Fixed content_id persistence

Cargo.lock

@@ -690,9 +690,9 @@ dependencies = [
 
 [[package]]
 name = "clap"
-version = "4.6.2"
+version = "4.6.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dd059f9da4f5c36b3787f65d38ccaab1cc315f07b01f89abc8359ee6a8205011"
+checksum = "0fb99565819980999fb7b4a1796046a5c949e6d4ff132cf5fadf5a641e20d776"
 dependencies = [
  "clap_builder",
  "clap_derive",
@@ -712,9 +712,9 @@ dependencies = [
 
 [[package]]
 name = "clap_derive"
-version = "4.6.1"
+version = "4.6.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f2ce8604710f6733aa641a2b3731eaa1e8b3d9973d5e3565da11800813f997a9"
+checksum = "32f2392eae7f16557a3d727ef3a12e57b2b2ca6f98566a5f4fb41ffe305df077"
 dependencies = [
  "heck",
  "proc-macro2",
@@ -1717,7 +1717,7 @@ dependencies = [
 [[package]]
 name = "federation-net"
 version = "0.1.0"
-source = "git+https://gt.hexor.cy/ab/frid.git#a897737978d476c223b66e65b960b70376084f2a"
+source = "git+https://gt.hexor.cy/ab/frid.git#8ee1db9cf89ea604c6b8a8c3fe089a714c1e321f"
 dependencies = [
  "blake3",
  "data-encoding",
@@ -1844,11 +1844,12 @@ checksum = "e6d5a32815ae3f33302d95fdcb2ce17862f8c65363dcfd29360480ba1001fc9c"
 
 [[package]]
 name = "furumusic"
-version = "0.6.2-fd"
+version = "0.6.5-fd"
 dependencies = [
  "anyhow",
  "async-trait",
  "base64 0.22.1",
+ "blake3",
  "chrono",
  "cot",
  "croner",
@@ -2470,9 +2471,9 @@ dependencies = [
 
 [[package]]
 name = "hyper"
-version = "1.10.1"
+version = "1.11.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "55281c53a1894c864990125767da440a4e630446785086f52523b20033b74498"
+checksum = "d22053281f852e11534f5198498373cbb59295120a20771d90f7ed1897490a72"
 dependencies = [
  "atomic-waker",
  "bytes",
@@ -3621,7 +3622,7 @@ dependencies = [
 [[package]]
 name = "music-dht"
 version = "0.1.0"
-source = "git+https://gt.hexor.cy/ab/frid.git#a897737978d476c223b66e65b960b70376084f2a"
+source = "git+https://gt.hexor.cy/ab/frid.git#8ee1db9cf89ea604c6b8a8c3fe089a714c1e321f"
 dependencies = [
  "async-trait",
  "blake3",

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "furumusic"
-version = "0.6.3-fd"
+version = "0.6.6-fd"
 edition = "2024"
 description = "Reusable web-app boilerplate: auth, OIDC/SSO, admin panel, user management, i18n, PostgreSQL"
 
@@ -16,6 +16,7 @@ reqwest = { version = "0.12", default-features = false, features = ["rustls-tls"
 tokio = { version = "1", features = ["sync", "fs", "io-util"] }
 tower = "0.5"
 base64 = "0.22"
+blake3 = "1"
 serde_json = "1"
 tracing = "0.1"
 tracing-subscriber = { version = "0.3", features = ["env-filter"] }

src/federation/mod.rs

@@ -15,6 +15,7 @@
 mod serve;
 mod storage;
 
+use std::collections::{HashMap, HashSet};
 use std::path::PathBuf;
 use std::sync::{Arc, OnceLock};
 use std::time::Duration;
@@ -42,10 +43,19 @@ struct Running {
     tasks: Vec<tokio::task::JoinHandle<()>>,
 }
 
