discourse/script/import_scripts/base.rb

646 lines
20 KiB
Ruby
Raw Normal View History

if ARGV.include?('bbcode-to-md')
# Replace (most) bbcode with markdown before creating posts.
# This will dramatically clean up the final posts in Discourse.
#
# In a temp dir:
#
2014-07-12 01:36:05 +08:00
# git clone https://github.com/nlalonde/ruby-bbcode-to-md.git
# cd ruby-bbcode-to-md
# gem build ruby-bbcode-to-md.gemspec
# gem install ruby-bbcode-to-md-0.0.13.gem
require 'ruby-bbcode-to-md'
end
require_relative '../../config/environment'
2015-01-24 01:19:46 +08:00
require_dependency 'url_helper'
require_dependency 'file_helper'
module ImportScripts; end
class ImportScripts::Base
2015-01-24 01:19:46 +08:00
include ActionView::Helpers::NumberHelper
def initialize
preload_i18n
@bbcode_to_md = true if ARGV.include?('bbcode-to-md')
2014-07-17 01:59:30 +08:00
@existing_groups = {}
@failed_groups = []
@existing_users = {}
@failed_users = []
@categories_lookup = {}
@existing_posts = {}
@topic_lookup = {}
@site_settings_during_import
@old_site_settings = {}
@start_time = Time.now
puts "loading existing groups..."
2014-07-17 01:59:30 +08:00
GroupCustomField.where(name: 'import_id').pluck(:group_id, :value).each do |group_id, import_id|
@existing_groups[import_id] = group_id
end
puts "loading existing users..."
UserCustomField.where(name: 'import_id').pluck(:user_id, :value).each do |user_id, import_id|
@existing_users[import_id] = user_id
end
puts "loading existing categories..."
CategoryCustomField.where(name: 'import_id').pluck(:category_id, :value).each do |category_id, import_id|
@categories_lookup[import_id] = category_id
end
puts "loading existing posts..."
PostCustomField.where(name: 'import_id').pluck(:post_id, :value).each do |post_id, import_id|
@existing_posts[import_id] = post_id
end
puts "loading existing topics..."
Post.joins(:topic).pluck("posts.id, posts.topic_id, posts.post_number, topics.slug").each do |p|
@topic_lookup[p[0]] = {
topic_id: p[1],
post_number: p[2],
url: Post.url(p[3], p[1], p[2]),
2014-08-18 19:04:08 +08:00
}
end
end
def preload_i18n
I18n.t("test")
ActiveSupport::Inflector.transliterate("test")
end
def perform
Rails.logger.level = 3 # :error, so that we don't create log files that are many GB
change_site_settings
execute
puts ""
update_bumped_at
update_last_posted_at
update_last_seen_at
update_feature_topic_users
update_category_featured_topics
update_topic_count_replies
reset_topic_counters
elapsed = Time.now - @start_time
puts '', "Done (#{elapsed.to_s} seconds)"
ensure
reset_site_settings
end
def change_site_settings
@site_settings_during_import = {
email_domains_blacklist: '',
min_topic_title_length: 1,
min_post_length: 1,
min_private_message_post_length: 1,
min_private_message_title_length: 1,
allow_duplicate_topic_titles: true,
disable_emails: true,
authorized_extensions: '*'
}
@site_settings_during_import.each do |key, value|
@old_site_settings[key] = SiteSetting.send(key)
SiteSetting.set(key, value)
end
RateLimiter.disable
end
def reset_site_settings
@old_site_settings.each do |key, value|
current_value = SiteSetting.send(key)
SiteSetting.set(key, value) unless current_value != @site_settings_during_import[key]
