== Feedzirra === Description Feedzirra is a feed library that is designed to get and update many feeds as quickly as possible. This includes using libcurl-multi through the taf2-curb[link:http://github.com/taf2/curb/tree/master] gem for faster http gets, and libxml through nokogiri[link:http://github.com/tenderlove/nokogiri/tree/master] and sax-machine[link:http://github.com/pauldix/sax-machine/tree/master] for faster parsing. It allows for easy customization of feed parsing options through the definition of custom parsing classes, and allows you to take as little or as much control as you want in updating feeds. Feedzirra makes it easy to figure out which content in feeds is new by providing simple 'backends' so that Feedzirra can track the last contents fetched from a particular feed. Out of the box, Feedzirra can store this information in the filesystem, Memcached or Tokyo Cabinet. If you want to keep track of new or updated feeds on your own, just use the default backend which will will let you set options for conditional fetching of feeds without the help of Feedzirra. === Installation For now Feedzirra exists only on github. It also has a few gem requirements that are only on github. Before you start you need to have libcurl[link:http://curl.haxx.se/] and libxml[link:http://xmlsoft.org/] installed. If you're on Leopard you have both. Otherwise, you'll need to grab them. Once you've got those libraries, these are the gems that get used: nokogiri, pauldix-sax-machine, taf2-curb (note that this is a fork that lives on github and not the Ruby Forge version of curb), and pauldix-feedzirra. The feedzirra gemspec has all the dependencies so you should be able to get up and running with the standard github gem install routine: gem sources -a http://gems.github.com # if you haven't already gem install pauldix-feedzirra ==== Troubleshooting Installation *NOTE:*Some people have been reporting a few issues related to installation. First, the Ruby Forge version of curb is not what you want. It will not work. Nor will the curl-multi gem that lives on Ruby Forge. You have to get the taf2-curb[link:http://github.com/taf2/curb/tree/master] fork installed. If you see this error when doing a require: /Library/Ruby/Site/1.8/rubygems/custom_require.rb:31:in `gem_original_require': no such file to load -- curb_core (LoadError) It means that the taf2-curb gem didn't build correctly. To resolve this you can do a git clone git://github.com/taf2/curb.git then run rake gem in the curb directory, then sudo gem install pkg/curb-0.2.4.0.gem. After that you should be good. If you see something like this when trying to run it: NoMethodError: undefined method `on_success' for # from ./lib/feedzirra/feed.rb:88:in `add_url_to_multi' This means that you are requiring curl-multi or the Ruby Forge version of Curb somewhere. You can't use those and need to get the taf2 version up and running. If you're on Debian or Ubuntu and getting errors while trying to install the taf2-curb gem, it could be because you don't have the latest version of libcurl installed. Do this to fix: sudo apt-get install libcurl4-gnutls-dev Another problem could be if you are running Mac Ports and you have libcurl installed through there. You need to uninstall it for curb to work! The version in Mac Ports is old and doesn't play nice with curb. If you're running Leopard, you can just uninstall and you should be golden. If you're on an older version of OS X, you'll then need to {download curl}[http://curl.haxx.se/download.html] and build from source. Then you'll have to install the taf2-curb gem again. You might have to perform the step above. If you're still having issues, please let me know on the mailing list. Also, {Todd Fisher (taf2)}[link:http://github.com/taf2] is working on fixing the gem install. Please send him a full error report. === Usage This experimental branch offers a new interface to feed fetching with persistent back-end stores. This allows you to easily run a script retrieving the feeds once per hour or once per day, and it will remember which feeds have been seen before and which are new. This features uses the Feedzirra::Reader interface. You can create a Feedzirra::Reader object after the Feedzirra library (with require 'feedzirra') is loaded as follows: reader = Feedzirra::Reader.new('http://www.woostercollective.com/rss/index.xml') feed = reader.fetch The Reader object can take a single URL or a list of URLs followed by a Hash of options. The options hash allows configuration of the backend store, as well as fetching options for the list of urls. Following is an example of configuration with the Memcache store connected to Tokyo Tyrant (the front-end for Tokyo Cabinet): reader = Feedzirra::Reader.new('http://www.pauldix.net/atom.xml', :backend => { :class => Feedzirra::Backend::Memcache, :port => 1978, :server => 'localhost' }) Other options that may be put in the options hash follow the original API described below. Running reader.fetch will first check the back-end store to see if this feed was fetched previously. If it was previously fetched, Feedzirra uses this information to avoid fetching the whole body if it has already been downloaded based on etag. If the feed has been updated, the new_entries will be populated based on the results of the last query. The back-end store will be updated with the results of every fetch, so Feedzirra will maintain state between executions. Feedzirra currently supports filesystem, memcache and a Ruby Hash structure-based back end that doesn't attempt to persist any information. Once you've retrieved a single feed, you can use the accessors below to query the results. # feed and entries accessors feed.title # => "Paul Dix Explains Nothing" feed.url # => "http://www.pauldix.net" feed.feed_url # => "http://feeds.feedburner.com/PaulDixExplainsNothing" feed.etag # => "GunxqnEP4NeYhrqq9TyVKTuDnh0" feed.last_modified # => Sat Jan 31 17:58:16 -0500 2009 # it's a Time object entry = feed.entries.first entry.title # => "Ruby Http Client Library Performance" entry.url # => "http://www.pauldix.net/2009/01/ruby-http-client-library-performance.html" entry.author # => "Paul Dix" entry.summary # => "..." entry.content # => "..." entry.published # => Thu Jan 29 17:00:19 UTC 2009 # it's a Time object entry.categories # => ["...", "..."] # sanitizing an entry's content entry.title.sanitize # => returns the title with harmful stuff escaped entry.author.sanitize # => returns the author with harmful stuff escaped entry.content.sanitize # => returns the content with harmful stuff escaped entry.content.sanitize! # => returns content with harmful stuff escaped and replaces original (also exists for author and title) entry.sanitize! # => sanitizes the entry's title, author, and content in place (as in, it changes the value to clean versions) feed.sanitize_entries! # => sanitizes all entries in place # updating a single feed updated_feed = Feedzirra::Feed.update(feed) # an updated feed has the following extra accessors updated_feed.updated? # returns true if any of the feed attributes have been modified. will return false if only new entries updated_feed.new_entries # a collection of the entry objects that are newer than the latest in the feed before update # fetching multiple feeds feed_urls = ["http://feeds.feedburner.com/PaulDixExplainsNothing", "http://feeds.feedburner.com/trottercashion"] feeds = Feedzirra::Reader.new(feed_urls).fetch # feeds is now a hash with the feed_urls as keys and the parsed feed objects as values. If an error was thrown # there will be a Fixnum of the http response code instead of a feed object # updating multiple feeds. if you're using a persistent back-end, Feedzirra uses that to determine which entries are ones that you haven't seen before updated_feeds = Feedzirra::reader.new(urls).fetch # defining custom behavior on failure or success. note that a return status of 304 (not updated) will call the on_success handler feed = Feedzirra::Reader.new("http://feeds.feedburner.com/PaulDixExplainsNothing", :on_success => lambda {|feed| puts feed.title }, :on_failure => lambda {|url, response_code, response_header, response_body| puts response_body }).fetch # if a collection was passed into the initializer, the handlers will be called for each one === Extending Feedzirra is easily extended with custom parsing classes and persistent back-ends. You'll have to read the source to find out how, though, because we still haven't written the documentation. :( === Benchmarks One of the goals of Feedzirra is speed. This includes not only parsing, but fetching multiple feeds as quickly as possible. I ran a benchmark getting 20 feeds 10 times using Feedzirra, rFeedParser, and FeedNormalizer. For more details the {benchmark code can be found in the project in spec/benchmarks/feedzirra_benchmarks.rb}[http://github.com/pauldix/feedzirra/blob/7fb5634c5c16e9c6ec971767b462c6518cd55f5d/spec/benchmarks/feedzirra_benchmarks.rb] feedzirra 5.170000 1.290000 6.460000 ( 18.917796) rfeedparser 104.260000 12.220000 116.480000 (244.799063) feed-normalizer 66.250000 4.010000 70.260000 (191.589862) The result of that benchmark is a bit sketchy because of the network variability. Running 10 times against the same 20 feeds was meant to smooth some of that out. However, there is also a {benchmark comparing parsing speed in spec/benchmarks/parsing_benchmark.rb}[http://github.com/pauldix/feedzirra/blob/7fb5634c5c16e9c6ec971767b462c6518cd55f5d/spec/benchmarks/parsing_benchmark.rb] on an atom feed. feedzirra 0.500000 0.030000 0.530000 ( 0.658744) rfeedparser 8.400000 1.110000 9.510000 ( 11.839827) feed-normalizer 5.980000 0.160000 6.140000 ( 7.576140) There's also a {benchmark that shows the results of using Feedzirra to perform updates on feeds}[http://github.com/pauldix/feedzirra/blob/45d64319544c61a4c9eb9f7f825c73b9f9030cb3/spec/benchmarks/updating_benchmarks.rb] you've already pulled in. I tested against 179 feeds. The first is the initial pull and the second is an update 65 seconds later. I'm not sure how many of them support etag and last-modified, so performance may be better or worse depending on what feeds you're requesting. feedzirra fetch and parse 4.010000 0.710000 4.720000 ( 15.110101) feedzirra update 0.660000 0.280000 0.940000 ( 5.152709) === Discussion I'd like feedback on the api and any bugs encountered on feeds in the wild. I've set up a {google group here}[http://groups.google.com/group/feedzirra]. === TODO This thing needs to hammer on many different feeds in the wild. I'm sure there will be bugs. I want to find them and crush them. I didn't bother using the test suite for feedparser. i wanted to start fresh. Here are some more specific TODOs. * Make a feedzirra-rails gem to integrate feedzirra seamlessly with Rails and ActiveRecord. * Add support for authenticated feeds. * Create a super sweet DSL for defining new parsers. * Test against Ruby 1.9.1 and fix any bugs. * I'm not keeping track of modified on entries. Should I add this? * Clean up the fetching code inside feed.rb so it doesn't suck so hard. * Make the feed_spec actually mock stuff out so it doesn't hit the net. * Readdress how feeds determine if they can parse a document. Maybe I should use namespaces instead? === LICENSE (The MIT License) Copyright (c) 2009: {Paul Dix}[http://pauldix.net] Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the 'Software'), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED 'AS IS', WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.