+struct ContentHashJob {
+    media_file_id: i64,
+    sha256_hash: String,
+    file_path: String,
+}
+
 pub struct Federation {
     /// Transport data directory; server-side DHT state and identity live in PostgreSQL.
     data_dir: PathBuf,
     database_url: std::sync::Mutex<String>,
+    storage_dir: std::sync::Mutex<String>,
+    content_cache: std::sync::Mutex<HashMap<i64, (String, String)>>,
+    content_pending: std::sync::Mutex<HashSet<i64>>,
     pool: tokio::sync::OnceCell<PgPool>,
     running: tokio::sync::Mutex<Option<Running>>,
     last_sync: std::sync::Mutex<Option<String>>,
@@ -69,6 +79,9 @@ pub fn handle() -> Arc<Federation> {
         Arc::new(Federation {
             data_dir: PathBuf::from(crate::media_paths::resolve_config_path("federation")),
             database_url: std::sync::Mutex::new(String::new()),
+            storage_dir: std::sync::Mutex::new(String::new()),
+            content_cache: std::sync::Mutex::new(Default::default()),
+            content_pending: std::sync::Mutex::new(Default::default()),
             pool: tokio::sync::OnceCell::new(),
             running: tokio::sync::Mutex::new(None),
             last_sync: std::sync::Mutex::new(None),
@@ -149,6 +162,7 @@ impl Federation {
     /// node. Called at boot and every time the admin settings are saved.
     pub async fn apply(self: &Arc<Self>, config: &AppConfig) {
         *lock(&self.database_url) = config.database_url.clone();
+        *lock(&self.storage_dir) = config.agent_storage_dir.clone();
         let network = config.federation_network_id.trim().to_string();
         if config.federation_enabled && !network.is_empty() {
             if let Err(err) = self.start(network, config.agent_storage_dir.clone()).await {
@@ -281,7 +295,7 @@ impl Federation {
         Ok(())
     }
 
-    async fn sync_once(&self, service: &MusicDhtService) -> Result<SyncStats> {
+    async fn sync_once(self: &Arc<Self>, service: &MusicDhtService) -> Result<SyncStats> {
         let specs = match self.collect_specs().await {
             Ok(specs) => specs,
             Err(err) => {
@@ -341,7 +355,7 @@ impl Federation {
 
     /// Everything the regular player shows, as DHT item specs: non-hidden
     /// artists, releases and tracks (a track also hides with its release).
-    async fn collect_specs(&self) -> Result<Vec<ItemSpec>> {
+    async fn collect_specs(self: &Arc<Self>) -> Result<Vec<ItemSpec>> {
         let pool = self.pool().await?;
         let mut specs = Vec::new();
 
@@ -362,6 +376,7 @@ impl Federation {
                 track_number: None,
                 disc_number: None,
                 duration_seconds: None,
+                content_id: None,
             });
         }
 
@@ -400,6 +415,7 @@ impl Federation {
                 track_number: None,
                 disc_number: None,
                 duration_seconds: None,
+                content_id: None,
             });
         }
 
@@ -426,16 +442,38 @@ impl Federation {
         }
         let tracks = sqlx::query(
             "SELECT t.id, t.title, COALESCE(t.year, r.year), t.duration_seconds,
-                    r.title, r.release_type, t.track_number, t.disc_number
+                    r.title, r.release_type, t.track_number, t.disc_number,
+                    t.audio_file_id, m.file_path, m.sha256_hash, c.content_id
              FROM furumusic__track t
              JOIN furumusic__release r ON r.id = t.release_id
+             JOIN furumusic__media_file m ON m.id = t.audio_file_id
+             LEFT JOIN furumusic__federation_content_id_cache c
+                    ON c.media_file_id = m.id AND c.sha256_hash = m.sha256_hash
              WHERE t.is_hidden = false AND r.is_hidden = false",
         )
         .fetch_all(&pool)
         .await?;
+        let storage_dir = lock(&self.storage_dir).clone();
+        let mut content_hash_jobs = Vec::new();
         for row in &tracks {
             let id: i64 = row.get(0);
             let duration: f64 = row.get(3);
+            let media_file_id: i64 = row.get(8);
+            let file_path: String = row.get(9);
+            let sha256_hash: String = row.get(10);
+            let cached_content_id: Option<String> = row.get(11);
+            let content_id = cached_content_id
+                .or_else(|| self.cached_content_id_for_media(media_file_id, &sha256_hash));
+            if content_id.is_none()
+                && !storage_dir.trim().is_empty()
+                && self.mark_content_hash_pending(media_file_id)
+            {
+                content_hash_jobs.push(ContentHashJob {
+                    media_file_id,
+                    sha256_hash,
+                    file_path,
+                });
+            }
             specs.push(ItemSpec {
                 local_key: format!("track:{id}"),
                 kind: ItemKind::Track,
@@ -448,12 +486,72 @@ impl Federation {
                 track_number: row.get(6),
                 disc_number: row.get(7),
                 duration_seconds: (duration > 0.0).then_some(duration),
+                content_id,
             });
         }
+        self.spawn_content_warmer(pool.clone(), storage_dir, content_hash_jobs);
 
