c748d45..c4335ebe3d 100644
--- a/system/scrapers/video/daum.xml
+++ b/system/scrapers/video/daum.xml
@@ -1,187 +1,275 @@
<?xml version="1.0" encoding="UTF-8"?>
-<scraper framework="1.1" date="2009-08-08" name="Daum" content="movies" thumb="daum.png" language="ko">
+<scraper framework="1.1" date="2009-12-28" name="Daum" content="movies" thumb="daum.png" language="ko">
<GetSettings dest="3">
<RegExp input="$$5" output="&lt;settings&gt;\1&lt;/settings&gt;" dest="3">
<RegExp input="$$1" output="&lt;setting label=&quot;포스터&quot; type=&quot;bool&quot; id=&quot;EnableThumbs&quot; default=&quot;true&quot;&gt;&lt;/setting&gt;" dest="5+">
<expression/>
</RegExp>
- <RegExp input="$$1" output="&lt;setting label=&quot;팬아트(고화질)&quot; type=&quot;bool&quot; id=&quot;EnableFanArt&quot; default=&quot;true&quot;&gt;&lt;/setting&gt;" dest="5+">
+ <RegExp input="$$1" output="&lt;setting label=&quot;팬아트&quot; type=&quot;bool&quot; id=&quot;EnableFanart&quot; default=&quot;true&quot;&gt;&lt;/setting&gt;" dest="5+">
+ <expression/>
+ </RegExp>
+ <RegExp input="$$1" output="&lt;setting label=&quot;TMDB팬아트 우선사용&quot; type=&quot;bool&quot; id=&quot;EnableTMDBFanart&quot; default=&quot;false&quot;&gt;&lt;/setting&gt;" dest="5+">
<expression/>
</RegExp>
<RegExp input="$$1" output="&lt;setting label=&quot;극작가&quot; type=&quot;bool&quot; id=&quot;EnableWriter&quot; default=&quot;true&quot;&gt;&lt;/setting&gt;" dest="5+">
<expression/>
</RegExp>
+ <RegExp input="$$1" output="&lt;setting label=&quot;전체줄거리&quot; type=&quot;bool&quot; id=&quot;EnableFullPlot&quot; default=&quot;true&quot;&gt;&lt;/setting&gt;" dest="5+">
+ <expression/>
+ </RegExp>
+ <RegExp input="$$1" output="&lt;setting label=&quot;배역 더보기&quot; type=&quot;bool&quot; id=&quot;EnableMoreCast&quot; default=&quot;true&quot;&gt;&lt;/setting&gt;" dest="5+">
+ <expression/>
+ </RegExp>
<expression noclean="1"/>
</RegExp>
</GetSettings>
<NfoUrl dest="3">
<RegExp input="$$1" output="&lt;url&gt;http://movie.daum.net/moviedetail/moviedetailMain.do?movieId=\1&lt;/url&gt;" dest="3">
- <expression>movie\.daum\.net[^\?]*\?movieId=([0-9]*)</expression>
+ <expression>movie\.daum\.net[^\?]*\?movieId=(\d*)</expression>
</RegExp>
</NfoUrl>
- <CreateSearchUrl dest="3" SearchStringEncoding="UTF-8">
- <RegExp input="$$1" output="http://movie.daum.net/search.do?type=movie&amp;q=\1" dest="3">
+ <CreateSearchUrl SearchStringEncoding="UTF-8" dest="3">
+ <RegExp input="$$1" output="&lt;url&gt;http://movie.daum.net/search.do?type=movie&amp;q=\1&lt;/url&gt;" dest="3">
<expression noclean="1"/>
</RegExp>
</CreateSearchUrl>
<GetSearchResults dest="8">
<RegExp input="$$5" 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="8">
- <RegExp input="$$1" output="&lt;entity&gt;&lt;title&gt;\2(\4,\3) &lt;/title&gt;&lt;url&gt;http://movie.daum.net/moviedetail/moviedetailMain.do?movieId=\1&lt;/url&gt;&lt;/entity&gt;" dest="5">
- <expression repeat="yes">class=&quot;fl srch&quot;&gt;&lt;a href=&quot;http://movie.daum.net/moviedetail/moviedetailMain.do\?movieId=([0-9]+)&quot;[^&gt;]*&gt;([^%]*)&lt;/a&gt;[ ]*\(([0-9]+)\)&lt;/span&gt;[^\(]*&quot;fs13&quot;&gt;([^&quot;]*)&lt;/span</expression>
