Modify/add site text associated with custom moderation flags?

scharnkn · November 13, 2024, 1:13pm

Hi,

I’ve been experimenting with the new “custom moderation flags” feature and I noticed that creating a new flag leads to missing text key errors in some cases.

For example when using the moderation option “Hide post”, the message send to the user expects the existence of a text key “flag_reasons.mycustomflag”.

Is it possible to add/define the missing keys? As far as I understand, the site text feature only allows to translate already present keys.

Any guidance or hints would be greatly appreciated!

tobiaseigen · December 1, 2024, 5:37pm

We are currently in the process of having newly added strings translated for the upcoming release. Is there a particular problem you’ve experiencedwith the strings in custom flags? I’m not entirely clear based on your question. Would it be possible for you to share a screenshot?What language is it that your forum is in? Does flagging work correctly in US English?

RGJ · December 1, 2024, 5:59pm

Email sent by an US English forum, custom flag “its_hostile”

Moin · December 1, 2024, 6:33pm

I don’t think the problem is related to a specific translation.

When a post is hidden because of flags from the community a personal message is sent to the author. This message contains a reason. For example:

These reasons are used when the message is created.

github.com

discourse/discourse/blob/main/config/locales/server.en.yml#L3239-L3244


      
          flag_reasons:
            off_topic: "Your post was flagged as **off-topic**: the community feels it is not a good fit for the topic, as currently defined by the title and the first post."
            inappropriate: "Your post was flagged as **inappropriate**: the community feels it is offensive, abusive, to be hateful conduct or a violation of [our community guidelines](%{base_path}/guidelines)."
            illegal: "Your post was flagged as  **illegal**: the community thinks it might be breaking the law."
            spam: "Your post was flagged as **spam**: the community feels it is an advertisement, something that is overly promotional in nature instead of being useful or relevant to the topic as expected."
            notify_moderators: "Your post was flagged **for moderator attention**: the community feels something about the post requires manual intervention by a staff member."

The code which adds this reason to the message uses the post_action_type

github.com

discourse/discourse/blob/efa2589204fadd731278e516ed0d6ed76c3e534b/app/models/post.rb#L651-L672


      
          # inform user
          if user.present?
            options = {
              url: url,
              edit_delay: SiteSetting.cooldown_minutes_after_hiding_posts,
              flag_reason:
                I18n.t(
                  "flag_reasons.#{post_action_type_view.types[post_action_type_id]}",
                  locale: SiteSetting.default_locale,
                  base_path: Discourse.base_path,
                ),
            }
          
            message = custom_message
            message = hiding_again ? :post_hidden_again : :post_hidden if message.nil?
          
            Jobs.enqueue_in(
              5.seconds,
              :send_system_message,
              user_id: user.id,

This file has been truncated. show original

That works for the default flag reasons. For example, flag_reasons.spam, as seen in the screenshot above.
The problem is that there is no such string for custom flag reasons. For example, if you create a “Testing” reason and enable Auto hide flagged content.

Then flag_reasons.custom_testing is used. However, this reason does not exist. The language doesn’t matter; it does not even exist in English because it’s a custom flag.

And the same reason is also used when the post is deleted (see screenshot in RGJ’s post)

github.com

discourse/discourse/blob/efa2589204fadd731278e516ed0d6ed76c3e534b/lib/post_destroyer.rb#L407-L417


      
          message_options: {
            flagged_post_raw_content: notify_responders ? options[:parent_post].raw : @post.raw,
            flagged_post_response_raw_content: @post.raw,
            url: notify_responders ? options[:parent_post].url : @post.url,
            flag_reason:
              I18n.t(
                "flag_reasons#{".responder" if notify_responders}.#{flag_type}",
                locale: SiteSetting.default_locale,
                base_path: Discourse.base_path,
              ),
          },

By the way for “responders” the illegal reason is missing too

github.com

discourse/discourse/blob/main/config/locales/server.en.yml#L3245-L3249


      
          responder:
            off_topic: "The post was flagged as **off-topic**: the community feels it is not a good fit for the topic, as currently defined by the title and the first post."
            inappropriate: "The post was flagged as **inappropriate**: the community feels it is offensive, abusive, to be hateful conduct or a violation of [our community guidelines](%{base_path}/guidelines)."
            spam: "The post was flagged as **spam**: the community feels it is an advertisement, something that is overly promotional in nature instead of being useful or relevant to the topic as expected."
            notify_moderators: "The post was flagged **for moderator attention**: the community feels something about the post requires manual intervention by a staff member."

tobiaseigen · December 2, 2024, 12:33am

Thanks @scharnkn for the report and thanks @moin for following up!

