]> Untitled Git - lemmy.git/blobdiff - crates/apub/src/objects/comment.rs
Replace TypedBuilder with Default in update forms (#3814)
[lemmy.git] / crates / apub / src / objects / comment.rs
index ae16b76fe0f85bf188234e9e7abe7cb52c64e64d..1b7724431200302ee0c153988a21649f366be1fc 100644 (file)
 use crate::{
-  extensions::context::lemmy_context,
-  fetcher::objects::{get_or_fetch_and_insert_comment, get_or_fetch_and_insert_post},
-  get_community_from_to_or_cc,
-  objects::{
-    check_object_domain,
-    check_object_for_community_or_site_ban,
-    create_tombstone,
-    get_object_from_apub,
-    get_or_fetch_and_upsert_person,
-    get_source_markdown_value,
-    set_content_and_source,
-    FromApub,
-    FromApubToForm,
-    ToApub,
+  activities::{verify_is_public, verify_person_in_community},
+  check_apub_id_valid_with_strictness,
+  mentions::collect_non_local_mentions,
+  objects::{read_from_string_or_source, verify_is_remote_object},
+  protocol::{
+    objects::{note::Note, LanguageTag},
+    InCommunity,
+    Source,
   },
-  NoteExt,
 };
-use activitystreams::{
-  object::{kind::NoteType, ApObject, Note, Tombstone},
-  prelude::*,
-  public,
+use activitypub_federation::{
+  config::Data,
+  kinds::{object::NoteType, public},
+  protocol::{values::MediaTypeMarkdownOrHtml, verification::verify_domains_match},
+  traits::Object,
+};
+use chrono::NaiveDateTime;
+use lemmy_api_common::{
+  context::LemmyContext,
+  utils::{local_site_opt_to_slur_regex, sanitize_html},
 };
-use anyhow::{anyhow, Context};
-use lemmy_api_common::blocking;
-use lemmy_db_queries::{Crud, DbPool};
 use lemmy_db_schema::{
   source::{
-    comment::{Comment, CommentForm},
+    comment::{Comment, CommentInsertForm, CommentUpdateForm},
     community::Community,
+    local_site::LocalSite,
     person::Person,
     post::Post,
   },
-  CommentId,
+  traits::Crud,
 };
 use lemmy_utils::{
-  location_info,
-  utils::{convert_datetime, remove_slurs},
-  LemmyError,
+  error::{LemmyError, LemmyErrorType},
+  utils::{markdown::markdown_to_html, slurs::remove_slurs, time::convert_datetime},
 };
-use lemmy_websocket::LemmyContext;
+use std::ops::Deref;
 use url::Url;
 
-#[async_trait::async_trait(?Send)]
-impl ToApub for Comment {
-  type ApubType = NoteExt;
-
-  async fn to_apub(&self, pool: &DbPool) -> Result<NoteExt, LemmyError> {
-    let mut comment = ApObject::new(Note::new());
+#[derive(Clone, Debug)]
+pub struct ApubComment(pub(crate) Comment);
 
-    let creator_id = self.creator_id;
-    let creator = blocking(pool, move |conn| Person::read(conn, creator_id)).await??;
+impl Deref for ApubComment {
+  type Target = Comment;
+  fn deref(&self) -> &Self::Target {
+    &self.0
+  }
+}
 
-    let post_id = self.post_id;
-    let post = blocking(pool, move |conn| Post::read(conn, post_id)).await??;
+impl From<Comment> for ApubComment {
+  fn from(c: Comment) -> Self {
+    ApubComment(c)
+  }
+}
 
-    let community_id = post.community_id;
-    let community = blocking(pool, move |conn| Community::read(conn, community_id)).await??;
+#[async_trait::async_trait]
+impl Object for ApubComment {
+  type DataType = LemmyContext;
+  type Kind = Note;
+  type Error = LemmyError;
 
-    // Add a vector containing some important info to the "in_reply_to" field
-    // [post_ap_id, Option(parent_comment_ap_id)]
-    let mut in_reply_to_vec = vec![post.ap_id.into_inner()];
+  fn last_refreshed_at(&self) -> Option<NaiveDateTime> {
+    None
+  }
 
-    if let Some(parent_id) = self.parent_id {
-      let parent_comment = blocking(pool, move |conn| Comment::read(conn, parent_id)).await??;
+  #[tracing::instrument(skip_all)]
+  async fn read_from_id(
+    object_id: Url,
+    context: &Data<Self::DataType>,
+  ) -> Result<Option<Self>, LemmyError> {
+    Ok(
+      Comment::read_from_apub_id(&mut context.pool(), object_id)
+        .await?
+        .map(Into::into),
+    )
+  }
 