+ <RegExp input="$$1" output="&lt;entity&gt;&lt;title&gt;\2&lt;/title&gt;&lt;year&gt;\3&lt;/year&gt;&lt;id&gt;\1&lt;/id&gt;&lt;url&gt;http://movie.daum.net/moviedetail/moviedetailMain.do?movieId=\1&lt;/url&gt;&lt;/entity&gt;" dest="5">
+ <expression clear="yes" repeat="yes">class="fl srch"&gt;&lt;a href="http://movie.daum.net/moviedetail/moviedetailMain.do\?movieId=(\d+)"[^&gt;]*&gt;([^"]*)&lt;/a&gt;\s*\((\d+)\)&lt;/span&gt;[^\(]*"fs13"&gt;\s*&lt;/span</expression>
+ </RegExp>
+ <RegExp input="$$1" output="&lt;entity&gt;&lt;title&gt;\2(\4)&lt;/title&gt;&lt;year&gt;\3&lt;/year&gt;&lt;id&gt;\1&lt;/id&gt;&lt;url&gt;http://movie.daum.net/moviedetail/moviedetailMain.do?movieId=\1&lt;/url&gt;&lt;/entity&gt;" dest="5+">
+ <expression repeat="yes">class="fl srch"&gt;&lt;a href="http://movie.daum.net/moviedetail/moviedetailMain.do\?movieId=(\d+)"[^&gt;]*&gt;([^"]*)&lt;/a&gt;\s*\((\d+)\)&lt;/span&gt;[^\(]*"fs13"&gt;\s*([^"]+)&lt;/span</expression>
</RegExp>
<expression noclean="1"/>
</RegExp>
</GetSearchResults>
<GetDetails dest="3">
<RegExp input="$$8" output="&lt;details&gt;\1&lt;/details&gt;" dest="3">
- <!-- Titles, year !-->
- <RegExp input="$$1" output="&lt;title&gt;\1&lt;/title&gt;&lt;year&gt;\2&lt;/year&gt;" dest="8">
- <expression noclean="1">class=&quot;title_kor&quot;[ ]*&gt;([^&lt;]+)&lt;/strong&gt;[^\(]*\(.*?&lt;?([0-9]+).*?title_AKA&quot;[ ]*&gt;[ \n]*&lt;span class=&quot;eng&quot;&gt;[ \n]*&lt;/span</expression>
+ <RegExp input="$$2" output="&lt;id&gt;\1&lt;/id&gt;" dest="8">
+ <expression/>
+ </RegExp>
+ <RegExp input="$$1" output="&lt;title&gt;\1&lt;/title&gt;&lt;year&gt;\2&lt;/year&gt;" dest="8+">
+ <expression trim="1">"title_kor"\s*&gt;([^&lt;]+)&lt;/strong&gt;[^\(]*\(&lt;a [^&gt;]*&gt;(\d+)&lt;/a&gt;\).*?"title_AKA"\s*&gt;\s*&lt;span class="eng"&gt;\s*&lt;/span</expression>
</RegExp>
- <RegExp input="$$1" output="&lt;title&gt;\1 (\3)&lt;/title&gt;&lt;year&gt;\2&lt;/year&gt;" dest="8">
- <expression noclean="1">class=&quot;title_kor&quot;[ ]*&gt;([^&lt;]+)&lt;/strong&gt;[^\(]*\(.*?&lt;?([0-9]+).*?title_AKA&quot;[ ]*&gt;[ \n]*&lt;span class=&quot;eng&quot;&gt;[ \n]*([^ \n][^&lt;]*)&lt;/span</expression>
+ <RegExp input="$$1" output="&lt;title&gt;\1 (\3)&lt;/title&gt;&lt;year&gt;\2&lt;/year&gt;" dest="8+">
+ <expression trim="1,3">"title_kor"\s*&gt;([^&lt;]+)&lt;/strong&gt;[^\(]*\(&lt;a [^&gt;]*&gt;(\d+)&lt;/a&gt;\).*?"title_AKA"\s*&gt;\s*&lt;span class="eng"&gt;\s*(\S[^&lt;]*)&lt;/span</expression>
</RegExp>
- <RegExp input="$$1" output="&lt;title&gt;\1 (\3)&lt;/title&gt;&lt;year&gt;\2&lt;/year&gt;" dest="8">
- <expression noclean="1">class=&quot;title_kor&quot;[ ]*&gt;([^&lt;]+)&lt;/strong&gt;[^\(]*\(.*?&lt;?([0-9]+).*?title_AKA&quot;[ ]*&gt;[ \n]*([^ \n].*?)&lt;span class=&quot;eng&quot;&gt;[ \n]*&lt;/span</expression>
+ <RegExp input="$$1" output="&lt;title&gt;\1 (\3)&lt;/title&gt;&lt;year&gt;\2&lt;/year&gt;" dest="8+">
+ <expression trim="1,3">"title_kor"\s*&gt;([^&lt;]+)&lt;/strong&gt;[^\(]*\(&lt;a [^&gt;]*&gt;(\d+)&lt;/a&gt;\).*?"title_AKA"\s*&gt;\s*(\S[^&lt;]*)&lt;span class="eng"&gt;\s*&lt;/span</expression>
</RegExp>
<RegExp input="$$1" output="&lt;title&gt;\1 (\3,\4)&lt;/title&gt;&lt;year&gt;\2&lt;/year&gt;" dest="8+">
