discourse/script/import_scripts/mbox/importer.rb

217 lines
5.4 KiB
Ruby

# frozen_string_literal: true
require_relative "../base"
require_relative "support/database"
require_relative "support/indexer"
require_relative "support/settings"
module ImportScripts::Mbox
class Importer < ImportScripts::Base
def initialize(settings_filename)
@settings = Settings.load(settings_filename)
super()
@database = Database.new(@settings.data_dir, @settings.batch_size)
end
def get_site_settings_for_import
settings = super
settings[:enable_staged_users] = true
settings[:incoming_email_prefer_html] = @settings.prefer_html
settings
end
protected
def execute
index_messages
if @settings.index_only
@skip_updates = true
else
SiteSetting.tagging_enabled = true if @settings.tags.present?
import_categories
import_users
import_posts
end
end
def index_messages
puts "", "creating index"
indexer = Indexer.new(@database, @settings)
indexer.execute
end
def import_categories
puts "", "creating categories"
rows = @database.fetch_categories
create_categories(rows) do |row|
{
id: row["name"],
name: row["name"],
parent_category_id: row["parent_category_id"].presence,
}
end
end
def import_users
puts "", "creating users"
total_count = @database.count_users
last_email = ""
batches do |offset|
rows, last_email = @database.fetch_users(last_email)
break if rows.empty?
next if all_records_exist?(:users, rows.map { |row| row["email"] })
create_users(rows, total: total_count, offset: offset) do |row|
{
id: row["email"],
email: row["email"],
name: row["name"],
trust_level: @settings.trust_level,
staged: @settings.staged,
active: !@settings.staged,
created_at: to_time(row["date_of_first_message"]),
}
end
end
end
def batches
super(@settings.batch_size)
end
def import_posts
puts "", "creating topics and posts"
total_count = @database.count_messages
last_row_id = 0
batches do |offset|
rows, last_row_id = @database.fetch_messages(last_row_id)
break if rows.empty?
next if all_records_exist?(:posts, rows.map { |row| row["msg_id"] })
create_posts(rows, total: total_count, offset: offset) do |row|
begin
if row["email_date"].blank?
puts "Date is missing. Skipping #{row["msg_id"]}"
nil
elsif row["in_reply_to"].blank?
map_first_post(row)
else
map_reply(row)
end
rescue => e
puts "Failed to map post for #{row["msg_id"]}", e, e.backtrace.join("\n")
end
end
end
end
def map_post(row)
user_id = user_id_from_imported_user_id(row["from_email"]) || Discourse::SYSTEM_USER_ID
{
id: row["msg_id"],
user_id: user_id,
created_at: to_time(row["email_date"]),
raw: format_raw(row, user_id),
raw_email: row["raw_message"],
via_email: true,
post_create_action: proc { |post| create_incoming_email(post, row) },
}
end
def format_raw(row, user_id)
body = row["body"] || ""
elided = row["elided"]
if row["attachment_count"].positive?
receiver = Email::Receiver.new(row["raw_message"])
user = User.find(user_id)
body = receiver.add_attachments(body, user)
end
if elided.present? && @settings.show_trimmed_content
body = "#{body}#{Email::Receiver.elided_html(elided)}"
end
body
end
def map_first_post(row)
subject = row["subject"]
tags = remove_tags!(subject)
mapped = map_post(row)
mapped[:category] = category_id_from_imported_category_id(row["category"])
mapped[:title] = subject.strip[0...255]
mapped[:tags] = tags if tags.present?
mapped
end
def map_reply(row)
parent = @lookup.topic_lookup_from_imported_post_id(row["in_reply_to"])
if parent.blank?
puts "Parent message #{row["in_reply_to"]} doesn't exist. Skipping #{row["msg_id"]}: #{row["subject"][0..40]}"
return nil
end
mapped = map_post(row)
mapped[:topic_id] = parent[:topic_id]
mapped
end
def remove_tags!(subject)
tag_names = []
remove_prefixes!(subject)
loop do
old_length = subject.length
@settings.tags.each do |tag|
tag_names << tag[:name] if subject.sub!(tag[:regex], "") && tag[:name].present?
end
remove_prefixes!(subject) if subject.length != old_length
break if subject.length == old_length
end
tag_names.uniq
end
def remove_prefixes!(subject)
# There could be multiple prefixes...
loop do
if subject.sub!(@settings.subject_prefix_regex, "")
subject.strip!
else
break
end
end
end
def create_incoming_email(post, row)
IncomingEmail.create(
message_id: row["msg_id"],
raw: row["raw_message"],
subject: row["subject"],
from_address: row["from_email"],
user_id: post.user_id,
topic_id: post.topic_id,
post_id: post.id,
)
end
def to_time(timestamp)
Time.zone.at(timestamp) if timestamp
end
end
end