Are there other places where the text key is unexpectly showing up instead of the expected text?

scharnkn · December 4, 2024, 11:40am

Hi everyone,

Thank you for providing helpful background information. I apologize for the initial lack of detail in my post.

Are there other places where the text key is unexpectly showing up instead of the expected text?

I haven’t encountered this issue elsewhere.

Best,
Katrin

Moin · December 4, 2024, 12:06pm

I noticed FIX: update flag reason message with default value (#30026) · discourse/discourse@28b4ff6 · GitHub
So I tried again:

Compared to, for example, “Your post was flagged as off-topic: the community feels it is not a good fit for the topic, as currently defined by the title and the first post,” the reason for a custom flag is only a keyword.
Maybe the “Your post was flagged as reason.” could also be used for custom flags.

And in the version for responders: “The post was flagged as reason.”

(I would also find a default text for ‘illegal’ helpful, as this is not a custom reason.)

tobiaseigen · December 4, 2024, 7:32pm

I’m inclined to agree with Moin…

for illegal, which is not a custom flag, do display the contents of topic_flag_types.illegal.description which exists. In US english, it is “This topic requires staff attention because I believe it contains content that is illegal.”
for custom flags, for clarity change reason to “The post was flagged as reason”

is there a technical limitation preventing us from including the reason description in the email?

Moin · December 4, 2024, 7:36pm

It does not. Well, the description does exist, but that is not the same as the reason. There are 3 texts:

The description shown to the user who flags the post.
The reason shown to the author of the post:

github.com

discourse/discourse/blob/main/config/locales/server.en.yml#L3241-L3246


      
          flag_reasons:
            off_topic: "Your post was flagged as **off-topic**: the community feels it is not a good fit for the topic, as currently defined by the title and the first post."
            inappropriate: "Your post was flagged as **inappropriate**: the community feels it is offensive, abusive, to be hateful conduct or a violation of [our community guidelines](%{base_path}/guidelines)."
            illegal: "Your post was flagged as  **illegal**: the community thinks it might be breaking the law."
            spam: "Your post was flagged as **spam**: the community feels it is an advertisement, something that is overly promotional in nature instead of being useful or relevant to the topic as expected."
            notify_moderators: "Your post was flagged **for moderator attention**: the community feels something about the post requires manual intervention by a staff member."

The reason shown to the responder to a deleted post when their post is also deleted and the setting is enabled. That one is missing for “illegal”:

github.com

discourse/discourse/blob/main/config/locales/server.en.yml#L3247-L3251


      
          responder:
            off_topic: "The post was flagged as **off-topic**: the community feels it is not a good fit for the topic, as currently defined by the title and the first post."
            inappropriate: "The post was flagged as **inappropriate**: the community feels it is offensive, abusive, to be hateful conduct or a violation of [our community guidelines](%{base_path}/guidelines)."
            spam: "The post was flagged as **spam**: the community feels it is an advertisement, something that is overly promotional in nature instead of being useful or relevant to the topic as expected."
            notify_moderators: "The post was flagged **for moderator attention**: the community feels something about the post requires manual intervention by a staff member."

tobiaseigen · December 4, 2024, 8:17pm

Oh, interesting. (2) and (3) are nearly identical - I wonder if we are over complicating things by having different texts for these two cases.

For now, looks like we can just create a new string for illegal and make sure it works, eg

responder: 
   illegal: "The post was flagged as  **illegal**: the community thinks it might be breaking the law."

We could also create a new string in (2) and (3) to cover all the custom flags, eg

flag_reasons: 
   custom: "Your post was flagged as **reason**.

responder: 
   custom: "The post was flagged as **reason**.

Topic		Replies	Views
Flag reason/s Support	4	479	November 10, 2022
"Subflags"/anotations as part of flags Feature	7	2312	April 24, 2016
The persistence of posts hidden by flags are a bane to our community UX completed	36	3436	June 8, 2023
Please distinguish flags for "bad" things from organizational and "please help" flags Feature moderation	13	876	October 19, 2023
People are confused when they see "Post hidden by community flags" but it was automatic false-positive spam detection UX	20	3061	November 23, 2017

Modify/add site text associated with custom moderation flags?

Related topics