- <expression noclean="1">class=&quot;title_kor&quot;[ ]*&gt;([^&lt;]+)&lt;/strong&gt;[^\(]*\(.*?&lt;?([0-9]+).*?title_AKA&quot;[ ]*&gt;([^&quot;]+)&quot;eng&quot;&gt;([^&lt;]*)&lt;/span</expression>
+ <expression trim="1,3,4">"title_kor"\s*&gt;([^&lt;]+)&lt;/strong&gt;[^\(]*\(&lt;a [^&gt;]*&gt;(\d+)&lt;/a&gt;\).*?"title_AKA"\s*&gt;\s*(\S[^&lt;]*)&lt;span class="eng"&gt;\s*(\S[^&lt;]*)&lt;/span</expression>
</RegExp>
- <!-- Director's names !-->
<RegExp input="$$7" output="&lt;director&gt;\1&lt;/director&gt;" dest="8+">
<RegExp input="$$1" output="\1" dest="7">
- <expression noclean="1">&lt;strong&gt;감독&lt;/strong&gt;(.*?)&lt;strong&gt;출연&lt;/strong&gt;</expression>
+ <expression noclean="1">&lt;strong&gt;감독&lt;/strong&gt;(.*)&lt;strong&gt;출연&lt;/strong&gt;</expression>
</RegExp>
- <expression noclean="1" repeat="yes">&lt;a href=[^&gt;]*&gt;([^&lt;]+)&lt;/a&gt;</expression>
+ <expression repeat="yes">&lt;a href=[^&gt;]*&gt;([^&lt;]*)&lt;/a&gt;</expression>
</RegExp>
- <!-- Studio !-->
<RegExp input="$$7" output="&lt;studio&gt;\1&lt;/studio&gt;" dest="8+">
<RegExp input="$$1" output="\1" dest="7">
- <expression noclean="1">&lt;strong&gt;제작/배급&lt;/strong&gt;(.*?)&lt;strong&gt;감독&lt;/strong&gt;</expression>
+ <expression noclean="1">&lt;strong&gt;제작/배급&lt;/strong&gt;(.*)&lt;strong&gt;감독&lt;/strong&gt;</expression>
</RegExp>
- <expression noclean="1" repeat="yes">&lt;a href=&quot;[^&gt;]*&gt;([^&lt;]+)&lt;/a&gt;</expression>
+ <expression repeat="yes">&lt;a href="[^&gt;]*&gt;([^&lt;]*)&lt;/a&gt;</expression>
</RegExp>
- <!-- Actors !-->
- <RegExp input="$$7" output="&lt;actor&gt;&lt;name&gt;\2&lt;/name&gt;&lt;role&gt;\3&lt;/role&gt;&lt;thumb&gt;\1&lt;/thumb&gt;&lt;/actor&gt;" dest="8+">
- <RegExp input="$$1" output="\1" dest="7">
- <expression noclean="1">출연.*?&quot;actors&quot;(.*)&quot;gradeWrap&quot;</expression>
- </RegExp>
- <expression noclean="1" repeat="yes">class=&quot;cast_text&quot;&gt;&lt;img id=&quot;[^&quot;]*&quot;src=&quot;(.[^&quot;]*)&quot;[^&gt;]*&gt;[^&gt;]*&gt;(.[^&lt;]*)&lt;/a&gt;[^\(]*\((.[^\)]*)\)</expression>
- </RegExp>
- <!-- Tag Line !-->
- <RegExp input="$$7" output="&lt;tagline&gt;\1&lt;/tagline&gt;" dest="8+">
+ <RegExp conditional="!EnableMoreCast" input="$$7" output="&lt;actor&gt;&lt;name&gt;\2&lt;/name&gt;&lt;role&gt;\3&lt;/role&gt;&lt;thumb&gt;\1&lt;/thumb&gt;&lt;/actor&gt;" dest="8+">
<RegExp input="$$1" output="\1" dest="7">
- <expression noclean="1">&quot;synopsis&quot;(.*)id=&quot;review&quot;</expression>
+ <expression noclean="1">&lt;strong&gt;출연&lt;/strong&gt;.*?"actors"(.*)"gradeWrap"</expression>
</RegExp>
- <expression>class=&quot;txt&quot;&gt;[^&gt;]*&gt;(.[^&lt;]*)&lt;/b&gt;</expression>
+ <expression repeat="yes">class="cast_text"&gt;&lt;img id="[^"]*"src="([^"]*)"[^&gt;]*&gt;[^&gt;]*&gt;([^&lt;]*)&lt;/a&gt;[^\(]*\(([^\)]*)\)</expression>
</RegExp>
- <!-- Plot !-->
- <RegExp input="$$1" output="&lt;plot&gt;\1&lt;/plot&gt;" dest="8+">
- <expression>&lt;div class=&quot;txt&quot;&gt;&lt;table&gt;&lt;tr&gt;&lt;td&gt;[^&gt;]*&gt;[^&gt;]*&gt;&lt;/a&gt;[^\t]*\t(.[^&lt;]*)&lt;</expression>
