Question

Background: I am playing around with the feedzirra plugin in a rails app, and I am attempting to utilize both the feed and entry accessors of the parsed feed in a single view.

To accomplish this, I have created two models: feed and feed_entry. They are as follows:

feed.rb:

class Feed < ActiveRecord::Base
  attr_accessible :title, :url, :feed_url, :last_modified, :guid
  has_many :feed_entries

  def self.update_from_feed(feed_url)
    feed = Feedzirra::Feed.fetch_and_parse(feed_url)
    add_feed(feed)
    FeedEntry.add_entries(feed.entries)
  end


  private

  def self.add_feed(feed)
    unless exists? :guid => feed.id
      create!(
        :title          => feed.title,
        :url            => feed.url,
        :feed_url       => feed.feed_url,
        :last_modified  => feed.last_modified,
        :guid           => feed.id
      )
    end
  end
end

feed_entry.rb:

class FeedEntry < ActiveRecord::Base
  attr_accessible :title, :url, :author, :summary, :content, :published, :guid
  belongs_to :feed

  def self.add_entries(entries)
    entries.each do |entry|
      unless exists? :guid => entry.id
        create!(
          :title        => entry.title,
          :url          => entry.url,
          :author       => entry.author,
          :summary      => entry.summary,
          :content      => entry.content,
          :published    => entry.published,
          :guid         => entry.id
        )
      end
    end
  end
end

Where the intention is to call Feed.update_from_feed(feed_url) as a cron job. In reality, I’m hoping to have the cron job call this function by passing in multiple feeds; but I figured I’d get one working first. For sake of testing, I’m currently calling this function from the console.

Questions:

NoMethodError: undefined method `id' for #<Feedzirra::Parser::RSS:0x00000100bcb0f0>

1. This may seem very simple, but in calling the update_from_feed function, I get the above error in my console. While new to ruby and rails, I was under the impression that every new entry in a table is given an id that can be referenced – in this scenario, by feed.id. Am I missing something here?

2. Is my 2-model approach to using this plugin a good direction to go in? Since I’m not able to verify the rest of my approach works until I puzzle out question 1, I thought I’d ask for a tip on best practice in the mean time.

3. The rest of my plan is to utilize the feed and entry accessors of the plugin within a single controller, and then reference those variables within the relative view somehow. Is this how I should be perceiving the MVC architecture and using it in a RESTful way? If so, how do I reference the feed and entry accessors in the FeedReader view?

This is what the FeedReader controller would look like:

class FeedEntriesController < ApplicationController
  def index
    @feed_entries = FeedEntry.all
    @feeds = Feed.all
  end
end

I know this was a bitch of a post, but any help would really be appreciated.

Answer 1

There are a couple of flaws in your approach.

Firstly, when you update your feed, you are checking whether or not exists (great) but then not dealing with the possibility that it does and then you try to collect new feeds anyway. You need something like;

feed = Feed.find_by_url(incoming_url) # try loading an existing feed first
feed = Feedzirra::Feed.fetch_and_parse(feed_url) if !feed # if not, create a new one

Assuming that the last Feedzirra command is the one to create a new feed. I can’t remember!

The two model idea seems like a sensible approach. When accessing your feeds via the controllers later, you should probably try nested resources. Then your URL would look something like;

/feed/1/entries

And your controller would be something like;

@feed = Feed.find(params[:feed_id])
@entries = @feed.entries

I haven’t answered all your questions, but hopefully it’s a start.