The branch, eden-pre has been updated
via abbe853f4a6d0c55786ecb7754ab646fb7ed7773 (commit)
from b2698e79fa49cdb077cf6bc6a9cd4e18b1bf5461 (commit)
- Log -----------------------------------------------------------------
http://xbmc.git.sourceforge.net/git/gitweb.cgi?p=xbmc/scrapers;a=commit;h=abbe853f4a6d0c55786ecb7754ab646fb7ed7773
commit abbe853f4a6d0c55786ecb7754ab646fb7ed7773
Author: olympia <[email protected]>
Date: Sat May 21 16:02:26 2011 +0200
[metadata.kinopoisk.ru] updated to version 1.0.33
diff --git a/metadata.kinopoisk.ru/addon.xml b/metadata.kinopoisk.ru/addon.xml
index 3b5829b..ff0bd83 100644
--- a/metadata.kinopoisk.ru/addon.xml
+++ b/metadata.kinopoisk.ru/addon.xml
@@ -1,8 +1,7 @@
<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<addon id="metadata.kinopoisk.ru"
name="KinoPoisk"
- version="1.0.28"
- provider-name="hamp + akara">
+ version="1.0.33">
<requires>
<import addon="xbmc.metadata" version="1.0"/>
<import addon="metadata.common.themoviedb.org" version="1.0.3"/>
diff --git a/metadata.kinopoisk.ru/changelog.Russian.txt
b/metadata.kinopoisk.ru/changelog.Russian.txt
index 8506a7e..af08fc8 100644
--- a/metadata.kinopoisk.ru/changelog.Russian.txt
+++ b/metadata.kinopoisk.ru/changelog.Russian.txt
@@ -1,3 +1,13 @@
+1.0.33:
+- Ðобавил заголовок Accept-Language
+1.0.32:
+- Ðзменил User-agent и поиÑковÑй ÑÑл
+1.0.31:
+- ÐÑпÑавил ÑÑейлеÑÑ
+1.0.30:
+- ÐÑпÑÐ°Ð²Ð»ÐµÐ½Ð¸Ñ [email protected]
+1.0.29:
+- ÐÑпÑÐ°Ð²Ð»ÐµÐ½Ð¸Ñ [email protected]
1.0.28:
- ÐÑпÑавлено полÑÑение поÑÑеÑов длÑ
ÑилÑмов Ñ Ð¾Ð´Ð½Ð¸Ð¼ поÑÑеÑом
1.0.27:
diff --git a/metadata.kinopoisk.ru/changelog.txt
b/metadata.kinopoisk.ru/changelog.txt
index af58280..d3f99f6 100644
--- a/metadata.kinopoisk.ru/changelog.txt
+++ b/metadata.kinopoisk.ru/changelog.txt
@@ -1,3 +1,13 @@
+1.0.33:
+- Added Accept-Language header
+1.0.32:
+- Changed User-agent and search url
+1.0.31:
+- Fixed trailers
+1.0.30:
+- Fixes by [email protected]
+1.0.29:
+- Fixes by [email protected]
1.0.28:
- Fixed poster scraping for movies with single poster
1.0.27:
@@ -18,7 +28,7 @@
- Fixed actors page parsings
1.0.19:
- Added option for scraping only actors with translated names
-1.0.18:
+1.0.18:
- New search web layout
- Removed "3D" from movie title
1.0.17:
diff --git a/metadata.kinopoisk.ru/kinopoisk.xml
b/metadata.kinopoisk.ru/kinopoisk.xml
index 38f52a3..1e9b3d7 100644
--- a/metadata.kinopoisk.ru/kinopoisk.xml
+++ b/metadata.kinopoisk.ru/kinopoisk.xml
@@ -1,13 +1,13 @@
<?xml version="1.0" encoding="WINDOWS-1251"?>
<scraper date="2010-02-23" framework="1.1">
<NfoUrl dest="3">
- <RegExp input="$$1"
output="<url>\1|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20i686%3B%20en-US%3B%20rv%3A1.8.1.14)%20Gecko%2F20080418%20Ubuntu%2F7.10%20(gutsy)%20Firefox%2F2.0.0.14</url>"
dest="3">
