merge peterc and fix tests

lib/pismo/internal_attributes.rb

@@ -3,7 +3,7 @@ module Pismo
   module InternalAttributes
     @@phrasie = Phrasie::Extractor.new
 
-    MONTHS_REGEX = %r{(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec|January|February|March|April|May|June|July|August|September|October|November|December)}i
+    MONTHS_REGEX = %r{(Jan|Feb|Mar|Apr|May|Jun|Jul|Aug|Sep|Oct|Nov|Dec|January|February|March|April|May|June|July|August|September|October|November|December)\.?}i
     DATETIME_REGEXEN = [
       /#{MONTHS_REGEX}\b\s+\d+\D{1,10}\d{4}/i,
       /(on\s+)?\d+\s+#{MONTHS_REGEX}\s+\D{0,10}\d+/i,
@@ -168,13 +168,12 @@ def datetime
       DATETIME_REGEXEN.detect {|r| datetime = @doc.to_html[r] }
 
       return unless datetime and datetime.length > 4
-
       # Clean up the string for use by Chronic
       datetime.strip!
       datetime.gsub!(/(Monday|Tuesday|Wednesday|Thursday|Friday|Saturday|Sunday|mon|tues|tue|weds|wed|thurs|thur|thu|fri|sat|sun)[^\w]*/i, '')
-      datetime.sub!(/(on\s+|\,)/, '')
+      datetime.sub!(/(on\s+|\,|\.)/, '')
       datetime.sub!(/(\d+)(th|st|rd)/, '\1')
-      Chronic.parse(datetime) || datetime
+      Chronic.parse(datetime, :context => :past) || datetime
     end
 
     # Returns the author of the page/content
@@ -264,6 +263,36 @@ def videos(limit = 1)
       reader_doc && !reader_doc.videos.empty? ? reader_doc.videos(limit) : nil
     end
 
+ # Returns the tags or categories of the page/content
+    def tags
+      css_selectors = [
+                       '.watch-info-tag-list a',  # YouTube
+                       '.entry .tags a',          # Livejournal
+                       'a[rel~=tag]',             # Wordpress and many others
+                       'a.tag',                   # Tumblr
+                       '.tags a',
+                       '.labels a',
+                       '.categories a',
+                       '.topics a'
+                      ]
+
+      tags = []
+
+      # grab the first one we get results from
+      css_selectors.each do |css_selector|
+        tags += @doc.css(css_selector)
+        break if tags.any?
+      end
+
+      # convert from Nokogiri Element objects to strings
+      tags.map!(&:inner_text)
+
+      # remove "#" from hashtag-like tags
+      tags.map! { |t| t.gsub(/^#/, '') }
+
+      tags
+    end
+
     # Returns the "keyword phrases" in the document (not the meta keywords - they're next to useless now)
     DEFAULT_KEYWORD_OPTIONS = { :limit => 20, :minimum_score => "1%" }
     def keywords(options = {})

test/corpus/metadata_expected.yaml

@@ -5,8 +5,9 @@
   :lede: I'm just aching to know if the new Apple tablet (insert caveats, weasel words and qualifiers here) is a potential Cintiq competitor. I don't think it will be, but you never know. It may also have a built in barometer and bird call generator. I'm never sure if Apple does themselves more good than harm with the secrecy and anticipation that surrounds the run-up to these announcements.
   :feeds:
     - http://www.readwriteweb.com/rss.xml
-    - http://www.readwriteweb.com/archives/2010/01/cartoon_apple_tablet_now_with_barometer_and_bird_c.xml
-:briancray:
+    - http://www.readwriteweb.com/archives/2010/01/cartoon_apple_tablet_now_with_barometer_and_bird_c.xml 
+  :tags: [apple tablet, steve jobs] 
+:briancray: 
   :title: 5 great examples of popular blog posts that you should know
   :feed: http://feeds.feedburner.com/briancray/blog
   :lede: "This is a mock post."
@@ -54,6 +55,7 @@
   :author: Peter Cooper
   :lede: CoffeeScript (GitHub repo) is a new programming language with a pure Ruby compiler.
   :feed: http://www.rubyinside.com/feed/
+  :tags: [Cool]
 :zefrank:
   :sentences: If there's anyone who knows how to marshal an online audience, it's Ze Frank. Ze is best-known for his 2006 program "The Show," in which he made a new 2-3 minute video every day for 1 year. Topics ranged from "fingers in food" to the mysteries of airport signage to a tour de force summary of creatives' addiction to un-executed ideas, aka brain crack.
   :title: "Ze Frank on Imaginary Audiences :: Articles :: The 99 Percent"
@@ -69,4 +71,8 @@
   :title: 18 Incredible CSS3 Effects You Have Never Seen Before
   :lede: "CSS3 is hot these days and will soon be available in most modern browser. Just recently, I started to become aware to the present of CSS3 around the web. I can see some of the websites such as twitter and designer portfolios websites are using it. Also, I have started to implement it to my own project as well and I really love it!"
   :sentences: CSS3 is hot these days and will soon be available in most modern browser. Just recently, I started to become aware to the present of CSS3 around the web. I can see some of the websites such as twitter and designer portfolios websites are using it.
-  :datetime: 2010-02-17 12:00:00 -07:00
+  :datetime: 2010-02-17 12:00:00 +00:00
+:thegoodbookblog:
+  :title: Signs Of Life
+  :datetime: 2012-07-25 12:00:00 +00:00
+  :tags: [Church Life, Marriage and Family, Ministry and Leadership]