+ <RegExp conditional="EnableMoreCast" input="$$1" output="&lt;url cache=&quot;daum-cast.htm&quot; function=&quot;GetDaumCast&quot;&gt;\1&lt;/url&gt;" dest="8+">
+ <expression>&lt;a class="tab3" href="([^"]*)"</expression>
</RegExp>
- <RegExp input="$$1" output="&lt;plot&gt;\1&lt;/plot&gt;" dest="8+">
- <expression>&lt;div class=&quot;txt&quot;&gt;&lt;table&gt;&lt;tr&gt;&lt;td&gt;[^\t]*\t(.[^&lt;]*)&lt;</expression>
+ <RegExp conditional="EnableWriter" input="$$1" output="&lt;url cache=&quot;daum-cast.htm&quot; function=&quot;GetDaumWriter&quot;&gt;\1&lt;/url&gt;" dest="8+">
+ <expression>&lt;a class="tab3" href="([^"]*)"</expression>
</RegExp>
- <!-- Single Poster !-->
- <RegExp conditional="!EnableThumbs" input="$$1" output="&lt;thumb&gt;\1&lt;/thumb&gt;" dest="8+">
- <expression>&lt;p class=&quot;poster&quot;&gt;&lt;a[^&lt;]*&lt;img src=&quot;([^&quot;]*)&quot;</expression>
+ <RegExp conditional="!EnableFullPlot" input="$$7" output="&lt;plot&gt;\1&lt;/plot&gt;" dest="8+">
+ <RegExp input="$$1" output="\1" dest="7">
+ <expression>&lt;div class="txt"&gt;&lt;table&gt;&lt;tr&gt;&lt;td&gt;(.*?)&lt;/td&gt;&lt;/tr&gt;&lt;/table&gt;\s*&lt;/div&gt;</expression>
+ </RegExp>
+ <expression>(\S.*\S)</expression>
</RegExp>
- <!-- Premiered !-->
- <RegExp input="$$7" output="&lt;premiered&gt;\1.\2&lt;/premiered&gt;" dest="8+">
+ <RegExp conditional="EnableFullPlot" input="$$1" output="&lt;url cache=&quot;daum-story.htm&quot; function=&quot;GetDaumFullPlot&quot;&gt;\1&lt;/url&gt;" dest="8+">
+ <expression>&lt;a class="tab2" href="([^"]*)"</expression>
+ </RegExp>
+ <RegExp input="$$7" output="&lt;premiered&gt;\1&lt;/premiered&gt;" dest="8+">
<RegExp input="$$1" output="\1" dest="7">
<expression noclean="1">&lt;strong&gt;요약정보&lt;/strong&gt;(.*)&lt;strong&gt;감독&lt;/strong&gt;</expression>
</RegExp>
- <expression noclean="1">([0-9]*)-([0-9]*)-[0-9]*</expression>
+ <expression>(\d{4}-\d{1,2}-\d{1,2})</expression>
</RegExp>
- <!-- Genres !-->
<RegExp input="$$7" output="&lt;genre&gt;\1&lt;/genre&gt;" dest="8+">
<RegExp input="$$1" output="\1" dest="7">
- <expression noclean="1">&lt;strong&gt;요약정보&lt;/strong&gt;(.*)&lt;strong&gt;감독&lt;/strong&gt;</expression>
+ <expression noclean="1">&lt;strong&gt;요약정보&lt;/strong&gt;(.*?)\|</expression>
</RegExp>
- <expression noclean="1" repeat="yes">&lt;a href=&apos;[^&apos;]*&apos;&gt;(.[^&lt;]*)&lt;/a&gt;</expression>
+ <expression repeat="yes">&lt;a href='[^']*'&gt;([^&lt;]*)&lt;/a&gt;</expression>
</RegExp>
- <!-- Runtime !-->
<RegExp input="$$7" output="&lt;runtime&gt;\1 분&lt;/runtime&gt;" dest="8+">
<RegExp input="$$1" output="\1" dest="7">
<expression noclean="1">&lt;strong&gt;요약정보&lt;/strong&gt;(.*)&lt;strong&gt;감독&lt;/strong&gt;</expression>
</RegExp>
- <expression noclean="1">([0-9]*) 분</expression>
+ <expression>(\d+) 분</expression>
</RegExp>
- <!-- User rating !-->
<RegExp input="$$1" output="&lt;rating&gt;\1&lt;/rating&gt;" dest="8+">
- <expression>&quot;star_big pink&quot;&gt;&lt;[^&gt;]*&gt;&lt;[^&gt;]*&gt;[^&gt;]*&gt;[^&gt;]*&gt;&lt;em&gt;(.[^&lt;]*)&lt;/em&gt;</expression>
+ <expression>"star_big [^"]+"&gt;.*?&gt;네티즌별점&lt;.*?&lt;em&gt;([\d\.]+)&lt;/em&gt;</expression>
</RegExp>
- <!-- User vote !-->