+ <RegExp input="$$1"
output="<url>\1|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20x86_64%3B%20rv%3A2.0.1)%20Gecko%2F20100101%20Firefox%2F4.0.1</url>"
dest="3">
<expression
noclean="1">(http://www\.kinopoisk\.ru/level/1/film/[0-9]*/)</expression>
</RegExp>
</NfoUrl>
<!-- Ôîðìèðóåì ïîèñêîâûé URl -->
<CreateSearchUrl SearchStringEncoding="CP1251" dest="3">
- <RegExp input="$$1"
output="<url>http://www.kinopoisk.ru/index.php?level=7&from=forma&result=adv&m_act%5Bfrom%5D=forma&m_act%5Bwhat%5D=content&m_act%5Bfind%5D=\1&m_act%5Byear%5D=$$2|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20i686%3B%20en-US%3B%20rv%3A1.8.1.14)%20Gecko%2F20080418%20Ubuntu%2F7.10%20(gutsy)%20Firefox%2F2.0.0.14</url>"
dest="3">
+ <RegExp input="$$1"
output="<url>http://s.kinopoisk.ru/index.php?level=7&from=forma&result=adv&m_act%5Bfrom%5D=forma&m_act%5Bwhat%5D=content&m_act%5Bfind%5D=\1&m_act%5Byear%5D=$$2|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20x86_64%3B%20rv%3A2.0.1)%20Gecko%2F20100101%20Firefox%2F4.0.1&Accept-Language=ru-ru,ru%3Bq=0.8,en-us%3Bq=0.5,en%3Bq=0.3</url>"
dest="3">
<expression noclean="1"/>
</RegExp>
</CreateSearchUrl>
@@ -34,16 +34,16 @@
<RegExp input="$$1"
output="<year>\1</year>" dest="7+">
<expression><a
href="/level/10/m_act%5Byear%5D/(\d+)/"</expression>
</RegExp>
- <RegExp input="$$1"
output="<url>http://www.kinopoisk.ru/level/1/film/\1|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20i686%3B%20en-US%3B%20rv%3A1.8.1.14)%20Gecko%2F20080418%20Ubuntu%2F7.10%20(gutsy)%20Firefox%2F2.0.0.14</url>"
dest="7+">
+ <RegExp input="$$1"
output="<url>http://www.kinopoisk.ru/level/1/film/\1/sr/1/|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20x86_64%3B%20rv%3A2.0.1)%20Gecko%2F20100101%20Firefox%2F4.0.1&Accept-Language=ru-ru,ru%3Bq=0.8,en-us%3Bq=0.5,en%3Bq=0.3</url>"
dest="7+">
<expression><a
href="/level/19/film/(\d+)/"></expression>
</RegExp>
<expression noclean="1" />
</RegExp>
<!--Êîä åñëè íàéäåíî íåñêîëüêî âàðèàíòîâ-->
- <RegExp conditional="!search_originaltitle" input="$$1"
output="<entity><title>\2</title><year>\3</year><url>http://www.kinopoisk.ru/level/1/film/\1|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20i686%3B%20en-US%3B%20rv%3A1.8.1.14)%20Gecko%2F20080418%20Ubuntu%2F7.10%20(gutsy)%20Firefox%2F2.0.0.14</url></entity>"
dest="5+">
+ <RegExp conditional="!search_originaltitle" input="$$1"
output="<entity><title>\2</title><year>\3</year><url>http://www.kinopoisk.ru/level/1/film/\1/sr/1/|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20x86_64%3B%20rv%3A2.0.1)%20Gecko%2F20100101%20Firefox%2F4.0.1&Accept-Language=ru-ru,ru%3Bq=0.8,en-us%3Bq=0.5,en%3Bq=0.3</url></entity>"
dest="5+">
<expression repeat="yes"><p
class="name"><a
href="http://www.kinopoisk.ru/level/1/film/(\d+)/sr/1/">(.+?)</a>[^>]*?class="year">(\d+)<</expression>
