include url content type when sending post over apub (fixes #2611)

Cargo.lock

@@ -2654,6 +2654,7 @@ dependencies = [
  "lemmy_db_views",
  "lemmy_db_views_actor",
  "lemmy_utils",
+ "mime",
  "moka",
  "once_cell",
  "reqwest",

Cargo.toml

@@ -129,6 +129,7 @@ tokio-postgres = "0.7.10"
 tokio-postgres-rustls = "0.10.0"
 urlencoding = "2.1.3"
 enum-map = "2.7"
+mime = "0.3.17"
 
 [dependencies]
 lemmy_api = { workspace = true }

crates/api_common/Cargo.toml

@@ -34,6 +34,7 @@ full = [
   "futures",
   "once_cell",
   "jsonwebtoken",
+  "mime"
 ]
 
 [dependencies]
@@ -62,6 +63,7 @@ actix-web = { workspace = true, optional = true }
 enum-map = { workspace = true }
 urlencoding = { workspace = true }
 async-trait = { workspace = true }
+mime = { workspace = true, optional = true }
 webpage = { version = "1.6", default-features = false, features = [
   "serde",
 ], optional = true }
@@ -70,7 +72,6 @@ jsonwebtoken = { version = "8.3.0", optional = true }
 # necessary for wasmt compilation
 getrandom = { version = "0.2.10", features = ["js"] }
 task-local-extensions = "0.1.4"
-mime = "0.3.17"
 
 [dev-dependencies]
 serial_test = { workspace = true }

crates/api_common/src/request.rs

@@ -29,6 +29,8 @@ pub fn client_builder(settings: &Settings) -> ClientBuilder {
 }
 
 /// Fetches metadata for the given link and optionally generates thumbnail.
+///
+/// TODO: consider caching the results as it will be called by api then apub send right after each other
 #[tracing::instrument(skip_all)]
 pub async fn fetch_link_metadata(
   url: &Url,
@@ -253,7 +255,7 @@ async fn generate_pictrs_thumbnail(
   }
 }
 
-// TODO: get rid of this
+// TODO: get rid of this by storing content type in db
 #[tracing::instrument(skip_all)]
 async fn is_image_content_type(client: &ClientWithMiddleware, url: &Url) -> Result<(), LemmyError> {
   let response = client.get(url.as_str()).send().await?;

crates/apub/Cargo.toml

@@ -37,10 +37,11 @@ async-trait = { workspace = true }
 anyhow = { workspace = true }
 reqwest = { workspace = true }
 once_cell = { workspace = true }
+serde_with = { workspace = true }
+mime = { workspace = true }
 html2md = "0.2.14"
 html2text = "0.6.0"
 stringreader = "0.1.1"
-serde_with = { workspace = true }
 enum_delegate = "0.2.0"
 moka = { version = "0.11", features = ["future"] }
 

crates/apub/src/objects/post.rs

@@ -112,6 +112,19 @@ impl Object for ApubPost {
     let community = Community::read(&mut context.pool(), community_id).await?;
     let language = LanguageTag::new_single(self.language_id, &mut context.pool()).await?;
 
+    let metadata = match &self.url {
+      Some(url) => fetch_link_metadata(url, false, &context)
+        .await
+        .unwrap_or_default(),
+      _ => Default::default(),
+    };
+    let attachment = self
+      .url
+      .clone()
+      .map(|url| Attachment::new(url, metadata.content_type))
+      .into_iter()
+      .collect();
+
     let page = Page {
       kind: PageType::Page,
       id: self.ap_id.clone().into(),
@@ -122,7 +135,7 @@ impl Object for ApubPost {
       content: self.body.as_ref().map(|b| markdown_to_html(b)),
       media_type: Some(MediaTypeMarkdownOrHtml::Html),
       source: self.body.clone().map(Source::new),
-      attachment: self.url.clone().map(Attachment::new).into_iter().collect(),
+      attachment,
       image: self.thumbnail_url.clone().map(ImageObject::new),
       comments_enabled: Some(!self.locked),
       sensitive: Some(self.nsfw),

crates/apub/src/protocol/objects/page.rs

@@ -72,24 +72,25 @@ pub struct Page {
 #[derive(Clone, Debug, Deserialize, Serialize)]
 #[serde(rename_all = "camelCase")]
 pub(crate) struct Link {
-  pub(crate) href: Url,
+  href: Url,
-  pub(crate) r#type: LinkType,
+  content_type: Option<String>,
+  r#type: LinkType,
 }
 
 #[derive(Clone, Debug, Deserialize, Serialize)]
 #[serde(rename_all = "camelCase")]
 pub(crate) struct Image {
   #[serde(rename = "type")]
-  pub(crate) kind: ImageType,
+  kind: ImageType,
-  pub(crate) url: Url,
+  url: Url,
 }
 
 #[derive(Clone, Debug, Deserialize, Serialize)]
 #[serde(rename_all = "camelCase")]
 pub(crate) struct Document {
   #[serde(rename = "type")]
-  pub(crate) kind: DocumentType,
+  kind: DocumentType,
-  pub(crate) url: Url,
+  url: Url,
 }
 
 #[derive(Clone, Debug, Deserialize, Serialize)]
@@ -167,9 +168,10 @@ impl Page {
 }
 
 impl Attachment {
-  pub(crate) fn new(url: DbUrl) -> Attachment {
+  pub(crate) fn new(url: DbUrl, content_type: Option<String>) -> Attachment {
     Attachment::Link(Link {
       href: url.into(),
+      content_type,
       r#type: Default::default(),
     })
   }