discourse/lib/cooked_post_processor.rb

196 lines
5.4 KiB
Ruby
Raw Normal View History

2013-02-06 03:16:51 +08:00
# Post processing that we can do after a post has already been cooked. For
# example, inserting the onebox content, or image sizes.
require_dependency 'oneboxer'
class CookedPostProcessor
2013-06-22 00:29:40 +08:00
include ActionView::Helpers::NumberHelper
2013-02-19 14:57:14 +08:00
2013-02-06 03:16:51 +08:00
def initialize(post, opts={})
@dirty = false
@opts = opts
@post = post
@doc = Nokogiri::HTML::fragment(post.cooked)
2013-02-19 14:57:14 +08:00
@size_cache = {}
@has_been_uploaded_cache = {}
2013-02-06 03:16:51 +08:00
end
def post_process
return unless @doc.present?
post_process_images
post_process_oneboxes
2013-02-06 03:16:51 +08:00
end
2013-02-26 00:42:20 +08:00
def post_process_images
2013-06-24 08:10:21 +08:00
images = @doc.css("img") - @doc.css(".onebox-result img")
2013-02-26 00:42:20 +08:00
return unless images.present?
2013-02-06 03:16:51 +08:00
2013-02-19 14:57:14 +08:00
images.each do |img|
# keep track of the original src
2013-02-19 14:57:14 +08:00
src = img['src']
# make sure the src is absolute (when working with locally uploaded files)
img['src'] = Discourse.base_url_no_prefix + img['src'] if img['src'] =~ /^\/[^\/]/
2013-02-06 03:16:51 +08:00
2013-04-13 22:31:20 +08:00
if src.present?
# make sure the img has both width and height attributes
update_dimensions!(img)
2013-06-17 07:00:25 +08:00
# retrieve the associated upload, if any
upload = get_upload_from_url(img['src'])
if upload.present?
2013-06-18 04:46:48 +08:00
# update reverse index
associate_to_post upload
2013-06-17 07:00:25 +08:00
# create a thumbnail
upload.create_thumbnail!
# optimize image
img['src'] = optimize_image(img)
# lightbox treatment
2013-06-22 00:29:40 +08:00
convert_to_link!(img, upload)
2013-06-17 07:00:25 +08:00
else
convert_to_link!(img)
end
# mark the post as dirty whenever the src has changed
@dirty |= src != img['src']
2013-02-06 03:16:51 +08:00
end
2013-04-13 22:31:20 +08:00
end
2013-02-19 14:57:14 +08:00
2013-04-13 22:31:20 +08:00
# Extract the first image from the first post and use it as the 'topic image'
if @post.post_number == 1
img = images.first
@post.topic.update_column :image_url, img['src'] if img['src'].present?
2013-02-19 14:57:14 +08:00
end
2013-04-13 22:31:20 +08:00
2013-02-19 14:57:14 +08:00
end
def post_process_oneboxes
args = { post_id: @post.id }
args[:invalidate_oneboxes] = true if @opts[:invalidate_oneboxes]
# bake onebox content into the post
result = Oneboxer.apply(@doc) do |url, element|
Oneboxer.onebox(url, args)
end
# mark the post as dirty whenever a onebox as been baked
@dirty |= result.changed?
end
2013-04-13 22:31:20 +08:00
def update_dimensions!(img)
return if img['width'].present? && img['height'].present?
w, h = get_size_from_image_sizes(img['src'], @opts[:image_sizes]) || image_dimensions(img['src'])
if w && h
img['width'] = w.to_s
img['height'] = h.to_s
@dirty = true
end
end
2013-02-21 09:07:36 +08:00
2013-06-17 07:00:25 +08:00
def get_upload_from_url(url)
if Upload.has_been_uploaded?(url)
if m = LocalStore.uploaded_regex.match(url)
Upload.where(id: m[:upload_id]).first
elsif Upload.is_on_s3?(url)
Upload.where(url: url).first
end
2013-06-17 07:00:25 +08:00
end
end
2013-06-18 04:46:48 +08:00
def associate_to_post(upload)
return if PostUpload.where(post_id: @post.id, upload_id: upload.id).count > 0
PostUpload.create({ post_id: @post.id, upload_id: upload.id })
rescue ActiveRecord::RecordNotUnique
# do not care if it's already associated
end
def optimize_image(img)
return img["src"]
# 1) optimize using image_optim
# 2) .png vs. .jpg
2013-02-19 14:57:14 +08:00
end
2013-06-22 00:29:40 +08:00
def convert_to_link!(img, upload=nil)
2013-02-19 14:57:14 +08:00
src = img["src"]
2013-04-13 22:31:20 +08:00
width, height = img["width"].to_i, img["height"].to_i
2013-02-19 14:57:14 +08:00
return unless src.present? && width > SiteSetting.auto_link_images_wider_than
2013-04-13 22:31:20 +08:00
original_width, original_height = get_size(src)
2013-02-19 14:57:14 +08:00
return unless original_width.to_i > width && original_height.to_i > height
parent = img.parent
while parent
return if parent.name == "a"
break unless parent.respond_to? :parent
parent = parent.parent
end
2013-02-26 00:42:20 +08:00
# not a hyperlink so we can apply
img['src'] = upload.thumbnail_url if (upload && upload.thumbnail_url.present?)
2013-02-19 14:57:14 +08:00
a = Nokogiri::XML::Node.new "a", @doc
img.add_next_sibling(a)
a["href"] = src
a["class"] = "lightbox"
a.add_child(img)
2013-06-22 00:29:40 +08:00
# some overlay informations
filename = upload ? upload.original_filename : File.basename(src)
informations = "#{original_width}x#{original_height}"
informations << " | #{number_to_human_size(upload.filesize)}" if upload
a.add_child create_span_node("filename", filename)
a.add_child create_span_node("informations", informations)
a.add_child create_span_node("expand")
# TODO: download
# TODO: views-count
2013-02-19 14:57:14 +08:00
@dirty = true
2013-06-22 00:29:40 +08:00
end
2013-02-19 14:57:14 +08:00
2013-06-22 00:29:40 +08:00
def create_span_node(klass, content=nil)
span = Nokogiri::XML::Node.new "span", @doc
span.content = content if content
span['class'] = klass
span
2013-02-06 03:16:51 +08:00
end
2013-02-19 14:57:14 +08:00
def get_size_from_image_sizes(src, image_sizes)
if image_sizes.present?
if dim = image_sizes[src]
ImageSizer.resize(dim['width'], dim['height'])
end
end
end
2013-02-06 03:16:51 +08:00
# Retrieve the image dimensions for a url
def image_dimensions(url)
w, h = get_size(url)
ImageSizer.resize(w, h) if w && h
2013-02-06 03:16:51 +08:00
end
def get_size(url)
# make sure s3 urls have a scheme (otherwise, FastImage will fail)
url = "http:" + url if Upload.is_on_s3? (url)
return unless is_valid_image_uri? url
# we can *always* crawl our own images
return unless SiteSetting.crawl_images? || Upload.has_been_uploaded?(url)
@size_cache[url] ||= FastImage.size(url)
rescue Zlib::BufError # FastImage.size raises BufError for some gifs
2013-02-06 03:16:51 +08:00
end
def is_valid_image_uri?(url)
uri = URI.parse(url)
%w(http https).include? uri.scheme
rescue URI::InvalidURIError
end
def dirty?
@dirty
end
def html
@doc.try(:to_html)
2013-02-06 03:16:51 +08:00
end
end