</RegExp>
- <RegExp conditional="search_originaltitle" input="$$1"
output="<entity><title>\2
[\4]</title><year>\3</year><url>http://www.kinopoisk.ru/level/1/film/\1|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20i686%3B%20en-US%3B%20rv%3A1.8.1.14)%20Gecko%2F20080418%20Ubuntu%2F7.10%20(gutsy)%20Firefox%2F2.0.0.14</url></entity>"
dest="5+">
+ <RegExp conditional="search_originaltitle" input="$$1"
output="<entity><title>\2
[\4]</title><year>\3</year><url>http://www.kinopoisk.ru/level/1/film/\1/sr/1/|User-Agent=Mozilla%2F5.0%20(X11%3B%20U%3B%20Linux%20x86_64%3B%20rv%3A2.0.1)%20Gecko%2F20100101%20Firefox%2F4.0.1&Accept-Language=ru-ru,ru%3Bq=0.8,en-us%3Bq=0.5,en%3Bq=0.3</url></entity>"
dest="5+">
<expression repeat="yes"><p
class="name"><a
href="http://www.kinopoisk.ru/level/1/film/(\d+)/sr/1/">(.+?)</a>[^>]*?class="year">(\d+)<.*?class="gray">(?:(?:\d+
ìèí)|)(.*?)(?:(?:, \d+ ìèí)|)<</expression>
</RegExp>
<!-- fix html entities from titles-->
@@ -121,7 +121,7 @@
</RegExp>
<!--COUNTRY-->
<RegExp input="$$1"
output="<country>\1</country>" dest="5+">
- <expression repeat="yes"><a
href="/level/10/m_act%5Bcountry%5D/.+?/"><img
src="/images/flags/flag.*?alt="(.*?)"</expression>
+ <expression repeat="yes"><a
href="/level/10/m_act%5Bcountry%5D/.+?/"><img
src="http://st.kinopoisk.ru/images/flags/flag.*?alt=";(.*?)"</expression>
</RegExp>
<!--YEAR-->
<RegExp input="$$1" output="\1" dest="13">
@@ -132,7 +132,7 @@
</RegExp>
<!-- premier year -->
<RegExp input="$$1" output="\1" dest="14">
-
<expression>href="/level/80/film/\d+/">.+?([0-9]+)</a></expression>
+
<expression>href="/level/80/film/\d+/".+?([0-9]+)</a></expression>
</RegExp>
<!-- êîëè÷åñòâî ãîëîñîâ íà imdb id -->
<RegExp input="$$1" output="\1 " dest="15">
@@ -147,11 +147,11 @@
</RegExp>
<!--TOP250-->
<RegExp input="$$1"
output="<top250>\1</top250>" dest="5+">
- <expression>Top250:
<a[^>]+>([0-9]*)</a></expression>
+ <expression><a href="/level/56/film/\d+/"
style="color:#555">([0-9]*).+?</a></expression>
</RegExp>
<!--Ðåéòèíã MPAA-->
<RegExp input="$$1"
output="<mpaa>\1</mpaa>" dest="5+">
- <expression><tr><td
class="type">[^<]+MPAA</td><td[^>]+><a
href='.+?/rn/(.+?)/'</expression>
+ <expression><tr><td
class="type">[^<]+MPAA</td><td[^>]+><a
href='.+?/rn/(.+?)/'</expression>
</RegExp>
<!--TAGLINE-->
<RegExp input="$$1"
output="<tagline>\1</tagline>" dest="5+">
@@ -163,7 +163,7 @@
</RegExp>
<!--Ðåéòèíã IMDB-->
<RegExp conditional="!kinopoisk_rating" input="$$1"
output="<rating>\1</rating><votes>\2</votes>" dest="5+">
- <expression>IMDB: ([0-9.]+)
\(([0-9\s]+)\)</div></expression>
+ <expression>IMDb: ([0-9.]+)
\(([0-9\s]+)\)</div></expression>
</RegExp>
<RegExp conditional="kinopoisk_rating" input="none"
output="" dest="2">
<expression />
@@ -225,7 +225,7 @@