<RegExp input="$$1" output="&lt;votes&gt;\1&lt;/votes&gt;" dest="8+">
- <expression>&lt;span class=&quot;fs11 fc3&quot;&gt;\(([0-9]*)명 참여\)&lt;/span&gt;</expression>
+ <expression>&lt;span class="fs11 fc3"&gt;\((\d+)명 참여\)&lt;/span&gt;</expression>
</RegExp>
- <!-- MPAA !-->
<RegExp input="$$1" output="&lt;mpaa&gt;\1&lt;/mpaa&gt;" dest="8+">
- <expression noclean="1">icon_rating_[0-9]*\.gif&quot; title=&quot;(.[^&quot;]*)&quot;</expression>
+ <expression>icon_rating_\d+\.gif" title="([^"]*)"</expression>
</RegExp>
- <!-- Poster !-->
- <RegExp conditional="EnableThumbs" input="$$1" output="&lt;url function=&quot;GetDaumPoster&quot;&gt;\1&lt;/url&gt;" dest="8+">
- <expression>&lt;p class=&quot;poster&quot;&gt;&lt;a href=&quot;(.[^&quot;]*)&quot;</expression>
+ <RegExp conditional="!EnableThumbs" input="$$1" output="&lt;thumb&gt;\1&lt;/thumb&gt;" dest="8+">
+ <expression>&lt;p class="poster"&gt;&lt;a [^&lt;]*&lt;img src="([^"]*)"</expression>
</RegExp>
- <!-- FanArts !-->
- <RegExp conditional="EnableFanArt" input="$$5" output="&lt;url function=&quot;GetTMDBId&quot;&gt;http://api.themoviedb.org/2.0/Movie.search?title=\1&amp;amp;api_key=57983e31fb435df4df77afb854740ea9&lt;/url&gt;" dest="8+">
- <RegExp input="$$1" output="\1" dest="7">
- <expression noclean="1" clear="yes">&quot;title_AKA&quot;[^&quot;]*&quot;eng&quot;&gt;([^&lt;]*)&lt;</expression>
+ <RegExp conditional="EnableThumbs" input="$$1" output="&lt;url cache=&quot;daum-photo.htm&quot; function=&quot;GetDaumPoster&quot;&gt;\1&lt;/url&gt;" dest="8+">
+ <expression>&lt;a class="tab4" href="([^"]*)"</expression>
+ </RegExp>
+ <RegExp conditional="EnableTMDBFanart" input="$$7" output="&lt;url function=&quot;GetTMDBFanartById&quot;&gt;http://www.google.com/search?q=site:imdb.com\1&lt;/url&gt;" dest="8+">
+ <RegExp input="$$1" output="\4 \2" dest="6">
+ <expression trim="1,3,4">"title_kor"\s*&gt;([^&lt;]*)&lt;/strong&gt;[^\(]*\(&lt;a [^&gt;]*&gt;(\d+)&lt;/a&gt;\).*?"title_AKA"\s*&gt;([^&lt;]*)&lt;span class="eng"&gt;([^&lt;]+)&lt;/span</expression>
</RegExp>
- <RegExp input="$$7" output="+\1" dest="5">
- <expression repeat="yes" clear="yes">([^ ,]+)</expression>
+ <RegExp input="$$6" output="+\1" dest="7">
+ <expression repeat="yes">([^ ,]+)</expression>
</RegExp>
- <expression noclean="1"/>
- </RegExp>
- <!-- FanArts !-->
- <RegExp conditional="EnableFanArt" input="$$1" output="&lt;url function=&quot;GetDAUMFanart&quot;&gt;\1&lt;/url&gt;" dest="8+">
- <expression noclean="1">&lt;a class=&quot;tab1&quot; href=&quot;(.[^&quot;]*)&quot;</expression>
+ <expression/>
</RegExp>
- <!-- Writer !-->
- <RegExp conditional="EnableWriter" input="$$1" output="&lt;url function=&quot;GetWriter&quot;&gt;\1&lt;/url&gt;" dest="8+">
- <expression noclean="1">&lt;a class=&quot;tab3&quot; href=&quot;(.[^&quot;]*)&quot;</expression>
+ <RegExp conditional="EnableFanart" input="$$1" output="&lt;url cache=&quot;daum-photo.htm&quot; function=&quot;GetDaumFanart&quot;&gt;\1&amp;order=recommend&lt;/url&gt;" dest="8+">
+ <expression>&lt;a class="tab4" href="([^"]*)"</expression>
</RegExp>
<expression noclean="1"/>
</RegExp>
</GetDetails>
- <GetDaumPoster dest="8">
- <RegExp input="$$9" output="&lt;details&gt;\1&lt;/details&gt;" dest="8">
