discourse/lib/pretty_text.rb

318 lines
9.9 KiB
Ruby
Raw Normal View History

2013-02-05 14:16:51 -05:00
require 'v8'
require 'nokogiri'
module PrettyText
def self.whitelist
{
elements: %w[
2013-02-05 14:16:51 -05:00
a abbr aside b bdo blockquote br caption cite code col colgroup dd div del dfn dl
dt em hr figcaption figure h1 h2 h3 h4 h5 h6 hgroup i img ins kbd li mark
ol p pre q rp rt ruby s samp small span strike strong sub sup table tbody td
tfoot th thead time tr u ul var wbr
],
attributes: {
2013-02-05 14:16:51 -05:00
:all => ['dir', 'lang', 'title', 'class'],
'aside' => ['data-post', 'data-full', 'data-topic'],
'a' => ['href'],
'blockquote' => ['cite'],
'col' => ['span', 'width'],
'colgroup' => ['span', 'width'],
'del' => ['cite', 'datetime'],
'img' => ['align', 'alt', 'height', 'src', 'width'],
'ins' => ['cite', 'datetime'],
'ol' => ['start', 'reversed', 'type'],
'q' => ['cite'],
'span' => ['style'],
'table' => ['summary', 'width', 'style', 'cellpadding', 'cellspacing'],
'td' => ['abbr', 'axis', 'colspan', 'rowspan', 'width', 'style'],
'th' => ['abbr', 'axis', 'colspan', 'rowspan', 'scope', 'width', 'style'],
'time' => ['datetime', 'pubdate'],
'ul' => ['type']
},
protocols: {
2013-02-05 14:16:51 -05:00
'a' => {'href' => ['ftp', 'http', 'https', 'mailto', :relative]},
'blockquote' => {'cite' => ['http', 'https', :relative]},
'del' => {'cite' => ['http', 'https', :relative]},
'img' => {'src' => ['http', 'https', :relative]},
'ins' => {'cite' => ['http', 'https', :relative]},
'q' => {'cite' => ['http', 'https', :relative]}
}
2013-02-25 11:42:20 -05:00
}
2013-02-05 14:16:51 -05:00
end
class Helpers
2013-02-25 11:42:20 -05:00
# function here are available to v8
2013-02-05 14:16:51 -05:00
def avatar_template(username)
return "" unless username
user = User.where(username_lower: username.downcase).first
if user
user.avatar_template
end
end
def is_username_valid(username)
return false unless username
username = username.downcase
return User.exec_sql('select 1 from users where username_lower = ?', username).values.length == 1
end
end
@mutex = Mutex.new
def self.mention_matcher
Regexp.new("(\@[a-zA-Z0-9_]{#{User.username_length.begin},#{User.username_length.end}})")
2013-02-25 11:42:20 -05:00
end
2013-02-05 14:16:51 -05:00
def self.app_root
Rails.root
end
def self.v8
return @ctx unless @ctx.nil?
@ctx = V8::Context.new
2013-02-25 11:42:20 -05:00
@ctx["helpers"] = Helpers.new
2013-02-05 14:16:51 -05:00
@ctx.load(app_root + "app/assets/javascripts/external/md5.js")
2013-02-05 14:16:51 -05:00
@ctx.load(app_root + "app/assets/javascripts/external/Markdown.Converter.js")
2013-02-25 11:42:20 -05:00
@ctx.load(app_root + "app/assets/javascripts/external/twitter-text-1.5.0.js")
2013-02-05 14:16:51 -05:00
@ctx.load(app_root + "lib/headless-ember.js")
2013-02-25 11:42:20 -05:00
@ctx.load(app_root + "app/assets/javascripts/external/rsvp.js")
2013-02-05 14:16:51 -05:00
@ctx.load(Rails.configuration.ember.handlebars_location)
#@ctx.load(Rails.configuration.ember.ember_location)
2013-04-15 01:37:31 -04:00
@ctx.load(app_root + "app/assets/javascripts/external_production/sugar-1.3.5.js")
2013-02-05 14:16:51 -05:00
@ctx.eval("var Discourse = {}; Discourse.SiteSettings = #{SiteSetting.client_settings_json};")
@ctx.eval("var window = {}; window.devicePixelRatio = 2;") # hack to make code think stuff is retina
@ctx.load(app_root + "app/assets/javascripts/discourse/components/bbcode.js")
@ctx.load(app_root + "app/assets/javascripts/discourse/components/utilities.js")
@ctx.load(app_root + "app/assets/javascripts/discourse/components/markdown.js")
2013-02-05 14:16:51 -05:00
# Load server side javascripts
if DiscoursePluginRegistry.server_side_javascripts.present?
DiscoursePluginRegistry.server_side_javascripts.each do |ssjs|
@ctx.load(ssjs)
end
end
@ctx['quoteTemplate'] = File.open(app_root + 'app/assets/javascripts/discourse/templates/quote.js.shbrs') {|f| f.read}
@ctx['quoteEmailTemplate'] = File.open(app_root + 'lib/assets/quote_email.js.shbrs') {|f| f.read}
2013-02-25 11:42:20 -05:00
@ctx.eval("HANDLEBARS_TEMPLATES = {
2013-02-05 14:16:51 -05:00
'quote': Handlebars.compile(quoteTemplate),
'quote_email': Handlebars.compile(quoteEmailTemplate),
};")
@ctx
end
def self.markdown(text, opts=nil)
# we use the exact same markdown converter as the client
2013-02-25 11:42:20 -05:00
# TODO: use the same extensions on both client and server (in particular the template for mentions)
2013-02-05 14:16:51 -05:00
baked = nil
2013-02-25 11:42:20 -05:00
@mutex.synchronize do
2013-02-05 14:16:51 -05:00
# we need to do this to work in a multi site environment, many sites, many settings
v8.eval("Discourse.SiteSettings = #{SiteSetting.client_settings_json};")
v8.eval("Discourse.BaseUrl = 'http://#{RailsMultisite::ConnectionManagement.current_hostname}';")
v8.eval("Discourse.getURL = function(url) {return '#{Discourse::base_uri}' + url};")
2013-02-05 14:16:51 -05:00
v8['opts'] = opts || {}
v8['raw'] = text
v8.eval('opts["mentionLookup"] = function(u){return helpers.is_username_valid(u);}')
v8.eval('opts["lookupAvatar"] = function(p){return Discourse.Utilities.avatarImg({username: p, size: "tiny", avatarTemplate: helpers.avatar_template(p)});}')
baked = v8.eval('Discourse.Markdown.markdownConverter(opts).makeHtml(raw)')
2013-02-05 14:16:51 -05:00
end
# we need some minimal server side stuff, apply CDN and TODO filter disallowed markup
2013-02-25 11:42:20 -05:00
baked = apply_cdn(baked, Rails.configuration.action_controller.asset_host)
2013-02-05 14:16:51 -05:00
baked
end
# leaving this here, cause it invokes v8, don't want to implement twice
def self.avatar_img(username, size)
r = nil
2013-02-25 11:42:20 -05:00
@mutex.synchronize do
2013-02-05 14:16:51 -05:00
v8['username'] = username
v8['size'] = size
v8.eval("Discourse.SiteSettings = #{SiteSetting.client_settings_json};")
v8.eval("Discourse.CDN = '#{Rails.configuration.action_controller.asset_host}';")
v8.eval("Discourse.BaseUrl = '#{RailsMultisite::ConnectionManagement.current_hostname}';")
2013-02-25 11:42:20 -05:00
r = v8.eval("Discourse.Utilities.avatarImg({ username: username, size: size });")
2013-02-05 14:16:51 -05:00
end
r
end
def self.apply_cdn(html, url)
return html unless url
image = /\.(jpg|jpeg|gif|png|tiff|tif)$/
doc = Nokogiri::HTML.fragment(html)
doc.css("a").each do |l|
href = l.attributes["href"].to_s
if href[0] == '/' && href =~ image
l["href"] = url + href
end
end
doc.css("img").each do |l|
src = l.attributes["src"].to_s
2013-02-25 11:42:20 -05:00
if src[0] == '/'
2013-02-05 14:16:51 -05:00
l["src"] = url + src
end
end
doc.to_s
end
def self.cook(text, opts={})
cloned = opts.dup
# we have a minor inconsistency
cloned[:topicId] = opts[:topic_id]
2013-02-25 11:42:20 -05:00
sanitized = Sanitize.clean(markdown(text.dup, cloned), PrettyText.whitelist)
if SiteSetting.add_rel_nofollow_to_user_content
2013-02-25 11:42:20 -05:00
sanitized = add_rel_nofollow_to_user_content(sanitized)
end
sanitized
end
2013-02-25 11:42:20 -05:00
def self.add_rel_nofollow_to_user_content(html)
whitelist = []
2013-02-11 03:01:33 -05:00
l = SiteSetting.exclude_rel_nofollow_domains
if l.present?
