DZone Snippets: Timmorgan's Code Snippets

Automatic Expiration of Rails Action Caching

timmorgan (Tim Morgan) — Wed, 16 Jul 2008 03:21:22 GMT

Would love to get some feedback on this...

No doubt, 5 minutes after posting this, someone will tell me of the built-in Rails way of doing this, but alas I could not find it.

Usage looks like this:



class PeopleController < ApplicationController

  caches_action :show, :for => 1.hour, :cache_path => Proc.new { |c| "people/#{c.params[:id]}_for_#{Person.logged_in.id}" }

end

The ":for => 1.hour" part is where the magic happens.

Basically, this bit of code adds a before_filter that checks the last modified time of the cache entry, and expires it if it is older than the specified time period.



module ActionController

  module Caching

    module Fragments

      # expire a cache key only if the block returns true or

      # if the age of the fragment is more than the specified age argument.

      def expire_fragment_by_mtime(key, age=nil, &block)

        block = Proc.new { |m| m < age.ago } unless block_given?

        if (m = cache_store.mtime(fragment_cache_key(key))) and block.call(m)

          expire_fragment(key)

        end

      end

    end

    module Actions

      module ClassMethods

        # adds an option :for

        # caches_action :show, :for => 2.hours, :cache_path => ...

        # :cache_path is required, unfortunately

        def caches_action_with_for(*actions)

          original_actions = actions.clone

          options = actions.extract_options!

          if for_time = options.delete(:for)

            cache_path = options[:cache_path]

            before_filter do |controller|

              cache_path = cache_path.call(controller) if cache_path.respond_to?(:call)

              controller.expire_fragment_by_mtime(cache_path, for_time)

            end

          end

          caches_action_without_for(*original_actions)

        end

        alias_method_chain :caches_action, :for

      end

    end

  end

end



# Add a method to grab the last modified time of the cache key.

# If you use a store other than the FileStore, you'll need to add

# a method like this to your store.

module ActiveSupport

  module Cache

    class FileStore

      def mtime(name)

        File.mtime(real_file_path(name)) rescue nil

      end

    end

  end

end

nslookup.rb

timmorgan (Tim Morgan) — Wed, 02 Jul 2008 20:47:57 GMT

Does a web-based DNS lookup. Useful for those of us behind corporate firewalls that block fun sites based on DNS queries, i.e. OpenDNS.



#!/usr/bin/env ruby



LOOKUP_URL = 'http://toolbar.netcraft.com/site_report?url=%s'

MATCH_RE = /IP address<\/b><\/td>([\d\.]+)<\/td>/

HOSTS_FILE_PATH = '/etc/hosts'



require 'open-uri'



host = ARGV.select { |arg| arg !~ /^\-/ }.first

update_hosts_file = ARGV.select { |arg| arg == '--hosts' }.any?



if open(LOOKUP_URL % host).read =~ MATCH_RE

  ip = $1

  puts ip

  if update_hosts_file

    File.open(HOSTS_FILE_PATH, 'a') do |file|

      file.write("\n#{ip} #{host}")

    end

  end

else

  puts 'There was an error looking up this host.'

end

Usage:

./nslookup.rb youtube.com

To add the entry to your hosts file:

sudo ./nslookup.rb youtube.com --hosts

Trac Wiki to GitHub Wiki

timmorgan (Tim Morgan) — Wed, 02 Apr 2008 02:05:02 GMT
A rough start of a script to help convert the wiki syntax of Trac pages to GitHub-friendly syntax.

Some TLC is still needed on each output page, but better than doing it all by hand.