- <RegExp input="$$1" output="\1" dest="7">
- <expression noclean="1">id=&quot;photo&quot;(.*)id=&quot;photoViewer&quot;</expression>
+ <GetDaumPoster dest="4">
+ <RegExp input="$$5" output="&lt;details&gt;\1&lt;/details&gt;" dest="4">
+ <RegExp input="$$1" output="&lt;url cache=&quot;daum-photo.htm&quot; function=&quot;GetDaumPosterPage&quot;&gt;\1&lt;/url&gt;" dest="5">
+ <expression noclean="1">&lt;a class="tab4" href="([^"]*)"</expression>
</RegExp>
- <RegExp input="$$7" output="&lt;thumb preview=&quot;http://\1/C155x225/\2&quot;&gt;http://\1/image/\2&lt;/thumb&gt;" dest="9">
- <expression noclean="1" repeat="yes">&lt;img src=&quot;http://([^/]+)/C155x225/([^&quot;]+)&quot;</expression>
+ <RegExp input="$$1" output="\1" dest="6">
+ <expression noclean="1">&lt;h3 class="tit"&gt;포스터 리스트&lt;/h3&gt;.*&lt;span class="current"&gt;(.*)&lt;!-- 포스터 End</expression>
+ </RegExp>
+ <RegExp input="$$6" output="&lt;url function=&quot;GetDaumPosterPage&quot;&gt;http://movie.daum.net/\1&lt;/url&gt;" dest="5+">
+ <expression repeat="yes" noclean="1">&lt;a href="([^"]*)"&gt;(\d+)&lt;/a&gt;</expression>
</RegExp>
<expression noclean="1"/>
</RegExp>
</GetDaumPoster>
- <GetTMDBId dest="8">
- <RegExp input="$$1" output="&lt;details&gt;&lt;url function=&quot;GetTMDBFanart&quot;&gt;http://api.themoviedb.org/2.0/Movie.getInfo?id=\1&amp;amp;api_key=57983e31fb435df4df77afb854740ea9&lt;/url&gt;&lt;/details&gt;" dest="8">
- <expression>&lt;id&gt;([0-9]*)&lt;/id&gt;</expression>
+ <GetDaumPosterPage dest="4">
+ <RegExp input="$$5" output="&lt;details&gt;\1&lt;/details&gt;" dest="4">
+ <RegExp input="$$1" output="\1" dest="6">
+ <expression noclean="1">&lt;h3 class="tit"&gt;포스터 리스트&lt;/h3&gt;(.*)&lt;!-- 포스터 End</expression>
+ </RegExp>
+ <RegExp input="$$6" output="&lt;thumb preview=&quot;http://\1/C155x225/\2&quot;&gt;http://\1/image/\2&lt;/thumb&gt;" dest="5">
+ <expression repeat="yes" noclean="1">&lt;img src="http://([^/]+)/C155x225/([^"]+)"</expression>
+ </RegExp>
+ <expression noclean="1"/>
+ </RegExp>
+ </GetDaumPosterPage>
+ <GetDaumFanart clearbuffers="no" dest="4">
+ <RegExp input="$$5" output="&lt;details&gt;\1&lt;/details&gt;" dest="4">
+ <RegExp input="$$1" output="\1" dest="6">
+ <expression noclean="1">&lt;h3 class="tit"&gt;포토 리스트&lt;/h3&gt;.*&lt;span class="current"&gt;(.*)&lt;!-- 포토 End</expression>
+ </RegExp>
+ <RegExp input="$$6" output="&lt;url function=&quot;GetDaumFanartPage2&quot;&gt;http://movie.daum.net\1&lt;/url&gt;" dest="5">
+ <expression repeat="yes" noclean="1">&lt;a href="([^"]*)"&gt;(\d+)&lt;/a&gt;</expression>
+ </RegExp>
+ <RegExp input="$$1" output="&lt;url cache=&quot;daum-photo.htm&quot; function=&quot;GetDaumFanartPage&quot;&gt;\1&lt;/url&gt;" dest="5+">
+ <expression noclean="1">&lt;a class="tab4" href="([^"]*)"</expression>
+ </RegExp>
+ <expression noclean="1"/>
</RegExp>
- </GetTMDBId>
- <GetTMDBFanart clearbuffers="no" dest="8">
- <RegExp input="$$1" output="&lt;thumb preview=&quot;http://themoviedb.org/image/backdrops/\1/\2_poster.\3&quot;&gt;http://themoviedb.org/image/backdrops/\1/\2.\3&lt;/thumb&gt;" dest="10+">
