forum_post 0.1 → 0.2

data/README

@@ -1,5 +1,15 @@
-== Forum_post GemPlugin
+== Forum_post_v0.2/ GemPlugin
 
-You should document your project here.
+the gem used to get the forum post
 
+install:
+gem install forum_post
+example:
+require 'rubygems'
+require 'forum_post'
+require 'open-uri'
+uri='http://topic.csdn.net/u/20110718/17/f1a523fb-8c65-4510-a094-daf7bd2698cf.html?50656'
+source=open(uri).read
+doc=ForumPost::Document.new(source)
+puts doc.content
 

data/Rakefile

@@ -6,7 +6,7 @@ require 'rake/rdoctask'
 require 'fileutils'
 include FileUtils
 
-version="0.1"
+version="0.2"
 name="forum_post"
 
 spec = Gem::Specification.new do |s|

data/lib/forum_post.rb

@@ -9,7 +9,7 @@ module ForumPost
 
     REGEXES = {
         :unlikelyCandidatesRe => /combx|community|disqus|extra|foot|header|menu|remark|rss|shoutbox|sidebar|sponsor|ad-break|agegate|pagination|pager|popup/i,
-        :maybe_post => /article|body|column|main|content|post|topic|text|info|message/i,
+        :maybe_post => /article|body|column|main|content|post|topic|text|info|message|item/i,
         :not_post => /author|head|avatar|profile|rank|user|uid/i,
         :divToPElementsRe => /<(a|blockquote|dl|div|img|ol|p|pre|table|ul)/i,
         :replaceBrsRe => /(<br[^>]*>[ \n\r\t]*){2,}/i,
@@ -32,21 +32,23 @@ module ForumPost
       remove_script_and_style!#去除scrpt和style
       remove_unlikely_candidates!#去除不可能是发帖内容的标签
       transform_misused_divs_into_p!#把一些噪音的div转换成p标签
-      if likely_posts.size > 0
-        #回帖数大于5
-        most_likely_post = most_likely_posts(likely_posts)
-        candidates=score_elem(most_likely_post)
-      else
-        #回帖数小于5
-        candidates=score_elem(@html)
-      end
-        best_elem=select_best(candidates)
-        best_elem.text
+			bests=[]
+			better_post = likely_posts
+			if better_post.size > 0
+			  debug("post>4")
+     		@html =  most_likely_posts(better_post)
+			else
+			  debug("post<4")
+		  end	
+	    bests=@html.css("div,tr,td") 
+			return @html.text.gsub(/\s(\s+)/,"")	if bests.size==0  
+     	candidates=score_elem(bests)
+      best_elem=select_best(candidates)
+      best_elem.text
     end
 
-    def score_elem(most_likely_post)
-      candidates=most_likely_post.css("div,tr,td") 
-      candidates.each do |elem|
+    def score_elem(bests)
+      bests.each do |elem|
         base_score=100
         str = "#{elem[:class]}#{elem[:id]}"
         base_score+=10 if str =~ REGEXES[:maybe_post]
@@ -54,7 +56,7 @@ module ForumPost
         base_score-=8  if elem_size(elem)<DEFAULT_OPTIONS[:min_text_length]
         elem["score"]=base_score.to_s
       end
-      candidates
+      bests
     end
 
     def select_best(candidates)
@@ -79,11 +81,11 @@ module ForumPost
       last_candidates.sort{|a,b| b["text_rate"].to_f<=>a["text_rate"].to_f}.first
     end
 
-    def most_likely_posts(likely_posts)
+    def most_likely_posts(better_post)
       most_likely_posts=[]
-      likely_posts.each do |q|
+      better_post.each do |q|
           flag=0
-          likely_posts.each do |p|
+          better_post.each do |p|
                if is_contain(p,q) == true && p != q
                   flag+=1
                end
@@ -93,6 +95,7 @@ module ForumPost
             debug("most_likelys:#{q.name}.#{q[:class]}.#{q[:id]}") 
           end
         end
+			debug("most_likely_posts:#{most_likely_posts.size}")	
       most_likely_posts.sort{|m,n| elem_size(n)<=>elem_size(m)}.first
     end
     
@@ -128,7 +131,7 @@ module ForumPost
       @html.css("*").each do |elem|
         str = "#{elem[:class]}#{elem[:id]}"
         if str =~ REGEXES[:unlikelyCandidatesRe] && str !~ REGEXES[:maybe_post] && elem.name.downcase != 'body'
-          debug("Removing unlikely candidate - #{str}")
+					debug("Removing unlikely candidate - #{str}")
           elem.remove
         end
       end
@@ -138,7 +141,7 @@ module ForumPost
       @html.css("*").each do |elem|
         if elem.name.downcase == "div"
           if elem.inner_html !~ REGEXES[:divToPElementsRe]
-            debug("Altering div(##{elem[:id]}.#{elem[:class]}) to p");
+						#debug("Altering div(##{elem[:id]}.#{elem[:class]}) to p");
             elem.name = "p"
           end
         end

data/lib/{forum_post → forum_post_v0.2}/init.rb

@@ -1,6 +1,6 @@
 require 'gem_plugin'
 
-# give this class the name you want for your command forum_post
+# give this class the name you want for your command forum_post_v0.2/
 class ChangeME < GemPlugin::Plugin "/somecategory"
 end
 

metadata

@@ -1,12 +1,12 @@
 --- !ruby/object:Gem::Specification 
 name: forum_post
 version: !ruby/object:Gem::Version 
-  hash: 9
+  hash: 15
   prerelease: false
   segments: 
   - 0
-  - 1
-  version: "0.1"
+  - 2
+  version: "0.2"
 platform: ruby
 authors: 
 - ltl3884
@@ -46,7 +46,7 @@ files:
 - README
 - Rakefile
 - lib/forum_post.rb
-- lib/forum_post/init.rb
+- lib/forum_post_v0.2/init.rb
 - resources/defaults.yaml
 has_rdoc: true
 homepage: