= feedme

* http://wiki.github.com/jdidion/feedme


A simple, flexible, and extensible RSS and Atom parser for Ruby. Based on the popular SimpleRSS library, but with many nice extra features.


* Parse RSS 0.91, 0.92, 1.0, and 2.0
* Parse Atom
* Parse all tags by default, or choose the tags you want to parse
* Access all attributes and content as if they were methods
* Access all values of tags that can appear multiple times
* Delicious syntactic sugar that makes it simple to get the data you want


The API is similar to SimpleRSS:

require 'rubygems'
require 'feedme'
require 'open-uri'

rss = FeedMe.parse open('http://slashdot.org/index.rdf')
rss.version # => 1.0
rss.channel.title # => "Slashdot"
rss.channel.link # => "http://slashdot.org/"
rss.items.first.link # => "http://books.slashdot.org/article.pl?sid=05/08/29/1319236&from=rss"

But since the parser can read Atom feeds as easily as RSS feeds, there are aliases that allow more atom like reading:

rss.feed.title # => "Slashdot"
rss.feed.link # => "http://slashdot.org/"
rss.entries.first.link # => "http://books.slashdot.org/article.pl?sid=05/08/29/1319236&from=rss"

Under the covers, all element values are stored in arrays. This means that you can access all content for an element that appears multiple times (i.e. category):

rss.items.first.category_array # => ["News for Nerds", "Technology"]
rss.items.first.category # => "News for Nerds"

You also have access to all the attributes as well as tag values:

rss.items.first.guid.isPermaLink # => "true"
rss.items.first.guid.content # => http://books.slashdot.org/article.pl?sid=05/08/29/1319236

FeedMe also adds some syntactic sugar that makes it easy to get the information you want:

rss.items.first.category? # => true
rss.items.first.category_count # => 2
rss.items.first.guid_value # => http://books.slashdot.org/article.pl?sid=05/08/29/1319236

There are two different parsers that you can use, depending on your needs. The default parser is "promiscuous," meaning that it parses all tags. There is also a strict parser that only parses tags specified in a list. Here is how you create the different types of parsers:

FeedMe.parse(source) # parse using the default (promiscuous) parser
FeedMe::ParserBuilder.new.parse(source) # equivalent to the previous line
FeedMe::StrictParserBuilder.new.parse(source) # only parse certain tags

The FeedMe class methods and the parser builder constructors also accept an options hash. Options are also passed on to the Parser constructor. Currently, only two options are available:

1. :empty_string_for_nil => false # return the empty string instead of a nil value
2. :error_on_missing_key => false # raise an error if a specified key or virtual method does not exist (otherwise nil is returned)

The strict parser can be extended by adding new tags to parse:

builder = FeedMe::StrictParserBuilder.new
builder.rss_tags << :some_new_tag
builder.rss_item_tags << :'item+myrel' # parse an item that has a custom rel type
builder.item_ext_tags << :feedburner_origLink # parse an extension tag - one that has a specific
# namespace (use '_', not ':', to separate namespace
# from attribute name)

Either parser can be extended by adding aliases to existing tags:

builder.aliases[:updated] => :pubDate # now you can always access the updated date using :updated,
# regardless of whether it's an RSS or Atom feed

If you don't know ahead of time what type of feed you'll be parsing, you can tell FeedMe to always emulate RSS or Atom. These methods just add a bunch of aliases:


Another bit of syntactic sugar are transformations. These are modifications that can be applied to feed content. There is a default transformation that can be applied by adding '!' to the tag name.

rss.entry.content # => <div>Some great stuff</div>
rss.entry.content! # => Some great stuff

The default transformation can be changed:

builder.default_transformation = [ :cleanHtml ]

Custom transformations are defined by mapping one or more transformation functions to a suffix:

builder.transformations['clean'] = [ :cleanHtml ]

rss.entry.content # => <div>This is a bunch of text</div><p></p></html>
rss.entry.content_clean # => <div>This is a bunch of text</div>

You can also/instead apply an arbitrary set of transformations via the transform method:

rss.entry.transform(:content, [ :clean, [ :trunc, 50 ] ])