2013-02-25 11:42:20 -05:00
whitelist = l.split(",")
end
2013-02-11 03:01:33 -05:00
site_uri = nil
doc = Nokogiri::HTML.fragment(html)
doc.css("a").each do |l|
href = l["href"].to_s
2013-02-25 11:42:20 -05:00
begin
uri = URI(href)
site_uri ||= URI(Discourse.base_url)
2013-02-25 11:42:20 -05:00
if !uri.host.present? ||
uri.host.ends_with?(site_uri.host) ||
2013-02-11 03:01:33 -05:00
whitelist.any?{|u| uri.host.ends_with?(u)}
# we are good no need for nofollow
else
l["rel"] = "nofollow"
end
rescue URI::InvalidURIError
2013-02-25 11:42:20 -05:00
# add a nofollow anyway
l["rel"] = "nofollow"
end
end
doc.to_html
2013-02-05 14:16:51 -05:00
end
def self.extract_links(html)
doc = Nokogiri::HTML.fragment(html)
links = []
doc.css("a").each do |l|
links << l.attributes["href"].to_s
end
doc.css("aside.quote").each do |a|
topic_id = a.attributes['data-topic']
2013-02-25 11:42:20 -05:00
url = "/t/topic/#{topic_id}"
if post_number = a.attributes['data-post']
url << "/#{post_number}"
end
links << url
end
2013-02-05 14:16:51 -05:00
links
end
class ExcerptParser < Nokogiri::XML::SAX::Document
2013-02-25 11:42:20 -05:00
class DoneException < StandardError; end
2013-02-05 14:16:51 -05:00
attr_reader :excerpt
2013-04-29 23:25:55 -04:00
def initialize(length,options)
2013-02-25 11:42:20 -05:00
@length = length
2013-02-05 14:16:51 -05:00
@excerpt = ""
@current_length = 0
2013-04-29 23:25:55 -04:00
@strip_links = options[:strip_links] == true
2013-02-05 14:16:51 -05:00
end
2013-04-29 23:25:55 -04:00
def self.get_excerpt(html, length, options)
me = self.new(length,options)
2013-02-05 14:16:51 -05:00
parser = Nokogiri::HTML::SAX::Parser.new(me)
2013-02-25 11:42:20 -05:00
begin
2013-02-05 14:16:51 -05:00
parser.parse(html) unless html.nil?
rescue DoneException
2013-02-25 11:42:20 -05:00
# we are done
2013-02-05 14:16:51 -05:00
end
me.excerpt
end
def start_element(name, attributes=[])
2013-02-25 11:42:20 -05:00
case name
2013-02-05 14:16:51 -05:00
when "img"
attributes = Hash[*attributes.flatten]
if attributes["alt"]
characters("[#{attributes["alt"]}]")
elsif attributes["title"]
characters("[#{attributes["title"]}]")
else
characters("[image]")
end
when "a"
2013-04-29 23:25:55 -04:00
unless @strip_links
c = "<a "
c << attributes.map{|k,v| "#{k}='#{v}'"}.join(' ')
c << ">"
characters(c, false, false, false)
@in_a = true
end
2013-02-05 14:16:51 -05:00
when "aside"
2013-02-25 11:42:20 -05:00
@in_quote = true
2013-02-05 14:16:51 -05:00
end
end
def end_element(name)
2013-02-25 11:42:20 -05:00
case name
2013-02-05 14:16:51 -05:00
when "a"
2013-04-29 23:25:55 -04:00
unless @strip_links
characters("</a>",false, false, false)
@in_a = false
end
2013-02-05 14:16:51 -05:00
when "p", "br"
characters(" ")
when "aside"
@in_quote = false
end
end
def characters(string, truncate = true, count_it = true, encode = true)
return if @in_quote
2013-02-25 11:42:20 -05:00
encode = encode ? lambda{|s| ERB::Util.html_escape(s)} : lambda {|s| s}
if count_it && @current_length + string.length > @length
length = [0, @length - @current_length - 1].max
@excerpt << encode.call(string[0..length]) if truncate
2013-02-05 14:16:51 -05:00
@excerpt << "&hellip;"
@excerpt << "</a>" if @in_a
raise DoneException.new
end
@excerpt << encode.call(string)
@current_length += string.length if count_it
end
end
2013-04-29 23:25:55 -04:00
def self.excerpt(html, max_length, options={})
ExcerptParser.get_excerpt(html, max_length, options)
2013-02-05 14:16:51 -05:00
end
end