From: Jamis Buck <jgb3@...>
Date: 2004-09-23T00:26:46+09:00
Subject: Re: mozilla bookmarks

--------------070106000601070906010503
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Dick Davies wrote:
> long shot but what the hell - don't suppose any of you good
> good people are sitting on a parser for Mozilla/Firefox bookmarks.html
> files, by any chance?

Funny you should ask. :) I've had this for awhile, and I can't even 
remember why I wrote it. It's pretty hacked together, and it's not a 
true "parser" (I just search for certain patterns in the bookmark file) 
and it is hardcoded (currently) for my own (obsolete) Phoenix bookmarks 
file, but it should be fairly straightforward to modify for your own 
purposes.

Hope this is at least close to what you are looking for... :)

- Jamis

-- 
Jamis Buck
jgb3@email.byu.edu
http://www.jamisbuck.org/jamis

"I use octal until I get to 8, and then I switch to decimal."

--------------070106000601070906010503
Content-Type: text/plain;
 name="bookmarks.rb"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="bookmarks.rb"

#!/usr/bin/ruby

class Item
  attr_accessor :last_modified
  attr_accessor :id
  attr_accessor :title
  attr_accessor :remarks

  def to_html_attr_list
    s = ""
    s << " LAST_MODIFIED=\"#{@last_modified}\"" if @last_modified
    s << " ID=\"#{@id}\"" if @id
    return s
  end
end

class Folder < Item
  attr_reader :items

  def initialize
    @items = Array.new
  end

  def dump( level = 0 )
    puts "#{' ' * level * 2}#{title}" if @title
    @items.each do |i|
      i.dump( level+1 )
    end
  end

  def sort!
    @items.sort! do |a,b|
      if a.type == b.type
        a.title.downcase <=> b.title.downcase if a.type == b.type
      elsif a.is_a? Folder
        -1
      elsif b.is_a? Folder
        1
      else
        raise "wrong type in folder"
      end
    end

    @items.each { |i| i.sort! if i.is_a? Folder }
  end

  def to_html( level, file )
    indent = " " * level * 4
    file.puts indent + "<DT><H3#{to_html_attr_list}>#{@title}</H3>" if @title
    file.puts indent + "<DD>#{@remarks}" if @remarks
    file.puts indent + "<HR>" if !@title # hack for top-level folder
    file.puts indent + "<DL><p>"

    @items.each do |i|
      i.to_html( level+1, file )
    end

    file.puts indent + "</DL><p>"
  end
end

class Bookmark < Item
  attr_accessor :last_visit
  attr_accessor :icon
  attr_accessor :last_charset
  attr_accessor :href

  def dump( level )
    print " " * level * 2
    print "'" + @title + "' => "
    puts @href
  end

  def to_html_attr_list
    s = super
    s << " LAST_VISIT=\"#{@last_visit}\"" if @last_visit
    s << " ICON=\"#{@icon}\"" if @icon
    s << " LAST_CHARSET=\"#{@last_charset}\"" if @last_charset
    s << " HREF=\"#{@href}\"" if @href
    return s
  end

  def to_html( level, file )
    indent = " " * level * 4
    file.puts indent + "<DT><A#{to_html_attr_list}>#{@title}</A>"
    file.puts indent + "<DD>#{@remarks}" if @remarks
  end
end

class BookmarkManager
  def initialize
    @top_folder = Folder.new
  end

  def build_attribute_hash( str )
    list = str.scan( /[_A-Z]+="[^"]*"/ )
    hash = Hash.new
    list.each do |item|
      item =~ /([_A-Z]+)="(.*)"/
      hash[ $1 ] = $2
    end
    hash
  end

  def append( bookmarks_file )
    folder_stack = [ @top_folder ]

    File.open( bookmarks_file, "r" ) do |file|
      # skip to the start of the bookmark data
      while ( line = file.gets.strip ) != "<DL><p>"; end

      last_item = nil
      while folder_stack.length > 0
        line = file.gets.strip

        case line
          when /<HR>/ then
            # separator...
            last_item = nil

          when /<DT><H3 (.*)>(.*)<\/H3>/
            last_item = folder = Folder.new
            attr_list = $1
            folder.title = $2
            attrs = build_attribute_hash( attr_list )
            folder.last_modified = attrs[ "LAST_MODIFIED" ]
            folder.id = attrs[ "ID" ]
            folder_stack.last.items.push folder
            folder_stack.push folder

          when /<DT><A (.*)>(.*)<\/A>/
            last_item = bookmark = Bookmark.new
            attr_list = $1
            bookmark.title = $2
            attrs = build_attribute_hash( attr_list )
            bookmark.last_modified = attrs[ "LAST_MODIFIED" ]
            bookmark.id = attrs[ "ID" ]
            bookmark.last_visit = attrs[ "LAST_VISIT" ]
            bookmark.icon = attrs[ "ICON" ]
            bookmark.last_charset = attrs[ "LAST_CHARSET" ]
            bookmark.href = attrs[ "HREF" ]
            folder_stack.last.items.push bookmark

          when /<\/DL><p>/
            folder_stack.pop
            last_item = nil

          when /<DD>(.*)/
            last_item.remarks = $1

          when /<DL><p>/
            # start of a list
        end
      end
    end

    @top_folder.sort!
  end

  def dump
    puts "Bookmarks:"
    @top_folder.dump
  end

  def to_html( file )
    file.puts "<!DOCTYPE NETSCAPE-Bookmark-file-1>"
    file.puts "<!-- This is an automatically generated file."
    file.puts "     It will be read and overwritten."
    file.puts "     DO NOT EDIT! -->"
    file.puts "<META HTTP-EQUIV=\"Content-Type\" CONTENT=\"text/html; charset=UTF-8\">"
    file.puts "<TITLE>Bookmarks</TITLE>"
    file.puts "<H1>Bookmarks</H1>"
    file.puts

    @top_folder.to_html( 0, file )
  end
end


mgr = BookmarkManager.new
mgr.append "/home/jgb3/.phoenix/default/d2isamzz.slt/bookmarks.html"
mgr.to_html( $stdout )

--------------070106000601070906010503--