Fix published date and author for topics created via wp-discourse-embed plugin

Support WordPress

AI-generated summary

The discussion revolves around fixing the published date and author for topics created via the wp-discourse-embed plugin. techAPJ shared a rake task that fetches WordPress data and updates the topic’s author and published date in Discourse. The task uses PostOwnerChanger and PostTimestampChanger to make the necessary changes.

techAPJ noted that the task was not included in lib/task/users.rake because it depends on an external plugin and should be removed after running to avoid git problems. pfaffman confirmed this understanding.

Canapin attempted to use the script to update post dates only but encountered an issue where the dates were not changed despite the script reporting success.

1

I recently installed wp-discourse-embed plugin on Quarter to Three WordPress blog.

As soon as I installed the plugin there was an influx of new topics being created for existing (old) WP posts.

I ran this rake task to fix the published date and author for topics that are linked to old WP posts:

desc "Fetch WP data"
task "users:fetch_wp_data" => [:environment] do

  puts "Fetching WP data"
  i = 0
  TopicEmbed.find_each do |topic_embed|
    begin
      html = open(topic_embed.embed_url).read
      raw_doc = Nokogiri::HTML(html)
    rescue
      next
    end

    topic = topic_embed.topic
    next if topic.nil?

    author_username = raw_doc.at('meta[@name="author"]')
    if author_username.present?
      begin
        author = User.where(username_lower: author_username[:content].strip).first
        if author.present? && author.username_lower != topic.user.username_lower
          PostOwnerChanger.new(post_ids: [topic.first_post.id], topic_id: topic.id, new_owner: User.find_by(username_lower: author.username_lower), acting_user: User.find_by(username_lower: "system"), skip_revision: true).change_owner!
        end
      rescue
        # handle error
      end
    end

    wp_post_date = raw_doc.at('meta[@property="article:published_time"]')
    if wp_post_date.present?
      begin
        timestamp = DateTime.parse(wp_post_date[:content].strip).to_i
        PostTimestampChanger.new(topic_id: topic.id, timestamp: timestamp).change!
      rescue ArgumentError
       # handle invalid date
      end
    end

    putc "."
    i += 1
  end

  puts "", "#{i} posts fetched and corrected!", ""
end

Commands I ran:

cd /var/discourse
./launcher enter app
vim lib/tasks/users.rake
(append the above rake task to `lib/tasks/users.rake` file)
rake users:fetch_wp_data
(remove the appended rake task from `lib/tasks/users.rake` file)
2

This helps me to figure out some more about how rake works and seeing PostTimestampChanger is helpful because it’s taking me a long while to figure out those internals.

My question started out as “why not just add this rake tool to Discourse?”

My answer is that users:fetch_wp_data isn’t included in lib/task/users.rake because it depends on an external plugin. And then I wondered why bother to delete it, but I think that’s because if we don’t remove it when we’re done, we’ll get into git problems when the next pull happens?

3

Yep, your answer is correct! :100:

4

Hi, I tried your script but removed this part:

author_username = raw_doc.at('meta[@name="author"]')
if author_username.present?
  begin
    author = User.where(username_lower: author_username[:content].strip).first
    if author.present? && author.username_lower != topic.user.username_lower
      PostOwnerChanger.new(post_ids: [topic.first_post.id], topic_id: topic.id, new_owner: User.find_by(username_lower: author.username_lower), acting_user: User.find_by(username_lower: "system"), skip_revision: true).change_owner!
    end
  rescue
    # handle error
  end
end

Because I only wanted to update the Discourse posts dates.
The script told me: 61 posts fetched and corrected! but the dates weren’t changed.

Any idea how to make this work?