end
RateLimiter.enable
end
# Implementation will do most of its work in its execute method.
# It will need to call create_users, create_categories, and create_posts.
def execute
raise NotImplementedError
end
# Get the Discourse Post id based on the id of the source record
def post_id_from_imported_post_id(import_id)
@existing_posts[import_id] || @existing_posts[import_id.to_s]
end
# Get the Discourse topic info (a hash) based on the id of the source record
def topic_lookup_from_imported_post_id(import_id)
post_id = post_id_from_imported_post_id(import_id)
post_id ? @topic_lookup[post_id] : nil
end
2014-07-17 01:59:30 +08:00
# Get the Discourse Group id based on the id of the source group
def group_id_from_imported_group_id(import_id)
@existing_groups[import_id] || @existing_groups[import_id.to_s] || find_group_by_import_id(import_id).try(:id)
end
def find_group_by_import_id(import_id)
GroupCustomField.where(name: 'import_id', value: import_id.to_s).first.try(:group)
end
# Get the Discourse User id based on the id of the source user
def user_id_from_imported_user_id(import_id)
2014-07-05 04:05:15 +08:00
@existing_users[import_id] || @existing_users[import_id.to_s] || find_user_by_import_id(import_id).try(:id)
end
def find_user_by_import_id(import_id)
UserCustomField.where(name: 'import_id', value: import_id.to_s).first.try(:user)
end
# Get the Discourse Category id based on the id of the source category
def category_id_from_imported_category_id(import_id)
@categories_lookup[import_id] || @categories_lookup[import_id.to_s]
end
def create_admin(opts={})
admin = User.new
admin.email = opts[:email] || "sam.saffron@gmail.com"
admin.username = opts[:username] || "sam"
admin.password = SecureRandom.uuid
admin.save!
admin.grant_admin!
admin.change_trust_level!(TrustLevel[4])
admin.email_tokens.update_all(confirmed: true)
admin
end
2014-07-17 01:59:30 +08:00
# Iterate through a list of groups to be imported.
# Takes a collection and yields to the block for each element.
# Block should return a hash with the attributes for each element.
# Required fields are :id and :name, where :id is the id of the
# group in the original datasource. The given id will not be used
# to create the Discourse group record.
def create_groups(results, opts={})
groups_created = 0
groups_skipped = 0
total = opts[:total] || results.size
results.each do |result|
g = yield(result)
if group_id_from_imported_group_id(g[:id])
groups_skipped += 1
else
new_group = create_group(g, g[:id])
if new_group.valid?
@existing_groups[g[:id].to_s] = new_group.id
groups_created += 1
else
@failed_groups << g
puts "Failed to create group id #{g[:id]} #{new_group.name}: #{new_group.errors.full_messages}"
end
end
print_status groups_created + groups_skipped + @failed_groups.length + (opts[:offset] || 0), total
end
return [groups_created, groups_skipped]
end
def create_group(opts, import_id)
opts = opts.dup.tap {|o| o.delete(:id) }
import_name = opts[:name]
opts[:name] = UserNameSuggester.suggest(import_name)
existing = Group.where(name: opts[:name]).first
return existing if existing and existing.custom_fields["import_id"].to_i == import_id.to_i
g = existing || Group.new(opts)
g.custom_fields["import_id"] = import_id
g.custom_fields["import_name"] = import_name
g.tap(&:save)
end
# Iterate through a list of user records to be imported.
# Takes a collection, and yields to the block for each element.
# Block should return a hash with the attributes for the User model.
# Required fields are :id and :email, where :id is the id of the
# user in the original datasource. The given id will not be used to
# create the Discourse user record.
2014-07-05 04:05:15 +08:00
def create_users(results, opts={})
users_created = 0
users_skipped = 0
2014-07-05 04:05:15 +08:00
total = opts[:total] || results.size
results.each do |result|
u = yield(result)
# block returns nil to skip a user
if u.nil?
users_skipped += 1
else
import_id = u[:id]
2014-08-18 19:04:08 +08:00
if user_id_from_imported_user_id(import_id)
users_skipped += 1
elsif u[:email].present?
new_user = create_user(u, import_id)
if new_user.valid?