<expression>href="/level/19/film/(\d+)/"</expression>
</RegExp>
<!--POSTERS-->
- <RegExp input="$$1"
conditional="posters_from_kinopoisk" output="<url
function="GMP">http://www.kinopoisk.ru\1</url>"; dest="5+">
+ <RegExp conditional="posters_from_kinopoisk"
input="$$1" output="<url
function="GMP">http://www.kinopoisk.ru\1</url>"; dest="5+">
<expression>href="(/level/17/film/([0-9]+)/)"</expression>
</RegExp>
<!--TRAILER-->
@@ -239,14 +239,14 @@
<!-- Åñëè ôèëüì ïðîèçâåä¸í â Ðîññèè, òî
âûáèðàåì ïåðâûé òðåéëåð -->
<RegExp input="$$9" output="$$4" dest="2">
<RegExp input="$$1" output="<url
function="GMT_Russia">http://www.kinopoisk.ru/level/16/film/\1</url>";
dest="4">
- <expression><a
href="/level/16/film/(\d+)/"></expression>
+ <expression><a
href="/level/16/film/(\d+)/</expression>
</RegExp>
<expression>^true$</expression>
</RegExp>
<!-- Åñëè ôèëüì ïðîèçâåä¸í íå â Ðîññèè, òî
âûáèðàåì ðóññêèé òðåéëåð -->
<RegExp input="$$9" output="$$3" dest="2">
<RegExp input="$$1" output="<url
function="GMT">http://www.kinopoisk.ru/level/16/film/\1</url>";
dest="3">
- <expression><a
href="/level/16/film/(\d+)/"></expression>
+ <expression><a
href="/level/16/film/(\d+)/</expression>
</RegExp>
<expression>^<country></expression>
</RegExp>
@@ -259,7 +259,7 @@
</GetDetails>
<GetPoster dest="4">
<RegExp input="$$1" output="<details><thumb
preview="http://st.kinopoisk.ru/images/poster/sm_im_\2.jpg">http://st.kinopoisk.ru/\1</thumb></details>";
dest="4">
- <expression repeat="no" noclean="1"><img style="border: 1px
#777 solid; position: absolute; margin-left: -5000px"
src="http://st.kinopoisk.ru/([^"]+?-(\d+).jpg)"</expression>
+ <expression repeat="no" noclean="1"><img style="border: 1px
#777 solid; position: absolute; margin-left: -5000px" id='image'
src="http://st.kinopoisk.ru/([^"]+?-(\d+).jpg)"</expression>
</RegExp>
</GetPoster>
<GMP dest="4">
@@ -274,7 +274,7 @@
<expression noclean="1"><div
class="navigator">(.*?)</div></expression>
</RegExp>
<RegExp input="$$7" output="<url
function="GMP_SINGLE">http://www.kinopoisk.ru\1</url>";
dest="8+">
- <expression repeat="yes" noclean="1"><li
><a
href="(/level/17/film/\d+/page/\d+/)">\d+</a></li></expression>
+ <expression repeat="yes" noclean="1"><li
><a
href="(/level/17/film/\d+/page/\d+/)">\d+</a></li></expression>
</RegExp>
<expression noclean="1"/>
</RegExp>
@@ -301,7 +301,7 @@
</RegExp>
<!-- HD TRAILERS-->
<RegExp conditional="download_hd" input="$$3"
output="<trailer type="HD"
quality="\2">\1</trailer>" dest="5+">
- <expression repeat="yes"><td
width=20><img src="/images/icon-hd2.gif".*?<a
href="(?:.*?)&type=trailer&link=(.*?)"
class="continue">(.*?)</a></expression>
+ <expression repeat="yes"><td
width=20><img src="http://st.kinopoisk.ru/images/icon-hd2.gif".*?<a
href="(?:.*?)&type=trailer&link=(.*?)"