Project lives here: github.com/seven1m/trac_wiki_to_github

#!/usr/bin/env ruby TRAC_DB_PATH = 'trac.db' OUT_PATH = 'wiki' GITHUB_WIKI_URL = '/seven1m/onebody/wikis/' require 'sqlite3' db = SQLite3::Database.new(TRAC_DB_PATH) pages = db.execute('select name, text from wiki w2 where version = (select max(version) from wiki where name = w2.name);') pages.each do |title, body| File.open(File.join(OUT_PATH, title.gsub(/\s/, '')), 'w') do |file| body.gsub!(/\{\{\{([^\n]+?)\}\}\}/, '\1') body.gsub!(/\{\{\{(.+?)\}\}\}/m, '\1 ') body.gsub!(/====\s(.+?)\s====/, 'h4. \1') body.gsub!(/===\s(.+?)\s===/, 'h3. \1') body.gsub!(/==\s(.+?)\s==/, 'h2. \1') body.gsub!(/=\s(.+?)\s=[\s\n]*/, '') body.gsub!(/\[(http[^\s\[\]]+)\s([^\[\]]+)\]/, '"\2":\1') body.gsub!(/\[([^\s]+)\s(.+)\]/, '"\2":' + GITHUB_WIKI_URL + '\1') body.gsub!(/([^"\/\!])(([A-Z][a-z0-9]+){2,})/, '\1[[\2]]') body.gsub!(/\!(([A-Z][a-z0-9]+){2,})/, '\1') body.gsub!(/'''(.+)'''/, '*\1*') body.gsub!(/''(.+)''/, '_\1_') body.gsub!(/^\s\*/, '*') body.gsub!(/^\s\d\./, '#') file.write(body) end end
Rails Notice/Warning Flash Message timmorgan (Tim Morgan) — Sun, 23 Mar 2008 00:01:03 GMT Somewhat lame, but handy nonetheless. <% if flash[:warning] or flash[:notice] %> class="warning"<% end %>> <%= flash[:warning] || flash[:notice] %> <% end %> Copy Public Key To Host In One Line timmorgan (Tim Morgan) — Wed, 19 Mar 2008 18:23:53 GMT ssh username@host "echo `cat ~/.ssh/id_dsa.pub` >> ~/.ssh/authorized_keys" Ruby Array#every method timmorgan (Tim Morgan) — Thu, 13 Mar 2008 02:46:54 GMT Chunks an array into smaller arrays of the specified size. Updated: Turns out a similar method is available in Rails called Array#in_groups_of. But still, I will leave this up because it seems elegant enough to me. class Array def every(count) chunks = [] each_with_index do |item, index| chunks << [] if index % count == 0 chunks.last << item end chunks end alias / every end Usage: (1..7).to_a.every(2) # => [[1, 2], [3, 4], [5, 6], [7]] (1..7).to_a / 3 # => [[1, 2, 3], [4, 5, 6], [7]] Rails MySQL/SQLite convenience methods timmorgan (Tim Morgan) — Mon, 18 Feb 2008 02:57:36 GMT Usage: Person.find :all, :conditions => ["#{sql_year 'birthday'} >= ?", year] SQLITE = true # or false def sql_concat(*args) SQLITE ? args.join(' || ') : "CONCAT(#{args.join(', ')})" end def sql_lcase(expr) SQLITE ? "LOWER(#{expr})" : "LCASE(#{expr})" end def sql_year(expr) SQLITE ? "CAST(STRFTIME('%y', #{expr}) as 'INTEGER')" : "YEAR(#{expr})" end def sql_month(expr) SQLITE ? "CAST(STRFTIME('%m', #{expr}) as 'INTEGER')" : "MONTH(#{expr})" end def sql_day(expr) SQLITE ? "CAST(STRFTIME('%d', #{expr}) as 'INTEGER')" : "DAY(#{expr})" end def sql_now SQLITE ? "CURRENT_TIMESTAMP" : "NOW()" end Rails Array#add_condition Method timmorgan (Tim Morgan) — Mon, 18 Feb 2008 02:54:44 GMT Lets you add a condition to a set of ActiveRecord conditions easily like this: conditions = ['active = ? and type = ?', true, 2] conditions.add_condition ['person_id = ?', 345] class Array def add_condition(condition, conjunction='and') if condition.is_a? Array if self.empty? (self << condition).flatten! else self[0] += " #{conjunction} " + condition.shift (self << condition).flatten! end elsif condition.is_a? String self[0] += " #{conjunction} " + condition else raise "don't know how to handle this condition type" end self end end Marshalize (Cache) ActiveRecord Query Results timmorgan (Tim Morgan) — Thu, 31 Jan 2008 01:37:39 GMT A quick way to cache results in a file and read from the file on subsequent requests instead of the database. Makes the initial query a bit slower, but later queries *much* faster. class MyCachedModel < ActiveRecord::Base class << self alias_method :rails_original_find_by_sql, :find_by_sql def find_by_sql(sql) cache_filename = Base64.encode64(sql) if File.exists? cache_filename Marshal.load(File.open(cache_filename)) else Marshal.dump(records = rails_original_find_by_sql(sql), File.open(cache_filename, 'w')) return records end end end end Ruby DBF Library Fixes timmorgan (Tim Morgan) — Tue, 29 Jan 2008 18:01:56 GMT A handful of fixes/enhancements for the "dbf" Rubygem. # example.rb require 'rubygems' require 'dbf' require 'dbf_fixes' table = DBF::Table.new('/path/to/table.dbf', :in_memory => false) table.each_record do |record| # do something here # with in_memory=>false tables, each_record is much more efficient # because it reads directly from disk end # dbf_fixes.rb module DBF class Record private # fix bug in DBF code (or workaround bug in FoxPro dbf files; I don't know :-) def initialize_values(columns) columns.each do |column| case column.type when 'I' # added by Tim - I don't understand this much, but it seems to work @attributes[column.name] = @data.read(column.length).unpack("I").first when 'N' # number @attributes[column.name] = column.decimal.zero? ? unpack_string(column).to_i : unpack_string(column).to_f when 'D' # date raw = unpack_string(column).strip unless raw.empty? begin parts = raw.match(DATE_REGEXP).to_a.slice(1,3).map {|n| n.to_i} @attributes[column.name] = Time.gm(*parts) rescue parts = raw.match(DATE_REGEXP).to_a.slice(1,3).map {|n| n.to_i} @attributes[column.name] = Date.new(*parts) end end when 'M' # memo starting_block = unpack_string(column).to_i @attributes[column.name] = read_memo(starting_block) when 'L' # logical @attributes[column.name] = unpack_string(column) =~ /^(y|t)$/i ? true : false else @attributes[column.name] = unpack_string(column).strip end end end # don't know why, but accessors stopped working for me. def define_accessors @table.columns.each do |column| underscored_column_name = underscore(column.name) if @table.options[:accessors] self.class.send :define_method, underscored_column_name do @attributes[column.name] end @@accessors_defined = true end end end end class Table # more efficient iterator (so we don't load everything) def each_record if options[:in_memory] and @records @records.each { |r| yield(r) } else 0.upto(@record_count - 1) do |n| seek_to_record(n) yield(DBF::Record.new(self)) unless deleted_record? end end end end end