Re: XPath

Michael Schuerig catalogs books, and asked for help:

----------8<----------

$:.unshift 'rexml/1.2.6/'
require 'rexml/document'
include REXML

OLD = <<DELIMITER
<biblio>
<entry type="Book">
<author>Thomas, David; Hunt, Andrew</author>
<language>english</language>
<publisher>Addison-Wesley</publisher>
<title>Programming Ruby. The Pragmatic Programmer's Guide</title>
<year>2000</year>
</entry>
<entry type="Book">
<author>Blammo, Blah</author>
<language>english</language>
<publisher>Hubbabubba</publisher>
<title>Foozboozer's Life</title>
<type>Book</type>
<year>2002</year>
</entry>
</biblio>
DELIMITER

CURRENT = <<DELIMITER
<biblio>
<entry type="Book">
<language>english</language>
<publisher>Addison-Wesley</publisher>
<title>Programming Ruby. The Pragmatic Programmer's Guide</title>
<type>Book</type>
<year>2000</year>
</entry>
<entry type="Book">
<author>Blammo, Blah</author>
<language>english</language>
<publisher>Hubbabubba</publisher>
<title>Foozboozer's Life</title>
<type>Book</type>
<year>2002</year>
</entry>
</biblio>
DELIMITER

#MS = you, Mr. [e-mail]

#MS: Read in both versions of the file
OLD_TREE  = Document.new OLD
goal_tree = Document.new CURRENT.dup

#puts OLD_TREE,goal_tree

XPath.each goal_tree, '/biblio/entry' do |entry|
#MS: For each "dubious" entry in the current file
   unless XPath.first(entry, '//author')
     title =  XPath.first(entry, '//title').text
     year  =  XPath.first(entry, '//year').text
#MS: Find a corresponding (based on title and year) entry
#MS:  in the original file
     XPath.each(OLD_TREE,"//entry") do |e|
       title_test = XPath.first(e, '//title').text == title
       year_test  = XPath.first(e, '//year').text == year
       if title_test and year_test
         @author = XPath.first(e, '//author')
       end
     end
     #/author .text
     #XPath.first(needed_entry author
     #puts needed_entry
#MS: Fix the current entry
     entry.insert_before XPath.first(entry, '//language'), @author
     #puts title
   end
end

print goal_tree.to_s.gsub(/>\s*</, ">\n<")

#MS: Dump a fixed file
=begin
<biblio>
<entry type='Book'>
<author>Thomas, David; Hunt, Andrew</author>
<language>english</language>
<publisher>Addison-Wesley</publisher>
<title>Programming Ruby. The Pragmatic Programmer&apos;s Guide</title>
<type>Book</type>
<year>2000</year>
</entry>
<entry type='Book'>
<author>Blammo, Blah</author>
<language>english</language>
<publisher>Hubbabubba</publisher>
<title>Foozboozer&apos;s Life</title>
<type>Book</type>
<year>2002</year>
</entry>
</biblio>
=end

# This is a pretty quirky way to do it.
# Part of the problem is my limited time,
#  part of it is that REXML's current XPath support is limited.
# 1. You and/or others can optimize the above, and make it more elegant.
# 2. If memory becomes a problem think about streaming APIs.
# 3. You might want to think about ISBNs.
# 4. You might want to wait for REXML 2.0
# 5. The thing should @ least give you a start ;)

----------8<----------


Tobi


-- 

* peace&love.
* http://www.pinkjuice.com/