discourse/app/models/post_analyzer.rb

181 lines
4.4 KiB
Ruby
Raw Normal View History

# frozen_string_literal: true
class PostAnalyzer
def initialize(raw, topic_id)
2017-07-27 20:20:09 -05:00
@raw = raw
@topic_id = topic_id
@onebox_urls = []
@found_oneboxes = false
2016-04-12 13:09:59 -05:00
end
def found_oneboxes?
@found_oneboxes
end
def has_oneboxes?
return false unless @raw.present?
cooked_stripped
found_oneboxes?
end
# What we use to cook posts
def cook(raw, opts = {})
cook_method = opts[:cook_method]
return raw if cook_method == Post.cook_methods[:raw_html]
if cook_method == Post.cook_methods[:email]
cooked = EmailCook.new(raw).cook(opts)
else
cooked = PrettyText.cook(raw, opts)
end
limit = SiteSetting.max_oneboxes_per_post
result =
Oneboxer.apply(cooked, extra_paths: ".inline-onebox-loading") do |url, element|
if opts[:invalidate_oneboxes]
Oneboxer.invalidate(url)
InlineOneboxer.invalidate(url)
end
next if element["class"] != Oneboxer::ONEBOX_CSS_CLASS
next if limit <= 0
limit -= 1
@onebox_urls << url
onebox = Oneboxer.cached_onebox(url)
@found_oneboxes = true if onebox.present?
onebox
end
if result.changed?
PrettyText.sanitize_hotlinked_media(result.doc)
cooked = result.to_html
end
cooked
end
# How many images are present in the post
def embedded_media_count
return 0 unless @raw.present?
# TODO - do we need to look for tags other than img, video and audio?
cooked_stripped
.css("img", "video", "audio")
.reject do |t|
if dom_class = t["class"]
(Post.allowed_image_classes & dom_class.split).count > 0
end
end
.count
end
2013-07-21 19:39:17 -05:00
# How many attachments are present in the post
def attachment_count
return 0 unless @raw.present?
2013-10-17 11:44:09 -05:00
attachments =
cooked_stripped.css("a.attachment[href^=\"#{Discourse.store.absolute_base_url}\"]")
attachments +=
cooked_stripped.css(
"a.attachment[href^=\"#{Discourse.store.relative_base_url}\"]",
) if Discourse.store.internal?
attachments.count
2013-07-21 19:39:17 -05:00
end
def raw_mentions
return [] if @raw.blank?
return @raw_mentions if @raw_mentions.present?
@raw_mentions = PrettyText.extract_mentions(cooked_stripped)
end
# from rack ... compat with ruby 2.2
def self.parse_uri_rfc2396(uri)
@parser ||= defined?(URI::RFC2396_Parser) ? URI::RFC2396_Parser.new : URI
@parser.parse(uri)
end
# Count how many hosts are linked in the post
def linked_hosts
all_links = raw_links + @onebox_urls
return {} if all_links.blank?
return @linked_hosts if @linked_hosts.present?
@linked_hosts = {}
2013-10-17 11:44:09 -05:00
all_links.each do |u|
2013-07-15 05:11:23 -05:00
begin
uri = self.class.parse_uri_rfc2396(u)
2013-07-15 05:11:23 -05:00
host = uri.host
@linked_hosts[host] ||= 1 unless host.nil?
rescue URI::Error
# An invalid URI does not count as a host
2013-07-15 05:11:23 -05:00
next
end
end
2013-10-17 11:44:09 -05:00
@linked_hosts
end
# Returns an array of all links in a post excluding mentions
def raw_links
return [] unless @raw.present?
return @raw_links if @raw_links.present?
@raw_links = []
cooked_stripped
.css("a")
.each do |l|
# Don't include @mentions in the link count
next if link_is_a_mention?(l)
# Don't include heading anchor in the link count
next if link_is_an_anchor?(l)
# Don't include hashtags in the link count
next if link_is_a_hashtag?(l)
@raw_links << l["href"].to_s
end
2013-10-17 11:44:09 -05:00
@raw_links
end
# How many links are present in the post
def link_count
raw_links.size + @onebox_urls.size
end
2018-06-07 00:28:18 -05:00
def cooked_stripped
@cooked_stripped ||=
begin
doc = Nokogiri::HTML5.fragment(cook(@raw, topic_id: @topic_id))
doc.css(
"pre .mention, aside.quote > .title, aside.quote .mention, aside.quote .mention-group, .onebox, .elided",
).remove
doc
end
2018-06-07 00:28:18 -05:00
end
private
2018-06-07 00:28:18 -05:00
def link_is_a_mention?(l)
href = l["href"].to_s
l["class"].to_s["mention"] &&
(
href.start_with?("#{Discourse.base_path}/u/") ||
href.start_with?("#{Discourse.base_path}/users/")
)
end
def link_is_an_anchor?(l)
l["class"].to_s["anchor"] && l["href"].to_s.start_with?("#")
end
def link_is_a_hashtag?(l)
href = l["href"].to_s
l["class"].to_s["hashtag"] &&
(
href.start_with?("#{Discourse.base_path}/c/") ||
href.start_with?("#{Discourse.base_path}/tag/")
)
2018-06-07 00:28:18 -05:00
end
end