discourse/script/import_scripts/vanilla_body_parser.rb
Rachel Carvalho 812e0d6b5e
FIX: improve Vanilla importing (#10478)
* ensure emails don't have spaces
* import banned users as suspended for 1k yrs
* upgrade users to TL2 if they have comments
* topic: import views, closed and pinned info
* import messages
* encode vanilla usernames for permalinks. Vanilla usernames can contain spaces and special characters.
* parse Vanilla's new rich body format
2020-08-24 16:19:57 -04:00

212 lines
5.9 KiB
Ruby

# frozen_string_literal: true
class VanillaBodyParser
def self.configure(lookup:, uploader:, host:, uploads_path:)
@@lookup = lookup
@@uploader = uploader
@@host = host
@@uploads_path = uploads_path
end
def initialize(row, user_id)
@row = row
@user_id = user_id
end
def parse
return clean_up(@row['Body']) unless rich?
full_text = json.each_with_index.map(&method(:parse_fragment)).join('')
normalize full_text
end
private
def clean_up(text)
text.gsub(/<\/?font[^>]*>/, '').gsub(/<\/?span[^>]*>/, '').gsub(/<\/?div[^>]*>/, '').gsub(/^ +/, '').gsub(/ +/, ' ')
end
def rich?
@row['Format'] == 'Rich'
end
def json
return nil unless rich?
@json ||= JSON.parse(@row['Body']).map(&:deep_symbolize_keys)
end
def parse_fragment(fragment, index)
text = fragment.keys.one? && fragment[:insert].is_a?(String) ? fragment[:insert] : rich_parse(fragment)
text = parse_code(text, fragment, index)
text = parse_list(text, fragment, index)
text
end
def rich_parse(fragment)
insert = fragment[:insert]
return parse_mention(insert[:mention]) if insert.respond_to?(:dig) && insert.dig(:mention, :userID)
return parse_formatting(fragment) if fragment[:attributes]
embed_type = insert.dig(:'embed-external', :data, :embedType)
quoting = embed_type == 'quote'
return parse_quote(insert) if quoting
embed = embed_type.in? ['image', 'link', 'file']
parse_embed(insert) if embed
end
def parse_mention(mention)
user = user_from_imported_id(mention[:userID])
username = user&.username || mention[:name]
"@#{username}"
end
def user_from_imported_id(imported_id)
user_id = @@lookup.user_id_from_imported_user_id(imported_id)
User.find(user_id) if user_id
end
def parse_formatting(fragment)
insert = fragment[:insert]
attributes = fragment[:attributes]
text = fragment[:insert]
text = "<a href=\"#{attributes[:link]}\">#{text}</a>" if attributes[:link]
text = "<i>#{text}</i>" if attributes[:italic]
text = "<b>#{text}</b>" if attributes[:bold]
text
end
# In the Quill format used by Vanilla Forums, a line is rendered as `code`
# when it's followed by a fragment with attributes: {'code-block': true}.
# So we open our ``` block when the next fragment has a 'code-block'
# attribute and the previous one didn't and we close the ``` block when
# the second next fragment does not contain the 'code-block' attribute
def parse_code(text, fragment, index)
next_fragment = next_fragment(index)
next_code = next_fragment.dig(:attributes, :'code-block')
if next_code
previous_fragment = previous_fragment(index)
previous_code = previous_fragment.dig(:attributes, :'code-block')
# if next is code and previous is not, prepend ```
text = "\n```#{text}" unless previous_code
end
current_code = fragment.dig(:attributes, :'code-block')
if current_code
second_next_fragment = second_next_fragment(index)
second_next_code = second_next_fragment.dig(:attributes, :'code-block')
# if current is code and 2 after is not, prepend ```
text = "\n```#{text}" unless second_next_code
end
text
end
def parse_list(text, fragment, index)
next_fragment = next_fragment(index)
next_list = next_fragment.dig(:attributes, :list, :type)
if next_list
# if next is list, prepend <li>
text = '<li>' + text
previous_fragment = previous_fragment(index)
previous_list = previous_fragment.dig(:attributes, :list, :type)
# if next is list and previous is not, prepend <ol> or <ul>
list_tag = next_list == 'ordered' ? '<ol>' : '<ul>'
text = "\n#{list_tag}\n#{text}" unless previous_list
end
current_list = fragment.dig(:attributes, :list, :type)
if current_list
# if current is list prepend </li>
tag_closings = '</li>'
second_next_fragment = second_next_fragment(index)
second_next_list = second_next_fragment.dig(:attributes, :list, :type)
# if current is list and 2 after is not, prepend </ol>
list_tag = current_list == 'ordered' ? '</ol>' : '</ul>'
tag_closings = "#{tag_closings}\n#{list_tag}" unless second_next_list
text = tag_closings + text
end
text
end
def next_fragment(index)
json[index + 1] || {}
end
def previous_fragment(index)
json[index - 1] || {}
end
def second_next_fragment(index)
json[index + 2] || {}
end
def parse_quote(insert)
embed = insert.dig(:'embed-external', :data)
import_post_id = "#{embed[:recordType]}##{embed[:recordID]}"
topic = @@lookup.topic_lookup_from_imported_post_id(import_post_id)
user = user_from_imported_id(embed.dig(:insertUser, :userID))
quote_info = topic && user ? "=\"#{user.username}, post: #{topic[:post_number]}, topic: #{topic[:topic_id]}\"" : ''
"[quote#{quote_info}]\n#{embed[:body]}\n[/quote]\n\n"""
end
def parse_embed(insert)
embed = insert.dig(:'embed-external', :data)
url = embed[:url]
if /https?\:\/\/#{@@host}\/uploads\/.*/.match?(url)
remote_path = url.scan(/uploads\/(.*)/)
path = File.join(@@uploads_path, remote_path)
upload = @@uploader.create_upload(@user_id, path, embed[:name])
if upload&.persisted?
return "\n" + @@uploader.html_for_upload(upload, embed[:name]) + "\n"
else
puts "Failed to upload #{path}"
puts upload.errors.full_messages.join(', ') if upload
end
end
"\n[#{embed[:name]}](#{url})\n"
end
def normalize(full_text)
code_matcher = /```(.*\n)+```/
code_block = full_text[code_matcher]
full_text[code_matcher] = '{{{CODE_BLOCK}}}' if code_block
full_text = double_new_lines(full_text)
full_text['{{{CODE_BLOCK}}}'] = code_block if code_block
full_text
end
def double_new_lines(text)
text.split("\n").map(&:strip).map(&:presence).compact.join("\n\n")
end
end