feat: match RSS results by IMDb id (#99)

crates/arr-indexer/src/search.rs

@@ -72,6 +72,10 @@ pub struct SearchRelease {
     /// TMDB id the tracker attached to the item, when it attached one. An ID
     /// the indexer supplies beats anything read off the release name.
     pub tmdb_id: Option<u32>,
+    /// `IMDb` id the tracker attached to the item, as it spelled it. Kept
+    /// verbatim because definitions differ on the `tt` prefix and on the
+    /// zero padding; matching (§6.2) normalises both sides.
+    pub imdb_id: Option<String>,
 }
 
 /// Results from one indexer in a multi-indexer search.
@@ -283,6 +287,7 @@ struct ReleaseBuilder {
     link: Option<String>,
     enclosure_url: Option<String>,
     tmdb_id: Option<u32>,
+    imdb_id: Option<String>,
 }
 
 impl ReleaseBuilder {
@@ -300,6 +305,7 @@ impl ReleaseBuilder {
             publish_date: self.publish_date,
             download_url,
             tmdb_id: self.tmdb_id,
+            imdb_id: self.imdb_id,
         })
     }
 
@@ -428,6 +434,12 @@ fn read_element_attributes(
                         .filter(|&id| id != 0)
                         .or(builder.tmdb_id);
                 }
+                Some("imdbid" | "imdb") => {
+                    builder.imdb_id = value
+                        .map(|value| value.trim().to_owned())
+                        .filter(|value| !value.is_empty())
+                        .or_else(|| builder.imdb_id.take());
+                }
                 _ => {}
             }
         }
@@ -760,6 +772,45 @@ mod tests {
         );
     }
 
+    /// `imdbid` is the id most Cardigann definitions carry, and it is kept
+    /// exactly as the tracker spelled it — matching normalises (§6.2).
+    #[test]
+    fn a_torznab_imdb_id_is_kept() {
+        let body = br#"
+            <rss version="2.0" xmlns:torznab="http://torznab.com/schemas/2015/feed">
+              <channel>
+                <item>
+                  <title>Dune.Part.Two.2024.2160p.WEB-DL</title>
+                  <guid>prefixed</guid>
+                  <link>https://indexer.invalid/download/prefixed</link>
+                  <torznab:attr name="imdbid" value="tt15239678"/>
+                </item>
+                <item>
+                  <title>Dune.2021.2160p.WEB-DL</title>
+                  <guid>bare</guid>
+                  <link>https://indexer.invalid/download/bare</link>
+                  <torznab:attr name="imdbid" value=" 1160419 "/>
+                </item>
+                <item>
+                  <title>Nosferatu.2024.2160p.WEB-DL</title>
+                  <guid>anonymous</guid>
+                  <link>https://indexer.invalid/download/anonymous</link>
+                  <torznab:attr name="imdbid" value=""/>
+                </item>
+              </channel>
+            </rss>
+        "#;
+
+        let releases = parse_releases(3, body).expect("feed structure is valid");
+
+        assert_eq!(releases[0].imdb_id.as_deref(), Some("tt15239678"));
+        assert_eq!(releases[1].imdb_id.as_deref(), Some("1160419"));
+        assert_eq!(
+            releases[2].imdb_id, None,
+            "an empty attribute is the tracker saying it does not know"
+        );
+    }
+
     /// A pack and one episode of that pack carry the same series title, so
     /// the tag on the release name is the only thing separating them.
     #[test]