-      in_reply_to_vec.push(parent_comment.ap_id.into_inner());
+  #[tracing::instrument(skip_all)]
+  async fn delete(self, context: &Data<Self::DataType>) -> Result<(), LemmyError> {
+    if !self.deleted {
+      let form = CommentUpdateForm {
+        deleted: Some(true),
+        ..Default::default()
+      };
+      Comment::update(&mut context.pool(), self.id, &form).await?;
     }
+    Ok(())
+  }
 
-    comment
-      // Not needed when the Post is embedded in a collection (like for community outbox)
-      .set_many_contexts(lemmy_context()?)
-      .set_id(self.ap_id.to_owned().into_inner())
-      .set_published(convert_datetime(self.published))
-      // NOTE: included community id for compatibility with lemmy v0.9.9
-      .set_many_tos(vec![community.actor_id.into_inner(), public()])
-      .set_many_in_reply_tos(in_reply_to_vec)
-      .set_attributed_to(creator.actor_id.into_inner());
+  #[tracing::instrument(skip_all)]
+  async fn into_json(self, context: &Data<Self::DataType>) -> Result<Note, LemmyError> {
+    let creator_id = self.creator_id;
+    let creator = Person::read(&mut context.pool(), creator_id).await?;
 
-    set_content_and_source(&mut comment, &self.content)?;
+    let post_id = self.post_id;
+    let post = Post::read(&mut context.pool(), post_id).await?;
+    let community_id = post.community_id;
+    let community = Community::read(&mut context.pool(), community_id).await?;
 
-    if let Some(u) = self.updated {
-      comment.set_updated(convert_datetime(u));
-    }
+    let in_reply_to = if let Some(comment_id) = self.parent_comment_id() {
+      let parent_comment = Comment::read(&mut context.pool(), comment_id).await?;
+      parent_comment.ap_id.into()
+    } else {
+      post.ap_id.into()
+    };
+    let language = LanguageTag::new_single(self.language_id, &mut context.pool()).await?;
+    let maa = collect_non_local_mentions(&self, community.actor_id.clone().into(), context).await?;
 
-    Ok(comment)
-  }
+    let note = Note {
+      r#type: NoteType::Note,
+      id: self.ap_id.clone().into(),
+      attributed_to: creator.actor_id.into(),
+      to: vec![public()],
+      cc: maa.ccs,
+      content: markdown_to_html(&self.content),
+      media_type: Some(MediaTypeMarkdownOrHtml::Html),
+      source: Some(Source::new(self.content.clone())),
+      in_reply_to,
+      published: Some(convert_datetime(self.published)),
+      updated: self.updated.map(convert_datetime),
+      tag: maa.tags,
+      distinguished: Some(self.distinguished),
+      language,
+      audience: Some(community.actor_id.into()),
+    };
 
-  fn to_tombstone(&self) -> Result<Tombstone, LemmyError> {
-    create_tombstone(
-      self.deleted,
-      self.ap_id.to_owned().into(),
-      self.updated,
-      NoteType::Note,
-    )
+    Ok(note)
   }
-}
 
-#[async_trait::async_trait(?Send)]
-impl FromApub for Comment {
-  type ApubType = NoteExt;
-
-  /// Converts a `Note` to `Comment`.
-  ///
-  /// If the parent community, post and comment(s) are not known locally, these are also fetched.
-  async fn from_apub(
-    note: &NoteExt,
-    context: &LemmyContext,
-    expected_domain: Url,
-    request_counter: &mut i32,
-    mod_action_allowed: bool,
-  ) -> Result<Comment, LemmyError> {
-    let comment: Comment = get_object_from_apub(
-      note,
-      context,
-      expected_domain,
-      request_counter,
-      mod_action_allowed,
-    )
-    .await?;
+  #[tracing::instrument(skip_all)]
+  async fn verify(
+    note: &Note,
+    expected_domain: &Url,
+    context: &Data<LemmyContext>,
+  ) -> Result<(), LemmyError> {
+    verify_domains_match(note.id.inner(), expected_domain)?;
+    verify_domains_match(note.attributed_to.inner(), note.id.inner())?;
+    verify_is_public(&note.to, &note.cc)?;
+    let community = note.community(context).await?;
 