@existing_users[import_id.to_s] = new_user.id
users_created += 1
else
@failed_users << u
puts "Failed to create user id: #{import_id}, username: #{new_user.username}, email: #{new_user.email}: #{new_user.errors.full_messages}"
end
else
@failed_users << u
puts "Skipping user id #{import_id} because email is blank"
end
end
2014-07-05 04:05:15 +08:00
print_status users_created + users_skipped + @failed_users.length + (opts[:offset] || 0), total
end
2014-07-05 04:05:15 +08:00
return [users_created, users_skipped]
end
def create_user(opts, import_id)
opts.delete(:id)
merge = opts.delete(:merge)
2014-07-17 01:59:30 +08:00
post_create_action = opts.delete(:post_create_action)
existing = User.where(email: opts[:email].downcase, username: opts[:username]).first
return existing if existing && (merge || existing.custom_fields["import_id"].to_i == import_id.to_i)
bio_raw = opts.delete(:bio_raw)
2014-08-18 19:04:08 +08:00
website = opts.delete(:website)
2015-01-24 01:19:46 +08:00
location = opts.delete(:location)
avatar_url = opts.delete(:avatar_url)
opts[:name] = User.suggest_name(opts[:email]) unless opts[:name]
if opts[:username].blank? ||
opts[:username].length < User.username_length.begin ||
opts[:username].length > User.username_length.end ||
opts[:username] =~ /[^A-Za-z0-9_]/ ||
opts[:username][0] =~ /[^A-Za-z0-9]/ ||
!User.username_available?(opts[:username])
opts[:username] = UserNameSuggester.suggest(opts[:username] || opts[:name] || opts[:email])
end
opts[:email] = opts[:email].downcase
opts[:trust_level] = TrustLevel[1] unless opts[:trust_level]
opts[:active] = opts.fetch(:active, true)
opts[:import_mode] = true
u = User.new(opts)
u.custom_fields["import_id"] = import_id
u.custom_fields["import_username"] = opts[:username] if opts[:username].present?
u.custom_fields["import_avatar_url"] = avatar_url if avatar_url.present?
begin
User.transaction do
u.save!
if bio_raw.present? || website.present? || location.present?
2014-08-18 19:04:08 +08:00
u.user_profile.bio_raw = bio_raw if bio_raw.present?
u.user_profile.website = website if website.present?
2015-01-24 01:19:46 +08:00
u.user_profile.location = location if location.present?
u.user_profile.save!
end
end
rescue
# try based on email
existing = User.find_by(email: opts[:email].downcase)
if existing
existing.custom_fields["import_id"] = import_id
existing.save!
u = existing
end
end
2014-07-17 01:59:30 +08:00
post_create_action.try(:call, u) if u.persisted?
u # If there was an error creating the user, u.errors has the messages
end
# Iterates through a collection to create categories.
# The block should return a hash with attributes for the new category.
# Required fields are :id and :name, where :id is the id of the
# category in the original datasource. The given id will not be used to
# create the Discourse category record.
# Optional attributes are position, description, and parent_category_id.
def create_categories(results)
results.each do |c|
params = yield(c)
2014-09-11 02:27:18 +08:00
# block returns nil to skip
next if params.nil? || category_id_from_imported_category_id(params[:id])
2015-01-24 01:19:46 +08:00
2014-09-11 02:27:18 +08:00
# Basic massaging on the category name
params[:name] = "Blank" if params[:name].blank?
params[:name].strip!
params[:name] = params[:name][0..49]
2014-08-18 19:04:08 +08:00
puts "\t#{params[:name]}"
2014-07-05 04:05:15 +08:00
# make sure categories don't go more than 2 levels deep
if params[:parent_category_id]
top = Category.find_by_id(params[:parent_category_id])
top = top.parent_category while top && !top.parent_category.nil?
params[:parent_category_id] = top.id if top
end
new_category = create_category(params, params[:id])
@categories_lookup[params[:id]] = new_category.id
end
end
def create_category(opts, import_id)
existing = Category.where("LOWER(name) = ?", opts[:name].downcase).first
return existing if existing && existing.parent_category.try(:id) == opts[:parent_category_id]
2014-07-17 01:59:30 +08:00
post_create_action = opts.delete(:post_create_action)
2014-08-18 19:04:08 +08:00
new_category = Category.new(
name: opts[:name],
user_id: opts[:user_id] || opts[:user].try(:id) || -1,
position: opts[:position],
description: opts[:description],
parent_category_id: opts[:parent_category_id]
)
2014-08-18 19:04:08 +08:00
new_category.custom_fields["import_id"] = import_id if import_id
new_category.save!
