63 lines
2.7 KiB
INI
Executable File
63 lines
2.7 KiB
INI
Executable File
**------------------------------------------------------------------------------------------------
|
|
* @header_start
|
|
* WebGrab+Plus ini for grabbing EPG data from TvGuide websites
|
|
* @Site: dw.com
|
|
* @MinSWversion: 1.1.1/53.16
|
|
* @Revision 2 - [01/10/2017] Netuddki
|
|
* - Fixed grabbing only a few hours a day (some code by Blackbear199)
|
|
* @Revision 1 - [31/05/2017] Netuddki
|
|
* - Complete rewrite for new site
|
|
* @Revision 0 - [05/05/2014] Jan van Straaten
|
|
* - creation
|
|
* @Remarks:
|
|
* @header_end
|
|
**------------------------------------------------------------------------------------------------
|
|
|
|
site {url=dw.com|timezone=UTC|maxdays=7|cultureinfo=en-GB|charset=UTF-8|titlematchfactor=90|firstshow=now}
|
|
|
|
url_index{url|http://www.dw.com/epg/data/##CHANNEL##/|subpage|}
|
|
url_index.headers {customheader=Accept-Encoding=gzip,deflate}
|
|
urldate.format {datestring|yyyy-MM-dd}
|
|
*
|
|
urldate.format {datenumber|JAVA|0}
|
|
subpage.format{list(format=D0 step=7200000 count=12)|'global_temp_1'}
|
|
*
|
|
scope.range{(urlindex)|end}
|
|
global_temp_1.modify {calculate(format=utcdate,java)|'urldate'}
|
|
end_scope
|
|
*
|
|
index_temp_1.modify {substring(type=regex)|'config_site_id' ".*?_(.*)"}
|
|
url_index.modify {replace|##CHANNEL##|'index_temp_1'}
|
|
index_variable_element.modify {substring(type=regex)|'config_site_id' "(\d+)_.*"}
|
|
*
|
|
index_showsplit.scrub {multi(includeblock="data-channel-id="'index_variable_element'"")|<div class="channelElements clearfix "|||}
|
|
*
|
|
scope.range{(splitindex)|end}
|
|
index_showsplit.modify {substring(type=regex)|"<div class=\"epgProgram\".*?<\/ul>"}
|
|
index_temp_1.modify {substring(type=regex)|'index_showsplit' "data-program-id=\"(\d+)\""}
|
|
index_temp_1.modify {cleanup(removeduplicates=equal,100 link="index_showsplit")}
|
|
end_scope
|
|
*
|
|
index_start.scrub {regex||data-begin-time="(.*?)"||}
|
|
index_stop.scrub {regex||data-end-time="(.*?)"||}
|
|
*
|
|
index_title.scrub {regex||<h2 class="title">(.*?)</a>||}
|
|
index_title.modify {cleanup(tags="<"">")}
|
|
*
|
|
index_subtitle.modify {substring(type=regex)|'index_title' "-\s(.*)"}
|
|
index_title.modify {remove|- 'index_subtitle'}
|
|
index_description.scrub {regex||<ul class="topics">(.*)||}
|
|
index_description.modify {cleanup(tags="<"">"}
|
|
index_description.modify {replace|\n| # }
|
|
index_description.modify {cleanup}
|
|
*
|
|
*
|
|
** _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
|
|
** ##### CHANNEL FILE CREATION (only to create the xxx-channel.xml file)
|
|
**
|
|
** @auto_xml_channel_start
|
|
*index_site_channel.modify {addstart|DW Deutsch\|DW Deutsch Plus\|DW English\|DW Espanol\|DW Arabic}
|
|
*index_site_id.scrub {|}
|
|
*index_site_id.modify {addstart|5_4765/1\|4_4765/1\|1_4757/2/\|3_6309/28\|2_9051/8}
|
|
** @auto_xml_channel_end
|