-    let post_id = comment.post_id;
-    let post = blocking(context.pool(), move |conn| Post::read(conn, post_id)).await??;
-    check_object_for_community_or_site_ban(note, post.community_id, context, request_counter)
-      .await?;
+    check_apub_id_valid_with_strictness(note.id.inner(), community.local, context).await?;
+    verify_is_remote_object(note.id.inner(), context.settings())?;
+    verify_person_in_community(&note.attributed_to, &community, context).await?;
+    let (post, _) = note.get_parents(context).await?;
     if post.locked {
-      // This is not very efficient because a comment gets inserted just to be deleted right
-      // afterwards, but it seems to be the easiest way to implement it.
-      blocking(context.pool(), move |conn| {
-        Comment::delete(conn, comment.id)
-      })
-      .await??;
-      Err(anyhow!("Post is locked").into())
-    } else {
-      Ok(comment)
+      return Err(LemmyErrorType::PostIsLocked)?;
     }
+    Ok(())
   }
-}
 
-#[async_trait::async_trait(?Send)]
-impl FromApubToForm<NoteExt> for CommentForm {
-  async fn from_apub(
-    note: &NoteExt,
-    context: &LemmyContext,
-    expected_domain: Url,
-    request_counter: &mut i32,
-    _mod_action_allowed: bool,
-  ) -> Result<CommentForm, LemmyError> {
-    let community = get_community_from_to_or_cc(note, context, request_counter).await?;
-    let ap_id = Some(check_object_domain(note, expected_domain, community.local)?);
-    let creator_actor_id = &note
-      .attributed_to()
-      .context(location_info!())?
-      .as_single_xsd_any_uri()
-      .context(location_info!())?;
-
-    let creator =
-      get_or_fetch_and_upsert_person(creator_actor_id, context, request_counter).await?;
-
-    let mut in_reply_tos = note
-      .in_reply_to()
-      .as_ref()
-      .context(location_info!())?
-      .as_many()
-      .context(location_info!())?
-      .iter()
-      .map(|i| i.as_xsd_any_uri().context(""));
-    let post_ap_id = in_reply_tos.next().context(location_info!())??;
-
-    // This post, or the parent comment might not yet exist on this server yet, fetch them.
-    let post = Box::pin(get_or_fetch_and_insert_post(
-      &post_ap_id,
-      context,
-      request_counter,
-    ))
-    .await?;
-
-    // The 2nd item, if it exists, is the parent comment apub_id
-    // For deeply nested comments, FromApub automatically gets called recursively
-    let parent_id: Option<CommentId> = match in_reply_tos.next() {
-      Some(parent_comment_uri) => {
-        let parent_comment_ap_id = &parent_comment_uri?;
-        let parent_comment = Box::pin(get_or_fetch_and_insert_comment(
-          &parent_comment_ap_id,
-          context,
-          request_counter,
-        ))
-        .await?;
-
-        Some(parent_comment.id)
-      }
-      None => None,
-    };
+  /// Converts a `Note` to `Comment`.
+  ///
+  /// If the parent community, post and comment(s) are not known locally, these are also fetched.
+  #[tracing::instrument(skip_all)]
+  async fn from_json(note: Note, context: &Data<LemmyContext>) -> Result<ApubComment, LemmyError> {
+    let creator = note.attributed_to.dereference(context).await?;
+    let (post, parent_comment) = note.get_parents(context).await?;
+
+    let content = read_from_string_or_source(&note.content, &note.media_type, &note.source);
 
-    let content = get_source_markdown_value(note)?.context(location_info!())?;
-    let content_slurs_removed = remove_slurs(&content);
+    let local_site = LocalSite::read(&mut context.pool()).await.ok();
+    let slur_regex = &local_site_opt_to_slur_regex(&local_site);
+    let content = remove_slurs(&content, slur_regex);
+    let content = sanitize_html(&content);
+    let language_id =
+      LanguageTag::to_language_id_single(note.language, &mut context.pool()).await?;
 
-    Ok(CommentForm {
+    let form = CommentInsertForm {
       creator_id: creator.id,
       post_id: post.id,
-      parent_id,
-      content: content_slurs_removed,
+      content,
       removed: None,
-      read: None,
-      published: note.published().map(|u| u.to_owned().naive_local()),
-      updated: note.updated().map(|u| u.to_owned().naive_local()),
-      deleted: None,
-      ap_id,
+      published: note.published.map(|u| u.naive_local()),
+      updated: note.updated.map(|u| u.naive_local()),
+      deleted: Some(false),
+      ap_id: Some(note.id.into()),
+      distinguished: note.distinguished,
       local: Some(false),
-    })
+      language_id,
+    };
+    let parent_comment_path = parent_comment.map(|t| t.0.path);
+    let comment = Comment::create(&mut context.pool(), &form, parent_comment_path.as_ref()).await?;
+    Ok(comment.into())
+  }
+}
+
+#[cfg(test)]
+pub(crate) mod tests {
+  #![allow(clippy::unwrap_used)]
+  #![allow(clippy::indexing_slicing)]
+
+  use super::*;
+  use crate::{
+    objects::{
+      community::{tests::parse_lemmy_community, ApubCommunity},
+      instance::ApubSite,
+      person::{tests::parse_lemmy_person, ApubPerson},
+      post::ApubPost,
+      tests::init_context,
+    },
+    protocol::tests::file_to_json_object,
+  };
+  use assert_json_diff::assert_json_include;
+  use html2md::parse_html;
+  use lemmy_db_schema::source::site::Site;
+  use serial_test::serial;
+
+  async fn prepare_comment_test(
+    url: &Url,
+    context: &Data<LemmyContext>,
+  ) -> (ApubPerson, ApubCommunity, ApubPost, ApubSite) {
+    // use separate counter so this doesnt affect tests
+    let context2 = context.reset_request_count();
+    let (person, site) = parse_lemmy_person(&context2).await;
+    let community = parse_lemmy_community(&context2).await;
+    let post_json = file_to_json_object("assets/lemmy/objects/page.json").unwrap();
+    ApubPost::verify(&post_json, url, &context2).await.unwrap();
+    let post = ApubPost::from_json(post_json, &context2).await.unwrap();
+    (person, community, post, site)
+  }
+
+  async fn cleanup(data: (ApubPerson, ApubCommunity, ApubPost, ApubSite), context: &LemmyContext) {
+    Post::delete(&mut context.pool(), data.2.id).await.unwrap();
+    Community::delete(&mut context.pool(), data.1.id)
+      .await
+      .unwrap();
+    Person::delete(&mut context.pool(), data.0.id)
+      .await
+      .unwrap();
+    Site::delete(&mut context.pool(), data.3.id).await.unwrap();
+    LocalSite::delete(&mut context.pool()).await.unwrap();
+  }
+
+  #[tokio::test]
+  #[serial]
+  pub(crate) async fn test_parse_lemmy_comment() {
+    let context = init_context().await;
+    let url = Url::parse("https://enterprise.lemmy.ml/comment/38741").unwrap();
+    let data = prepare_comment_test(&url, &context).await;
+
+    let json: Note = file_to_json_object("assets/lemmy/objects/note.json").unwrap();
+    ApubComment::verify(&json, &url, &context).await.unwrap();
+    let comment = ApubComment::from_json(json.clone(), &context)
+      .await
+      .unwrap();
+
+    assert_eq!(comment.ap_id, url.into());
+    assert_eq!(comment.content.len(), 14);
+    assert!(!comment.local);
+    assert_eq!(context.request_count(), 0);
+
+    let comment_id = comment.id;
+    let to_apub = comment.into_json(&context).await.unwrap();
+    assert_json_include!(actual: json, expected: to_apub);
+
+    Comment::delete(&mut context.pool(), comment_id)
+      .await
+      .unwrap();
+    cleanup(data, &context).await;
+  }
+
+  #[tokio::test]
+  #[serial]
+  async fn test_parse_pleroma_comment() {
+    let context = init_context().await;
+    let url = Url::parse("https://enterprise.lemmy.ml/comment/38741").unwrap();
+    let data = prepare_comment_test(&url, &context).await;
+
+    let pleroma_url =
+      Url::parse("https://queer.hacktivis.me/objects/8d4973f4-53de-49cd-8c27-df160e16a9c2")
+        .unwrap();
+    let person_json = file_to_json_object("assets/pleroma/objects/person.json").unwrap();
+    ApubPerson::verify(&person_json, &pleroma_url, &context)
+      .await
+      .unwrap();
+    ApubPerson::from_json(person_json, &context).await.unwrap();
+    let json = file_to_json_object("assets/pleroma/objects/note.json").unwrap();
+    ApubComment::verify(&json, &pleroma_url, &context)
+      .await
+      .unwrap();
+    let comment = ApubComment::from_json(json, &context).await.unwrap();
+
+    assert_eq!(comment.ap_id, pleroma_url.into());
+    assert_eq!(comment.content.len(), 64);
+    assert!(!comment.local);
+    assert_eq!(context.request_count(), 1);
+
+    Comment::delete(&mut context.pool(), comment.id)
+      .await
+      .unwrap();
+    cleanup(data, &context).await;
+  }
+
+  #[tokio::test]
+  #[serial]
+  async fn test_html_to_markdown_sanitize() {
+    let parsed = parse_html("<script></script><b>hello</b>");
+    assert_eq!(parsed, "**hello**");
   }
 }