Class: AsyncFeedbag

Inherits:
Object
  • Object
show all
Defined in:
lib/async-feedbag.rb

Constant Summary collapse

CONTENT_TYPES = 
[
  "application/x.atom+xml",
  "application/atom+xml",
  "application/xml",
  "text/xml",
  "application/rss+xml",
  "application/rdf+xml",
  "application/json",
  "application/feed+json"
].freeze
FEED_SCHEME_RE = 
%r{^feed://}
RedirectionError = 
Class.new(StandardError)
XML_RE = 
/.xml$/
SERVICE_FEED_XPATH = 
"//link[@rel='alternate' or @rel='service.feed'][@href][@type]"
JSON_FEED_XPATH = 
"//link[@rel='alternate' and @type='application/json'][@href]"
FEED_RE = 
%r{(\.(rdf|xml|rss)(\?([\w'\-%]?(=[\w'\-%.]*)?(&|#|\+|;)?)+)?(:[\w'\-%]+)?$|feed=(rss|atom)|(atom|feed)/?$)}i

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(options: nil) ⇒ AsyncFeedbag

Returns a new instance of AsyncFeedbag.



46
47
48
49
50
 
# File 'lib/async-feedbag.rb', line 46

def initialize(options: nil)
  @feeds = []
  @options = options || {}
  @options["User-Agent"] ||= "AsyncFeedbag/#{VERSION}"
end

Class Method Details

.feed?(url) ⇒ Object 



34
35
36
 
# File 'lib/async-feedbag.rb', line 34

def feed?(url)
  new.feed?(url)
end

.find(url, options = {}) ⇒ Object

Parameters:

  • url (String)
  • options (Hash) (defaults to: {}) 


41
42
43
 
# File 'lib/async-feedbag.rb', line 41

def find(url, options = {})
  new(options: options).find(url, **options)
end

Instance Method Details

#add_feed(feed_url, orig_url, base_uri = nil) ⇒ Object 



165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
 
# File 'lib/async-feedbag.rb', line 165

def add_feed(feed_url, orig_url, base_uri = nil)
  url = feed_url.sub(/^feed:/, "").strip

  if base_uri
    url = URI.parse(base_uri).merge(feed_url).to_s
  end

  begin
    uri = URI.parse(url)
  rescue
    puts "Error with `#{url}'"
    exit 1
  end
  unless uri.absolute?
    orig = URI.parse(orig_url)
    url = orig.merge(url).to_s
  end

  # verify url is really valid
  @feeds.push(url) unless @feeds.include?(url)
end

#feed?(url) ⇒ Boolean

Returns:

  • (Boolean) 


53
54
55
56
57
58
59
60
61
62
63
64
 
# File 'lib/async-feedbag.rb', line 53

def feed?(url)
  # use LWR::Simple.normalize some time
  url_uri = URI.parse(url)
  url = "#{url_uri.scheme or "http"}://#{url_uri.host}#{url_uri.path}"
  url << "?#{url_uri.query}" if url_uri.query

  # hack:
  url.sub!(FEED_SCHEME_RE, "http://")

  res = AsyncFeedbag.find(url)
  (res.size == 1) && (res.first == url)
end

#find(url, _options = {}) ⇒ Object 



71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
 
# File 'lib/async-feedbag.rb', line 71

def find(url, _options = {})
  url_uri = URI.parse(url)
  url = nil
  if url_uri.scheme.nil?
    url = "http://#{url_uri}"
  elsif url_uri.scheme == "feed"
    return add_feed(url_uri.to_s.sub(FEED_SCHEME_RE, "http://"), nil)
  else
    url = url_uri.to_s
  end

  # check if feed_valid is avail
  begin
    require "feed_validator"
    v = W3C::FeedValidator.new
    v.validate_url(url)
    return add_feed(url, nil) if v.valid?
  rescue LoadError
    # scoo
  rescue REXML::ParseException
    # usually indicates timeout
    # TODO: actually find out timeout. use Terminator?
    # $stderr.puts "Feed looked like feed but might not have passed validation or timed out"
  rescue => e
    warn "#{e.class} error occurred with: `#{url}': #{e.message}"
  end

  retries = 2
  begin
    headers = @options.slice("User-Agent")
    Sync do
      response = AsyncInternetWithRedirect.get(url, headers)
      if response.redirection?
        original_uri = URI.parse(url)
        uri = URI.parse(response.headers["location"])
        if uri.host == original_uri.host
          url = response.headers["location"]
          raise RedirectionError
        end
      end

      content_type = response.headers["content-type"].gsub(/;.*$/, "").downcase
      next add_feed(url, nil) if CONTENT_TYPES.include?(content_type)

      doc = Nokogiri::HTML(response.read)

      @base_uri = (doc.at("base")["href"] if doc.at("base") && doc.at("base")["href"])

      # first with links
      (doc / "atom:link").each do |l|
        next unless l["rel"] && l["href"].present?

        add_feed(l["href"], url, @base_uri) if l["type"] && CONTENT_TYPES.include?(l["type"].downcase.strip) && (l["rel"].downcase == "self")
      end

      doc.xpath(SERVICE_FEED_XPATH).each do |l|
        add_feed(l["href"], url, @base_uri) if CONTENT_TYPES.include?(l["type"].downcase.strip)
      end

      doc.xpath(JSON_FEED_XPATH).each do |e|
        add_feed(e["href"], url, @base_uri) if looks_like_feed?(e["href"])
      end

      (doc / "a").each do |a|
        next unless a["href"]

        add_feed(a["href"], url, @base_uri) if looks_like_feed?(a["href"]) && (a["href"].include?("/") || a["href"] =~ /#{url_uri.host}/)

        next unless a["href"]

        add_feed(a["href"], url, @base_uri) if looks_like_feed?(a["href"])
      end

      # Added support for feeds like http://tabtimes.com/tbfeed/mashable/full.xml
      add_feed(url, nil) if url.match(XML_RE) && doc.root && doc.root["xml:base"] && (doc.root["xml:base"].strip == url.strip)
    ensure
      response&.close
    end
  rescue RedirectionError
    retries -= 1
    retry if retries >= 0
  rescue Timeout::Error => e
    warn "Timeout error occurred with `#{url}: #{e}'"
  rescue => e
    warn "#{e.class} error occurred with: `#{url}': #{e.message}"
  end
  return @feeds
end

#looks_like_feed?(url) ⇒ Boolean

Returns:

  • (Boolean) 


161
162
163
 
# File 'lib/async-feedbag.rb', line 161

def looks_like_feed?(url)
  FEED_RE.match?(url)
end