Class: WebPageParser::BbcNewsPageParserV4

Inherits:
BbcNewsPageParserV3 show all
Defined in:
lib/web-page-parser/parsers/bbc_news_page_parser.rb

Constant Summary collapse

CONTENT_RE = 
ORegexp.new('<div class=.story-body.>(.*?)<!-- / story\-body', 'm')
STRIP_PAGE_BOOKMARKS = 
ORegexp.new('<div id="page-bookmark-links-head".+?</div>', 'm')
STRIP_STORY_DATE = 
ORegexp.new('<span class="date".+?</span>', 'm')
STRIP_STORY_LASTUPDATED = 
ORegexp.new('<span class="time\-text".+?</span>', 'm')
STRIP_STORY_TIME = 
ORegexp.new('<span class="time".+?</span>', 'm')
TITLE_RE = 
ORegexp.new('<h1 class="story\-header">(.+?)</h1>', 'm')
STRIP_CAPTIONS_RE2 = 
ORegexp.new('<div class=.caption.+?</div>','m')
STRIP_HIDDEN_A = 
ORegexp.new('<a class=.hidden.+?</a>','m')
STRIP_STORY_FEATURE = 
ORegexp.new('<div class=.story\-feature.+?</div>', 'm')
STRIP_HYPERPUFF_RE = 
ORegexp.new('<div class=.embedded-hyper.+?<div class=.hyperpuff.+?</div>.+?</div>', 'm')
STRIP_MARKETDATA_RE = 
ORegexp.new('<div class=.market\-data.+?</div>', 'm')
STRIP_EMBEDDEDHYPER_RE = 
ORegexp.new('<div class=.embedded\-hyper.+?</div>', 'm')

Constants inherited from BbcNewsPageParserV3

WebPageParser::BbcNewsPageParserV3::ICONV, WebPageParser::BbcNewsPageParserV3::STRIP_FEATURES_RE, WebPageParser::BbcNewsPageParserV3::STRIP_MARKET_DATA_WIDGET_RE

Constants inherited from BbcNewsPageParserV2

WebPageParser::BbcNewsPageParserV2::DATE_RE, WebPageParser::BbcNewsPageParserV2::PARA_RE, WebPageParser::BbcNewsPageParserV2::STRIP_BLOCKS_RE, WebPageParser::BbcNewsPageParserV2::STRIP_CAPTIONS_RE, WebPageParser::BbcNewsPageParserV2::STRIP_COMMENTS_RE, WebPageParser::BbcNewsPageParserV2::STRIP_TAGS_RE, WebPageParser::BbcNewsPageParserV2::WHITESPACE_RE

Constants inherited from BaseParser

WebPageParser::BaseParser::DATE_RE, WebPageParser::BaseParser::HTML_ENTITIES_DECODER, WebPageParser::BaseParser::ICONV, WebPageParser::BaseParser::KILL_CHARS_RE

Instance Attribute Summary

Attributes inherited from BaseParser

#guid, #page, #url

Instance Method Summary collapse

Methods inherited from BaseParser

#content, #date, #decode_entities, #hash, #initialize, #title

Constructor Details

This class inherits a constructor from WebPageParser::BaseParser

Instance Method Details

#content_processor ⇒ Object 



122
123
124
125
126
127
128
129
130
131
132
133
134
135
 
# File 'lib/web-page-parser/parsers/bbc_news_page_parser.rb', line 122

def content_processor
  @content = STRIP_PAGE_BOOKMARKS.gsub(@content, '')
  @content = STRIP_STORY_DATE.gsub(@content, '')
  @content = STRIP_STORY_LASTUPDATED.gsub(@content, '')
  @content = STRIP_STORY_TIME.gsub(@content, '')
  @content = TITLE_RE.gsub(@content, '')
  @content = STRIP_CAPTIONS_RE2.gsub(@content, '')
  @content = STRIP_HIDDEN_A.gsub(@content, '')
  @content = STRIP_STORY_FEATURE.gsub(@content, '')
  @content = STRIP_HYPERPUFF_RE.gsub(@content, '')
  @content = STRIP_MARKETDATA_RE.gsub(@content, '')
  @content = STRIP_EMBEDDEDHYPER_RE.gsub(@content, '')
  super
end