- <expression clear="yes" noclean="1" repeat="yes">&lt;backdrop size=&quot;original&quot;&gt;.*?/([0-9]+)/([^\.]+)\.([a-zA-Z]*)[^&lt;]*&lt;/backdrop&gt;</expression>
+ </GetDaumFanart>
+ <GetDaumFanartPage dest="4">
+ <!-- append result from GetTMDBFanart/GetDaumFanartPage2 calls -->
+ <RegExp input="$$11$$5$$10" output="&lt;details&gt;&lt;fanart&gt;\1&lt;/fanart&gt;&lt;/details&gt;" dest="4">
+ <RegExp input="$$1" output="\1" dest="6">
+ <expression noclean="1">&lt;h3 class="tit"&gt;포토 리스트&lt;/h3&gt;(.*)&lt;!-- 포토 End</expression>
+ </RegExp>
+ <RegExp input="$$6" output="&lt;thumb preview=&quot;http://\1/S150x125/\2&quot;&gt;http://\1/image/\2&lt;/thumb&gt;" dest="5">
+ <expression repeat="yes" noclean="1">&lt;img src="http://([^/]+)/S150x125/([^"]+)"</expression>
+ </RegExp>
+ <expression noclean="1"/>
</RegExp>
- </GetTMDBFanart>
- <GetDAUMFanart clearbuffers="no" dest="8">
- <RegExp input="$$10$$11" output="&lt;details&gt;&lt;fanart&gt;\1&lt;/fanart&gt;&lt;/details&gt;" dest="8">
- <RegExp input="$$1" output="&lt;thumb preview=&quot;http://\1/C93x70/\2&quot;&gt;http://\1/image/\2&lt;/thumb&gt;" dest="11">
- <expression clear="yes" noclean="1" repeat="yes">imgArray\.push\(&quot;http://(.[^/]*)/C93x70/(.[^&quot;]*)&quot;\)</expression>
+ </GetDaumFanartPage>
+ <GetDaumFanartPage2 clearbuffers="no" dest="4">
+ <RegExp input="$$5" output="&lt;details&gt;&lt;/details&gt;" dest="4">
+ <RegExp input="$$1" output="\1" dest="6">
+ <expression noclean="1">&lt;h3 class="tit"&gt;포토 리스트&lt;/h3&gt;(.*)&lt;!-- 포토 End</expression>
+ </RegExp>
+ <RegExp input="$$6" output="&lt;thumb preview=&quot;http://\1/S150x125/\2&quot;&gt;http://\1/image/\2&lt;/thumb&gt;" dest="10+">
+ <expression repeat="yes" noclean="1">&lt;img src="http://([^/]+)/S150x125/([^"]+)"</expression>
+ </RegExp>
+ <expression noclean="1"/>
+ </RegExp>
+ </GetDaumFanartPage2>
+ <GetDaumCast dest="4">
+ <RegExp input="$$6" output="&lt;details&gt;\1&lt;/details&gt;" dest="4">
+ <RegExp input="$$1" output="\1" dest="5">
+ <expression noclean="1">&lt;h5\s*&gt;\s*주연&lt;/h5&gt;(.*)&lt;h5\s*&gt;\s*제작</expression>
+ </RegExp>
+ <RegExp input="$$5" output="&lt;actor&gt;&lt;name&gt;\2&lt;/name&gt;&lt;role&gt;\3&lt;/role&gt;&lt;thumb&gt;\1&lt;/thumb&gt;&lt;/actor&gt;" dest="6">
+ <expression clear="yes" repeat="yes">&lt;img src="([^"]*)"[^&gt;]*&gt;&lt;/a&gt;&lt;/span&gt;\s*&lt;span class="name"[^&gt;]*&gt;&lt;a href[^&gt;]*&gt;([^&lt;]*)&lt;/a&gt;&lt;/span&gt;\s*&lt;span class="etcs"&gt;\s*([^&lt;]*) 역\s*&lt;/span&gt;</expression>
+ </RegExp>
+ <RegExp input="$$5" output="&lt;actor&gt;&lt;name&gt;\2&lt;/name&gt;&lt;role&gt;\4&lt;/role&gt;&lt;thumb&gt;\1&lt;/thumb&gt;&lt;/actor&gt;" dest="6+">
+ <expression repeat="yes">&lt;img src="([^"]*)"[^&gt;]*&gt;&lt;/a&gt;&lt;/span&gt;\s*&lt;span class="name"[^&gt;]*&gt;&lt;a href[^&gt;]*&gt;([^&lt;]*)&lt;/a&gt;&lt;/span&gt;\s*&lt;span class="aka fc4"&gt;&lt;span class="eng"&gt;\(([^\)]*)\)&lt;/span&gt;&lt;/span&gt;\s*&lt;span class="etcs"&gt;\s*([^&lt;]*) 역\s*&lt;/span&gt;</expression>
+ </RegExp>
+ <expression noclean="1"/>
+ </RegExp>
+ </GetDaumCast>
+ <GetDaumWriter dest="4">
+ <RegExp input="$$6" output="&lt;details&gt;\1&lt;/details&gt;" dest="4">
