From 1c4bc154c9a4953a2c3df6ac9c2414dea4639c3e Mon Sep 17 00:00:00 2001 From: Arpit Jalan Date: Thu, 14 Jul 2016 19:07:25 +0530 Subject: [PATCH] add SimplePress import script --- script/import_scripts/simplepress.rb | 192 +++++++++++++++++++++++++++ 1 file changed, 192 insertions(+) create mode 100644 script/import_scripts/simplepress.rb diff --git a/script/import_scripts/simplepress.rb b/script/import_scripts/simplepress.rb new file mode 100644 index 00000000000..2f4e20ab758 --- /dev/null +++ b/script/import_scripts/simplepress.rb @@ -0,0 +1,192 @@ +require 'mysql2' +require File.expand_path(File.dirname(__FILE__) + "/base.rb") + +class ImportScripts::SimplePress < ImportScripts::Base + + SIMPLE_PRESS_DB ||= ENV['SIMPLEPRESS_DB'] || "simplepress" + TABLE_PREFIX = "wp_sf" + BATCH_SIZE ||= 1000 + + def initialize + super + + @client = Mysql2::Client.new( + host: "localhost", + username: "root", + database: SIMPLE_PRESS_DB, + ) + + SiteSetting.max_username_length = 50 + end + + def execute + import_users + import_categories + import_topics_and_posts + end + + def import_users + puts "", "importing users..." + + last_user_id = -1 + total_users = mysql_query("SELECT COUNT(*) count FROM wp_users WHERE user_email LIKE '%@%'").first["count"] + + batches(BATCH_SIZE) do |offset| + users = mysql_query(<<-SQL + SELECT ID id, user_nicename, display_name, user_email, user_registered, user_url + FROM wp_users + WHERE user_email LIKE '%@%' + AND id > #{last_user_id} + ORDER BY id + LIMIT #{BATCH_SIZE} + SQL + ).to_a + + break if users.empty? + + last_user_id = users[-1]["id"] + user_ids = users.map { |u| u["id"].to_i } + + next if all_records_exist?(:users, user_ids) + + user_ids_sql = user_ids.join(",") + + users_description = {} + mysql_query(<<-SQL + SELECT user_id, meta_value description + FROM wp_usermeta + WHERE user_id IN (#{user_ids_sql}) + AND meta_key = 'description' + SQL + ).each { |um| users_description[um["user_id"]] = um["description"] } + + create_users(users, total: total_users, offset: offset) do |u| + { + id: u["id"].to_i, + username: u["user_nicename"], + email: u["user_email"].downcase, + name: u["display_name"], + created_at: u["user_registered"], + website: u["user_url"], + bio_raw: users_description[u["id"]] + } + end + end + end + + def import_categories + puts "", "importing categories..." + + categories = mysql_query(<<-SQL + SELECT forum_id, forum_name, forum_seq, forum_desc, parent + FROM #{TABLE_PREFIX}forums + ORDER BY forum_id + SQL + ) + + create_categories(categories) do |c| + category = { id: c['forum_id'], name: CGI.unescapeHTML(c['forum_name']), description: CGI.unescapeHTML(c['forum_desc']), position: c['forum_seq'] } + if (parent_id = c['parent'].to_i) > 0 + category[:parent_category_id] = category_id_from_imported_category_id(parent_id) + end + category + end + end + + def import_topics_and_posts + puts "", "creating topics and posts" + + total_count = mysql_query("SELECT count(*) count from #{TABLE_PREFIX}posts").first["count"] + + topic_first_post_id = {} + + batches(BATCH_SIZE) do |offset| + results = mysql_query(" + SELECT p.post_id id, + p.topic_id topic_id, + t.forum_id category_id, + t.topic_name title, + p.post_index post_index, + p.user_id user_id, + p.post_content raw, + p.post_date post_time + FROM #{TABLE_PREFIX}posts p, + #{TABLE_PREFIX}topics t + WHERE p.topic_id = t.topic_id + ORDER BY p.post_date + LIMIT #{BATCH_SIZE} + OFFSET #{offset}; + ") + + break if results.size < 1 + + next if all_records_exist? :posts, results.map {|m| m['id'].to_i} + + create_posts(results, total: total_count, offset: offset) do |m| + skip = false + mapped = {} + + mapped[:id] = m['id'] + mapped[:user_id] = user_id_from_imported_user_id(m['user_id']) || -1 + mapped[:raw] = process_simplepress_post(m['raw'], m['id']) + mapped[:created_at] = Time.zone.at(m['post_time']) + + if m['post_index'] == 1 + mapped[:category] = category_id_from_imported_category_id(m['category_id']) + mapped[:title] = CGI.unescapeHTML(m['title']) + topic_first_post_id[m['topic_id']] = m['id'] + else + parent = topic_lookup_from_imported_post_id(topic_first_post_id[m['topic_id']]) + if parent + mapped[:topic_id] = parent[:topic_id] + else + puts "Parent post #{first_post_id} doesn't exist. Skipping #{m["id"]}: #{m["title"][0..40]}" + skip = true + end + end + + skip ? nil : mapped + end + end + end + + def process_simplepress_post(raw, import_id) + s = raw.dup + + # convert the quote line + s.gsub!(/\[quote='([^']+)'.*?pid='(\d+).*?\]/) { + "[quote=\"#{convert_username($1, import_id)}, " + post_id_to_post_num_and_topic($2, import_id) + '"]' + } + + # :) is encoded as :) + s.gsub!(/(?:.*)/, '\1') + + # Some links look like this: http://www.onegameamonth.com + s.gsub!(/(.+)<\/a>/, '[\2](\1)') + + # Many phpbb bbcode tags have a hash attached to them. Examples: + # [url=https://google.com:1qh1i7ky]click here[/url:1qh1i7ky] + # [quote="cybereality":b0wtlzex]Some text.[/quote:b0wtlzex] + s.gsub!(/:(?:\w{8})\]/, ']') + + # Remove mybb video tags. + s.gsub!(/(^\[video=.*?\])|(\[\/video\]$)/, '') + + s = CGI.unescapeHTML(s) + + # phpBB shortens link text like this, which breaks our markdown processing: + # [http://answers.yahoo.com/question/index ... 223AAkkPli](http://answers.yahoo.com/question/index?qid=20070920134223AAkkPli) + # + # Work around it for now: + s.gsub!(/\[http(s)?:\/\/(www\.)?/, '[') + + s + end + + def mysql_query(sql) + @client.query(sql, cache_rows: false) + end + +end + +ImportScripts::SimplePress.new.perform