2014-08-18 19:04:08 +08:00
2014-07-17 01:59:30 +08:00
post_create_action.try(:call, new_category)
2014-08-18 19:04:08 +08:00
new_category
end
def created_post(post)
# override if needed
end
# Iterates through a collection of posts to be imported.
# It can create topics and replies.
# Attributes will be passed to the PostCreator.
# Topics should give attributes title and category.
# Replies should provide topic_id. Use topic_lookup_from_imported_post_id to find the topic.
def create_posts(results, opts={})
skipped = 0
created = 0
total = opts[:total] || results.size
results.each do |r|
params = yield(r)
# block returns nil to skip a post
if params.nil?
skipped += 1
else
import_id = params.delete(:id).to_s
if post_id_from_imported_post_id(import_id)
skipped += 1 # already imported this post
else
begin
new_post = create_post(params, import_id)
if new_post.is_a?(Post)
@existing_posts[import_id] = new_post.id
@topic_lookup[new_post.id] = {
post_number: new_post.post_number,
topic_id: new_post.topic_id,
url: new_post.url,
}
created_post(new_post)
created += 1
else
skipped += 1
puts "Error creating post #{import_id}. Skipping."
puts new_post.inspect
end
rescue Discourse::InvalidAccess => e
skipped += 1
puts "InvalidAccess creating post #{import_id}. Topic is closed? #{e.message}"
rescue => e
2014-07-05 04:05:15 +08:00
skipped += 1
puts "Exception while creating post #{import_id}. Skipping."
puts e.message
puts e.backtrace.join("\n")
2014-07-05 04:05:15 +08:00
end
end
end
print_status skipped + created + (opts[:offset] || 0), total
end
return [created, skipped]
end
2014-06-26 07:11:52 +08:00
def create_post(opts, import_id)
user = User.find(opts[:user_id])
2014-07-17 01:59:30 +08:00
post_create_action = opts.delete(:post_create_action)
opts = opts.merge(skip_validations: true)
opts[:import_mode] = true
2014-06-26 07:11:52 +08:00
opts[:custom_fields] ||= {}
opts[:custom_fields]['import_id'] = import_id
if @bbcode_to_md
opts[:raw] = opts[:raw].bbcode_to_md(false) rescue opts[:raw]
end
2014-07-05 04:05:15 +08:00
post_creator = PostCreator.new(user, opts)
post = post_creator.create
2014-07-17 01:59:30 +08:00
post_create_action.try(:call, post) if post
2014-07-05 04:05:15 +08:00
post ? post : post_creator.errors.full_messages
end
2014-07-17 01:59:30 +08:00
# Creates an upload.
# Expects path to be the full path and filename of the source file.
def create_upload(user_id, path, source_filename)
tmp = Tempfile.new('discourse-upload')
src = File.open(path)
FileUtils.copy_stream(src, tmp)
src.close
tmp.rewind
2015-02-04 01:44:18 +08:00
Upload.create_for(user_id, tmp, source_filename, tmp.size)
2014-07-17 01:59:30 +08:00
ensure
tmp.close rescue nil
tmp.unlink rescue nil
end
# Iterate through a list of bookmark records to be imported.
# Takes a collection, and yields to the block for each element.
# Block should return a hash with the attributes for the bookmark.
# Required fields are :user_id and :post_id, where both ids are
# the values in the original datasource.
def create_bookmarks(results, opts={})
bookmarks_created = 0
bookmarks_skipped = 0
total = opts[:total] || results.size
user = User.new
post = Post.new
results.each do |result|
params = yield(result)
# only the IDs are needed, so this should be enough
user.id = user_id_from_imported_user_id(params[:user_id])
post.id = post_id_from_imported_post_id(params[:post_id])
if user.id.nil? || post.id.nil?