class="continue">(.*?)</a></expression>
</RegExp>
<!-- Ïûòàåìñÿ âûáðàòü ïðåäïî÷èòàåìîå êà÷åñòâî-->
<RegExp input="$INFO[TrailerQ]" output="$$6" dest="2">
@@ -341,7 +341,7 @@
<GMT dest="4">
<RegExp input="$$2" output="<details>\1</details>"
dest="4">
<RegExp input="$$1" output="<trailer
tilte="\1"><html>\2</html></trailer>" dest="3">
- <expression repeat="yes"
noclean="2">src="/images/flags/flag-2.gif".*?class="all">(.*?)</a>(.*?)<!--
/ðîëèê</expression>
+ <expression repeat="yes"
noclean="2">src="http://st.kinopoisk.ru/images/flags/flag-2.gif".*?class="all"[^>]*>(.*?)</a>(.*?)<!--
/ðîëèê</expression>
</RegExp>
<RegExp input="$$3" output="\1" dest="3">
<expression
noclean="1">tilte="Òðåéëåð.*?"(.*?)</trailer></expression>
@@ -352,7 +352,7 @@
</RegExp>
<!-- HD TRAILERS-->
<RegExp conditional="download_hd" input="$$3"
output="<trailer type="HD"
quality="\2">\1</trailer>" dest="5+">
- <expression repeat="yes"><td
width=20><img src="/images/icon-hd2.gif".*?<a
href="(?:.*?)&type=trailer&link=(.*?)"
class="continue">(.*?)</a></expression>
+ <expression repeat="yes"><td
width=20><img src="http://st.kinopoisk.ru/images/icon-hd2.gif".*?<a
href="(?:.*?)&type=trailer&link=(.*?)"
class="continue">(.*?)</a></expression>
</RegExp>
<!-- Ïûòàåìñÿ âûáðàòü ïðåäïî÷èòàåìîå êà÷åñòâî-->
<RegExp input="$INFO[TrailerQ]" output="$$6" dest="2">
@@ -400,7 +400,7 @@
</RegExp>
<!-- HD TRAILERS-->
<RegExp conditional="download_hd" input="$$3"
output="<trailer type="HD"
quality="\2">\1</trailer>" dest="5+">
- <expression repeat="yes"><td
width=20><img src="/images/icon-hd2.gif".*?<a
href="(?:.*?)&type=trailer&link=(.*?)"
class="continue">(.*?)</a></expression>
+ <expression repeat="yes"><td
width=20><img src="http://st.kinopoisk.ru/images/icon-hd2.gif".*?<a
href="(?:.*?)&type=trailer&link=(.*?)"
class="continue">(.*?)</a></expression>
</RegExp>
<!-- Ïûòàåìñÿ âûáðàòü ïðåäïî÷èòàåìîå êà÷åñòâî-->
<RegExp conditional="orig_trailer"
input="$INFO[TrailerQ]" output="$$6" dest="2+">
@@ -462,15 +462,15 @@
<expression clear="yes"
noclean="1"><tr><td colspan=3><a
name="actor"></td></tr>.+?</table>(.*?)<table</expression>
</RegExp>
<RegExp conditional="!rus_actor_names" input="$$9"
output="<n>\2<r>\3<t>http://st.kinopoisk.ru/images/actor/\1</e>";
dest="9">
- <expression repeat="yes" trim="3"><img
src="http://st.kinopoisk.ru/images/(?:sm_actor/([0-9]+.jpg)|no-poster.gif)"
alt=.+?<p><a
href="/level/4/people/\d+/">(.*?)</a><b>.*?</b>...
([^<,]+)</expression>
+ <expression repeat="yes"
trim="3">title="/images/(?:sm_actor/([0-9]+.jpg)|no-poster.gif)" alt=.+?<a
href="/level/4/people/\d+/">(.*?)</a>.+?"role">...
([^<]+)</expression>
</RegExp>
<RegExp conditional="rus_actor_names" input="$$9"
output="<n>\2<r>\3<t>http://st.kinopoisk.ru/images/actor/\1</e>";
dest="9">
- <expression repeat="yes" trim="3"><img
src="http://st.kinopoisk.ru/images/(?:sm_actor/([0-9]+.jpg)|no-poster.gif)"
alt=.+?<p><a
href="/level/4/people/\d+/">([à-ÿÀ-ß][^<]*)</a><b>.*?</b>...
