wp2txt 0.5.1 → 0.5.02

data/README.md

@@ -30,20 +30,25 @@ Command line options are as follows:
 
     Usage: wp2txt [options]
     where [options] are:
-                         --input-file, -i:   Wikipedia dump file with .bz2 (compressed) or .txt (uncompressed) format
-                     --output-dir, -o <s>:   Output directory (default: Present working directory)
-                        --convert-off, -c:   Output XML (without converting to plain text)
-                           --list-off, -l:   Exclude list items from output
-                        --heading-off, -d:   Exclude section titles from output
-                          --title-off, -t:   Exclude page titles from output
-          --table-off, --no-table-off, -a:   Exclude page titles from output (default: true)
-    --template-off, --no-template-off, -e:   Remove template notations from output (default: true)
-                       --redirect-off, -r:   Not show redirect destination
-                       --strip-marker, -s:   Remove symbols prefixed to list items, definitions, etc.
-                       --category-off, -g:   Not show article category information
-                      --file-size, -f <i>:   Approximate size (in MB) of each output file (default: 10)
-                            --version, -v:   Print version and exit
-                               --help, -h:   Show this message
+                          --input-file, -i:   Wikipedia dump file with .bz2 (compressed) or .txt
+                                              (uncompressed) format
+                      --output-dir, -o <s>:   Output directory (default:
+                                              /Users/yohasebe/Dropbox/code/wp2txt)
+                         --convert-off, -c:   Output XML (without converting to plain text)
+                            --list-off, -l:   Exclude list items from output
+                         --heading-off, -d:   Exclude section titles from output
+                           --title-off, -t:   Exclude page titles from output
+           --table-off, --no-table-off, -a:   Exclude page titles from output (default: true)
+     --template-off, --no-template-off, -e:   Remove template notations from output (default:
+                                              true)
+                        --redirect-off, -r:   Not show redirect destination
+                        --strip-marker, -s:   Remove symbols prefixed to list items, definitions,
+                                              etc.
+                        --category-off, -g:   Not show article category information
+                       --file-size, -f <i>:   Approximate size (in MB) of each output file
+                                              (default: 10)
+                             --version, -v:   Print version and exit
+                                --help, -h:   Show this message
 
 ### Limitations ###
 

data/lib/wp2txt/article.rb

@@ -70,9 +70,8 @@ module Wp2txt
     @@blank_line_regex = Regexp.new('^\s*$')
 
     @@redirect_regex = Regexp.new('#(?:REDIRECT|転送)\s+\[\[(.+)\]\]', Regexp::IGNORECASE)
-
-    category_patterns = ["Category", "Categoria"].join("|")
-    @@category_regex = Regexp.new('[\{\[\|\b](?:' + category_patterns + ')\:(.*?)[\}\]\|\b]', Regexp::IGNORECASE)
+    
+    @@category_regex = Regexp.new('[\{\[\|\b](?:C|c)ategory\:(.*?)[\}\]\|\b]')
 
     def initialize(text, title = "", strip_tmarker = false)
       @title = title.strip

data/lib/wp2txt/version.rb

@@ -1,3 +1,3 @@
 module Wp2txt
-  VERSION = "0.5.1"
+  VERSION = "0.5.02"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: wp2txt
 version: !ruby/object:Gem::Version
-  version: 0.5.1
+  version: 0.5.02
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-01-16 00:00:00.000000000 Z
+date: 2013-01-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec