code translation extractor added

tasks/extract_code_translations.rb

@@ -0,0 +1,34 @@
+require 'nokogiri'
+
+# html = open('http://www.biblegateway.com/passage/?search=Mateo1-2&version=NVI')
+# doc = Nokogiri::HTML(html.read)
+
+link = 'language_table.html'
+
+# doc = Nokogiri::HTML(open(link))
+
+doc = Nokogiri::HTML(open(link).read, nil, 'utf-8')
+# doc.encoding = 'utf-8'
+
+codes = doc.css('tr td:first').map(&:content)
+
+puts codes
+
+hash = {}
+current_code = ''
+
+doc.css('tr').each do |tr|
+	td = tr.css('td').first
+	puts td
+	if td
+		code = td.content 
+		hash[code] = []
+
+		tr.css('td').each do |td|
+			label = td.content
+			hash[code] << label if hash[code] && label != code
+		end
+	end
+end
+
+puts hash

tasks/language_extract.html

@@ -0,0 +1,73 @@
+<table class="wikitable sortable">
+<tr>
+<th>ISO 639-1</th>
+<th>bg</th>
+<th>cs</th>
+<th>da</th>
+<th>de</th>
+<th>el</th>
+<th>en</th>
+<th>es</th>
+<th>et</th>
+<th>fi</th>
+</tr>
+<tr>
+<td>ab</td>
+<td>абхазки</td>
+<td>abchazština</td>
+<td>Abkhazian<sup class="Template-Fact" style="white-space:nowrap;">[<i><a href="/wiki/Wikipedia:Citation_needed" title="Wikipedia:Citation needed"><span title="This claim needs references to reliable sources from December 2010">citation needed</span></a></i>]</sup></td>
+<td>Abchasisch</td>
+<td>αμπχαζικά</td>
+<td><a href="/wiki/Abkhazian_language" title="Abkhazian language" class="mw-redirect">Abkhazian</a></td>
+<td>abjazio</td>
+<td>abhaasi</td>
+<td>abhaasi</td>
+</tr>
+<tr>
+<td>af</td>
+<td>африканс</td>
+<td>afrikánština</td>
+<td>Afrikaans</td>
+<td>Afrikaans</td>
+<td>αφρικάνς</td>
+<td><a href="/wiki/Afrikaans_language" title="Afrikaans language" class="mw-redirect">Afrikaans</a></td>
+<td>afrikaans</td>
+<td>afrikaani</td>
+<td>afrikaans</td>
+</tr>
+<tr>
+<td>an</td>
+<td>арагонски</td>
+<td>aragonština</td>
+<td>Aragonesisk</td>
+<td>Aragonesisch</td>
+<td>γλώσσα της Aragon</td>
+<td><a href="/wiki/Aragonese_language" title="Aragonese language">Aragonese</a></td>
+<td>aragonés</td>
+<td>aragoni</td>
+<td>aragonia</td>
+</tr>
+<tr>
+<td>ar</td>
+<td>арабски</td>
+<td>arabština</td>
+<td>Arabisk</td>
+<td>Arabisch</td>
+<td>αραβικά</td>
+<td><a href="/wiki/Arabic_language" title="Arabic language">Arabic</a></td>
+<td>árabe</td>
+<td>araabia</td>
+<td>arabia</td>
+</tr>
+<tr>
+<td>as</td>
+<td>асамски</td>
+<td>ásámština</td>
+<td>Assamesisk</td>
+<td>Assamesisch</td>
+<td>ασαμέζικα</td>
+<td><a href="/wiki/Assamese_language" title="Assamese language">Assamese</a></td>
+<td>asamés</td>
+<td>assami</td>
+<td>assami</td>
+</tr>