zetaben-Html2Feedbooks 1.0.5 → 1.0.6

data/bin/html2fb.rb

@@ -14,6 +14,7 @@ include HTML2FB
 options = {}
 options[:conf] = "conf.yaml"
 options[:preview] = true
+options[:conv] = true
 OptionParser.new do |opts|
 	opts.banner = "Usage: html2fb [options] URL"
 
@@ -23,6 +24,9 @@ OptionParser.new do |opts|
 	opts.on("-s", "-s","Send to feedbooks") do |f|
 		options[:preview] = !f
 	end
+	opts.on("-nc", "--no-conv","No charset conversion") do |f|
+		options[:conv] = !f
+	end
 end.parse!
 
 valid=false
@@ -40,7 +44,7 @@ while !valid
 	print "URL : " if entry.nil? || entry==''
 	entry=STDIN.readline.strip unless valid
 end
-conf=Conf.new(options[:conf])
+conf=Conf.new(options[:conf],options[:conv])
 content=Downloader.download(url)
 #puts content.size
 doc=Parser.new(conf).parse(content)

data/lib/conf.rb

@@ -2,13 +2,14 @@ require 'yaml'
 
 module HTML2FB
 	class Conf
-		def initialize(file)
+		def initialize(file,conv)
 			['','./',"#{File.dirname(__FILE__)}/","#{File.dirname(__FILE__)}/../confs/"].each do |p|
 				f=p+file
 				begin
 					if File.readable?(f) && File.exists?(f)
 						@conf=File.open(f,'r'){|txt| YAML::load(txt)}
 						puts "loaded config file : "+f
+						@conf['conv']=conv
 						return 
 					end
 				rescue Exception => e 

data/lib/parser.rb

@@ -15,6 +15,25 @@ module HTML2FB
 		def parse(txt)
 			puts "Parsing HTML"
 			pdoc=Hpricot(txt)
+			if @conf['conv']
+				mc=pdoc/'meta[@http-equiv="Content-Type"]'
+				if mc.size>0
+					charset=mc.first.attributes['content'].split(';').find do |s|
+						s.strip[0,7]=='charset'
+					end
+					unless charset.nil?
+						tc=charset.split('=').last.strip
+					end
+
+					unless tc.nil? 
+						puts "Trying to convert source encoding from #{tc} to utf-8"
+						require 'iconv'
+						pdoc=Hpricot(Iconv.conv('utf-8',tc.downcase,txt))
+
+					end
+
+				end
+			end
 			doc=Document.new
 			puts "Removing garbage elements"
 			remove_objs(pdoc)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: zetaben-Html2Feedbooks
 version: !ruby/object:Gem::Version 
-  version: 1.0.5
+  version: 1.0.6
 platform: ruby
 authors: 
 - Benoit Larroque