([^<,]+)</expression>
+ <expression repeat="yes"
trim="3">title="/images/(?:sm_actor/([0-9]+.jpg)|no-poster.gif)" alt=.+?<a
href="/level/4/people/\d+/">([à-ÿÀ-ß][^<]*)</a>.*?"role">...
([^<]+)</expression>
</RegExp>
<RegExp input="$$9"
output="<actor><name>\1</name><role>\2</role><thumb>\3</thumb></actor>"
dest="7">
<expression repeat="yes"
trim="3"><n>([^<]+)<r>(?:èãðàåò
|)([^<]+)<t>(?:(http://st.kinopoisk.ru/images/actor/[0-9]+.jpg)</e>|http://st.kinopoisk.ru/images/actor/</e>)</expression>
</RegExp>
- <!-- remove from actor roles -->
+ <!-- óáèðàåì èç ðîëåé àêò¸ðîâ -->
<RegExp input="$$7" output="\1&nbsp;" dest="7">
<expression noclean="1"/>
</RegExp>
@@ -482,13 +482,13 @@
<expression clear="yes"
noclean="1"><tr><td colspan=3><a
name="director"></td></tr>.+?</table>(.*?)<table</expression>
</RegExp>
<RegExp input="$$8"
output="<director>\1</director>" dest="7+">
- <expression repeat="yes" trim="3"><img
src="[^"]+" alt=.+?<p><a
href="/level/4/people/\d+/">([^<]+)</a></expression>
+ <expression repeat="yes"
trim="3">"name"><a
href="/level/4/people/\d+/">([^<]+)</a></expression>
</RegExp>
<RegExp input="$$1" output="\1" dest="8">
<expression clear="yes"
noclean="1"><tr><td colspan=3><a
name="writer"></td></tr>(.*?)<table</expression>
</RegExp>
<RegExp input="$$8"
output="<credits>\1</credits>" dest="7+">
- <expression repeat="yes" trim="3"><img
src="[^"]+" alt=.+?<p><a
href="/level/4/people/\d+/">([^<]+)</a></expression>
+ <expression repeat="yes"
trim="3">"name"><a
href="/level/4/people/\d+/">([^<]+)</a></expression>
</RegExp>
<expression noclean="1"/>
</RegExp>
-----------------------------------------------------------------------
Summary of changes:
metadata.kinopoisk.ru/addon.xml | 3 +-
metadata.kinopoisk.ru/changelog.Russian.txt | 10 +++++
metadata.kinopoisk.ru/changelog.txt | 12 ++++++-
metadata.kinopoisk.ru/kinopoisk.xml | 48 +++++++++++++-------------
4 files changed, 46 insertions(+), 27 deletions(-)
hooks/post-receive
--
Scrapers
------------------------------------------------------------------------------
What Every C/C++ and Fortran developer Should Know!
Read this article and learn how Intel has extended the reach of its
next-generation tools to help Windows* and Linux* C/C++ and Fortran
developers boost performance applications - including clusters.
http://p.sf.net/sfu/intel-dev2devmay
_______________________________________________
Xbmc-addons mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/xbmc-addons
![http://st.kinopoisk.ru/images/poster/sm_im_\2.jpg">http://st.kinopoisk.ru/\1</thumb></details>"](http://st.kinopoisk.ru/images/poster/sm_im_\2.jpg">http://st.kinopoisk.ru/\1</thumb></details>"){kind=link}
![http://st.kinopoisk.ru/images/icon-hd2.gif".*?<a](http://st.kinopoisk.ru/images/icon-hd2.gif".*?<a){kind=link}
![http://st.kinopoisk.ru/images/flags/flag-2.gif".*?class="all"[^>]*>](http://st.kinopoisk.ru/images/flags/flag-2.gif".*?class="all"[^>]*>){kind=link}
![http://st.kinopoisk.ru/images/actor/[0-9]+.jpg](http://st.kinopoisk.ru/images/actor/[0-9]+.jpg){kind=link}