+ <RegExp input="$$1" output="\1" dest="5">
+ <expression noclean="1">&lt;h5\s*&gt;\s*각본&lt;/h5&gt;(.*?)&lt;h5\s*&gt;</expression>
+ </RegExp>
+ <RegExp input="$$5" output="&lt;credits&gt;\1&lt;/credits&gt;" dest="6">
+ <expression repeat="yes">class="em b"&gt;([^&lt;]*)&lt;</expression>
+ </RegExp>
+ <expression noclean="1"/>
+ </RegExp>
+ </GetDaumWriter>
+ <GetDaumFullPlot dest="4">
+ <RegExp input="$$5" output="&lt;details&gt;\1&lt;/details&gt;" dest="4">
+ <RegExp input="$$6" output="&lt;plot&gt;\1&lt;/plot&gt;" dest="5">
+ <RegExp input="$$1" output="\1" dest="6">
+ <expression>&lt;a name="synopsis"&gt;.*?&lt;div class="txt"&gt;(.*?)&lt;/div&gt;</expression>
+ </RegExp>
+ <expression>(\S.*\S)</expression>
+ </RegExp>
+ <RegExp input="$$6" output="&lt;tagline&gt;\1&lt;/tagline&gt;" dest="5+">
+ <RegExp input="$$1" output="\1" dest="6">
+ <expression clear="yes">&lt;a name="tagline"&gt;.*?&lt;div class="txt"&gt;(.*?)&lt;/div&gt;</expression>
+ </RegExp>
+ <expression>(\S.*\S)</expression>
</RegExp>
<expression noclean="1"/>
</RegExp>
- </GetDAUMFanart>
- <GetWriter dest="8">
- <RegExp input="$$1" output="\1" dest="7">
- <expression noclean="1">&lt;h5[ ]*&gt;[ ]*각본&lt;/h5&gt;(.*)&lt;h5[ ]*&gt;촬영&lt;/h5&gt;</expression>
+ </GetDaumFullPlot>
+ <!-- Modify after copying from common/tmdb.xml -->
+ <GetTMDBFanartById dest="4">
+ <RegExp input="$$5" output="&lt;details&gt;\1&lt;/details&gt;" dest="4">
+ <RegExp input="$$1" output="&lt;url function=&quot;GetTMDBFanart&quot; cache=&quot;tmdb-images-tt\1.xml&quot;&gt;http://api.themoviedb.org/2.1/Movie.getImages/en/xml/57983e31fb435df4df77afb854740ea9/tt\1&lt;/url&gt;" dest="5">
+ <expression>/tt([0-9]+)</expression>
+ </RegExp>
+ <RegExp input="$$1" output="&lt;url function=&quot;GetTMDBFanart&quot; cache=&quot;tmdb-images-\1.xml&quot;&gt;http://api.themoviedb.org/2.1/Movie.getImages/en/xml/57983e31fb435df4df77afb854740ea9/\1&lt;/url&gt;" dest="5">
+ <expression>&lt;id&gt;([0-9]*)&lt;/id&gt;</expression>
+ </RegExp>
+ <expression noclean="1"/>
</RegExp>
- <RegExp input="$$6" output="&lt;details&gt;\1&lt;/details&gt;" dest="8">
- <RegExp input="$$7" output="&lt;credits&gt;\1&lt;/credits&gt;" dest="6">
- <expression noclean="1" repeat="yes">class=&quot;em b&quot;&gt;(.[^&lt;]*)&lt;</expression>
+ </GetTMDBFanartById>
+ <GetTMDBFanart clearbuffers="no" dest="5">
+ <RegExp input="$$2" output="&lt;details&gt;\1&lt;/details&gt;" dest="5">
+ <RegExp input="$$1" output="&lt;thumb&gt;\1&lt;/thumb&gt;" dest="3">
+ <expression noclean="1" repeat="yes">&lt;backdrop[^&gt;]*&gt;(.*?)&lt;/backdrop&gt;</expression>
+ </RegExp>
+ <RegExp input="$$3" output="&lt;thumb preview=&quot;\2&quot;&gt;\1\3&lt;/thumb&gt;" dest="4">
+ <expression noclean="1,3" repeat="yes">&lt;thumb&gt;(.*?)url=&quot;([^&quot;]*)&quot; size=&quot;poster&quot;(.*?)&lt;/thumb&gt;</expression>
+ </RegExp>
+ <RegExp input="$$4" output="&lt;thumb\1&gt;\2&lt;/thumb&gt;" dest="11">
+ <expression repeat="yes" noclean="1">&lt;thumb([^&gt;]*)&gt;.*?url=&quot;([^&quot;]*)&quot; size=&quot;original&quot;.*?&lt;/thumb&gt;</expression>
</RegExp>
<expression noclean="1"/>
</RegExp>
- </GetWriter>
+ </GetTMDBFanart>
</scraper>