1
0
Fork 0
forked from fedi/mastodon
mastodon/app/services/fetch_oembed_service.rb
ThibG b364175e1b Fix link crawler not specifying accepted content-type (#12646)
The link crawler expects HTML documents, so set the `Accept`
header accordingly.

Fixes #12618
2019-12-18 16:56:06 +01:00

101 lines
2.5 KiB
Ruby

# frozen_string_literal: true
class FetchOEmbedService
ENDPOINT_CACHE_EXPIRES_IN = 24.hours.freeze
attr_reader :url, :options, :format, :endpoint_url
def call(url, options = {})
@url = url
@options = options
if @options[:cached_endpoint]
parse_cached_endpoint!
else
discover_endpoint!
end
fetch!
end
private
def discover_endpoint!
return if html.nil?
@format = @options[:format]
page = Nokogiri::HTML(html)
if @format.nil? || @format == :json
@endpoint_url ||= page.at_xpath('//link[@type="application/json+oembed"]')&.attribute('href')&.value
@format ||= :json if @endpoint_url
end
if @format.nil? || @format == :xml
@endpoint_url ||= page.at_xpath('//link[@type="text/xml+oembed"]')&.attribute('href')&.value
@format ||= :xml if @endpoint_url
end
return if @endpoint_url.blank?
@endpoint_url = (Addressable::URI.parse(@url) + @endpoint_url).to_s
cache_endpoint!
rescue Addressable::URI::InvalidURIError
@endpoint_url = nil
end
def parse_cached_endpoint!
cached = @options[:cached_endpoint]
return if cached[:endpoint].nil? || cached[:format].nil?
@endpoint_url = Addressable::Template.new(cached[:endpoint]).expand(url: @url).to_s
@format = cached[:format]
end
def cache_endpoint!
url_domain = Addressable::URI.parse(@url).normalized_host
endpoint_hash = {
endpoint: @endpoint_url.gsub(/(=(http[s]?(%3A|:)(\/\/|%2F%2F)))([^&]*)/i, '={url}'),
format: @format,
}
Rails.cache.write("oembed_endpoint:#{url_domain}", endpoint_hash, expires_in: ENDPOINT_CACHE_EXPIRES_IN)
end
def fetch!
return if @endpoint_url.blank?
body = Request.new(:get, @endpoint_url).perform do |res|
res.code != 200 ? nil : res.body_with_limit
end
validate(parse_for_format(body)) if body.present?
rescue Oj::ParseError, Ox::ParseError
nil
end
def parse_for_format(body)
case @format
when :json
Oj.load(body, mode: :strict)&.with_indifferent_access
when :xml
Ox.load(body, mode: :hash_no_attrs)&.with_indifferent_access&.dig(:oembed)
end
end
def validate(oembed)
oembed if oembed[:version] == '1.0' && oembed[:type].present?
end
def html
return @html if defined?(@html)
@html = @options[:html] || Request.new(:get, @url).add_headers('Accept' => 'text/html').perform do |res|
res.code != 200 || res.mime_type != 'text/html' ? nil : res.body_with_limit
end
end
end