bookmarks_skipped += 1
puts "Skipping bookmark for user id #{params[:user_id]} and post id #{params[:post_id]}"
else
begin
PostAction.act(user, post, PostActionType.types[:bookmark])
bookmarks_created += 1
rescue PostAction::AlreadyActed
bookmarks_skipped += 1
end
print_status bookmarks_created + bookmarks_skipped + (opts[:offset] || 0), total
end
end
end
def close_inactive_topics(opts={})
num_days = opts[:days] || 30
puts '', "Closing topics that have been inactive for more than #{num_days} days."
query = Topic.where('last_posted_at < ?', num_days.days.ago).where(closed: false)
total_count = query.count
closed_count = 0
query.find_each do |topic|
topic.update_status('closed', true, Discourse.system_user)
closed_count += 1
print_status(closed_count, total_count)
end
end
def update_bumped_at
puts "", "updating bumped_at on topics"
Post.exec_sql("update topics t set bumped_at = COALESCE((select max(created_at) from posts where topic_id = t.id and post_type != #{Post.types[:moderator_action]}), bumped_at)")
end
def update_last_posted_at
puts "", "updating last posted at on users"
sql = <<-SQL
WITH lpa AS (
SELECT user_id, MAX(posts.created_at) AS last_posted_at
FROM posts
GROUP BY user_id
)
UPDATE users
SET last_posted_at = lpa.last_posted_at
FROM users u1
JOIN lpa ON lpa.user_id = u1.id
WHERE u1.id = users.id
AND users.last_posted_at <> lpa.last_posted_at
SQL
User.exec_sql(sql)
end
# scripts that are able to import last_seen_at from the source data should override this method
def update_last_seen_at
puts "", "updating last seen at on users"
User.exec_sql("UPDATE users SET last_seen_at = created_at WHERE last_seen_at IS NULL")
User.exec_sql("UPDATE users SET last_seen_at = last_posted_at WHERE last_posted_at IS NOT NULL")
end
def update_feature_topic_users
puts "", "updating featured topic users"
total_count = Topic.count
progress_count = 0
Topic.find_each do |topic|
topic.feature_topic_users
progress_count += 1
print_status(progress_count, total_count)
end
end
def reset_topic_counters
puts "", "resetting topic counters"
total_count = Topic.count
progress_count = 0
Topic.find_each do |topic|
Topic.reset_highest(topic.id)
progress_count += 1
print_status(progress_count, total_count)
end
end
def update_category_featured_topics
puts "", "updating featured topics in categories"
total_count = Category.count
progress_count = 0
Category.find_each do |category|
CategoryFeaturedTopic.feature_topics_for(category)
progress_count += 1
print_status(progress_count, total_count)
end
end
def update_topic_count_replies
puts "", "updating user topic reply counts"
total_count = User.real.count
progress_count = 0
User.real.find_each do |u|
u.user_stat.update_topic_reply_count
u.user_stat.save!
progress_count += 1
print_status(progress_count, total_count)
end
end
def update_tl0
User.all.each do |user|
user.change_trust_level!(0) if Post.where(user_id: user.id).count == 0
end
end
2015-01-24 01:19:46 +08:00
def html_for_upload(upload, display_filename)
if FileHelper.is_image?(upload.url)
embedded_image_html(upload)
else
attachment_html(upload, display_filename)
end
end
def embedded_image_html(upload)
%Q[<img src="#{upload.url}" width="#{[upload.width, 640].compact.min}" height="#{[upload.height,480].compact.min}"><br/>]
end
def attachment_html(upload, display_filename)
"<a class='attachment' href='#{upload.url}'>#{display_filename}</a> (#{number_to_human_size(upload.filesize)})"
end
def print_status(current, max)
print "\r%9d / %d (%5.1f%%) " % [current, max, ((current.to_f / max.to_f) * 100).round(1)]
end
def batches(batch_size)
offset = 0
loop do
yield offset
offset += batch_size
end
end
end