You can create your own transformation function. The following is an example of a transformation function that takes an argument. Note that transformation function names may only contain alphanumeric characters. Argument values are specified at the end separated by underscores.

builder.transformation_fns[:wrap] => proc {|str, col|
str.gsub(/(.{1,#{col}})( +|$\n?)|(.1,1,#{col})/, "\\1\\3\n").strip
builder.transformations['wrap'] = [ :wrap_10 ]

rss.entry.content = This is a bunch of text
rss.entry.content_wrap = This is a
bunch of

The transformation functions available by default are:

1. :stripHtml - described above
2. :cleanHtml - ** Requires FeedNormalizer (which in turn requires Hypricot) **

rss.entry_array[0].content # => 1 > 2
rss.entry_array[0].content! # => 1 &gt; 2

rss.entry_array[1].content # => <div>Some great stuff</div><p></p></html>
rss.entry_array[1].content! # => <div>Some great stuff</div>

3. :wrap - takes number of columns as a parameter. Respects word boundaries. Example of :wrap_10:

rss.entry.content # => This is a bunch of text
rss.entry.content! # => This is a
bunch of

4. :trunc - truncates text to a certain length. Example of :trunc_10:

rss.entries.first.content # => This is a long long long sentence
rss.entries.first.content! # => This is a

5. :truncHtml - truncates the content inside the first set of HTML tags, but preserves the tags. ** Requires ActiveSupport and Hpricot ** Example of :truncHtml_10:

rss.entries.first.content # => <div>This is a long long long sentence</div></html>
rss.entries.first.content! # => <div>This is a </div></html>

6. :regexp - apply a regular expression and extract the capture groups

rss.entries.first.content # => This is a long long long entry
rss.entries.first.transform(:content, [ :regexp, /(This is a long ).*(entry)/ ]) # => This is a long entry

In order to prevent clashes between tag/attribute names and the parser class' instance variables, all instance variables are prefixed with 'fm_'. They are:

fm_source # the original, unparsed source
fm_options # the options passed to the parser constructor
fm_type # the feed type
fm_tags # the tags the parser looks for in the source
fm_parsed # the list of tags the parser actually found
fm_unparsed # the list of tags that appeared in the feed but were not parsed (useful for debugging)

Additionally, there are several variables that are available at every level of the parse tree:

fm_builder # the ParserBuilder that created the parser
fm_parent # the container of the current level of the parse tree
fm_tag_name # the name of the rss/atom tag whose content is contained in this level of the tree

=== A word on RSS/Atom Versions

RSS has undergone much revision since Netscape 0.90 was released in 1999. The current accepted specification is maintained by the RSS Advisory board (http://www.rssboard.org/rss-specification). The current version (as of this writing) is 2.0.11, although the actual schema has not changed since 2.0.1.

Atom is an IETF standard (http://www.w3.org/2005/Atom) and so far there is a single version of the specification (commonly referred to as 1.0).

FeedMe does its best to support RSS and Atom versions currently in use. It specifically does *NOT* support any Netscape version of RSS.

Due to various incompatibilities between different RSS versions, it is strongly recommended that you examine the version attribute of the feed (as shown in the Usage section). Mark Pilgrim has an excellent article on RSS version incompatibility: http://diveintomark.org/archives/2004/02/04/incompatible-rss.


* gem install jdidion-feedme (Add GitHub as a gem source: gem sources -a http://gems.github.com)
* http://github.com/jdidion/feedme/downloads

To use certain features of FeedMe, some dependencies are required:
* To use the :truncHtml transformation for truncating HTML content, ActiveSupport and Hpricot are required

sudo gem install activesupport
sudo gem install hpricot

* To use the :cleanHtml for sanitizing HTML, FeedNormalizer and Hpricot are required

sudo gem install feed-normalizer
sudo gem install hpricot


This work is licensed under the Creative Commons Attribution 3.0 United States License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/us/ or send a letter to Creative Commons, 171 Second Street, Suite 300, San Francisco, California, 94105, USA.