         Ok(specs)
     }
 
+    fn cached_content_id_for_media(&self, media_file_id: i64, sha256_hash: &str) -> Option<String> {
+        if let Some((cached_hash, content_id)) = lock(&self.content_cache).get(&media_file_id)
+            && cached_hash == sha256_hash
+        {
+            return Some(content_id.clone());
+        }
+        None
+    }
+
+    fn mark_content_hash_pending(&self, media_file_id: i64) -> bool {
+        lock(&self.content_pending).insert(media_file_id)
+    }
+
+    fn spawn_content_warmer(
+        self: &Arc<Self>,
+        pool: PgPool,
+        storage_dir: String,
+        jobs: Vec<ContentHashJob>,
+    ) {
+        if jobs.is_empty() {
+            return;
+        }
+        let fed = Arc::clone(self);
+        tokio::spawn(async move {
+            let total = jobs.len();
+            let mut stored = 0usize;
+            for job in jobs {
+                let job_storage_dir = storage_dir.clone();
+                let job_file_path = job.file_path.clone();
+                let content_id = tokio::task::spawn_blocking(move || {
+                    audio_content_id(&job_storage_dir, &job_file_path)
+                })
+                .await
+                .ok()
+                .flatten();
+                lock(&fed.content_pending).remove(&job.media_file_id);
+                if let Some(content_id) = content_id {
+                    lock(&fed.content_cache).insert(
+                        job.media_file_id,
+                        (job.sha256_hash.clone(), content_id.clone()),
+                    );
+                    if let Err(err) =
+                        persist_content_id(&pool, job.media_file_id, &job.sha256_hash, &content_id)
+                            .await
+                    {
+                        tracing::warn!(
+                            media_file_id = job.media_file_id,
+                            "federation content-id cache write failed: {err:#}"
+                        );
+                    } else {
+                        stored += 1;
+                    }
+                }
+            }
+            tracing::info!(total, stored, "federation content-id cache warm finished");
+        });
+    }
+
     /// Live status for the admin page.
     pub async fn status(&self) -> Value {
         let guard = self.running.lock().await;
@@ -511,6 +609,41 @@ impl Federation {
     }
 }
 
+async fn persist_content_id(
+    pool: &PgPool,
+    media_file_id: i64,
+    sha256_hash: &str,
+    content_id: &str,
+) -> Result<()> {
+    sqlx::query(
+        "INSERT INTO furumusic__federation_content_id_cache
+         (media_file_id, sha256_hash, content_id, updated_at)
+         VALUES ($1, $2, $3, $4)
+         ON CONFLICT (media_file_id) DO UPDATE SET
+            sha256_hash = EXCLUDED.sha256_hash,
+            content_id = EXCLUDED.content_id,
+            updated_at = EXCLUDED.updated_at",
+    )
+    .bind(media_file_id)
+    .bind(sha256_hash)
+    .bind(content_id)
+    .bind(now_iso())
+    .execute(pool)
+    .await?;
+    Ok(())
+}
+
+fn audio_content_id(storage_dir: &str, file_path: &str) -> Option<String> {
+    if storage_dir.trim().is_empty() {
+        return None;
+    }
+    let path = crate::media_paths::resolve_media_file_path(storage_dir, file_path);
+    let mut file = std::fs::File::open(path).ok()?;
+    let mut hasher = blake3::Hasher::new();
+    std::io::copy(&mut file, &mut hasher).ok()?;
+    Some(format!("b3:{}", hasher.finalize().to_hex()))
+}
+
 async fn stop_running(running: Option<Running>) {
     let Some(running) = running else { return };
     for task in &running.tasks {

src/federation/serve.rs

@@ -101,6 +101,8 @@ struct CatalogTrack {
     track_number: Option<i32>,
     disc_number: Option<i32>,
     duration_seconds: Option<f64>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    content_id: Option<String>,
     item_id: String,
 }
 
@@ -612,32 +614,36 @@ async fn build_catalog(pool: &PgPool, own: &EndpointId, artist: &str) -> Result<
     for release_row in release_rows {
         let release_id: i64 = release_row.get(0);
         let track_rows = sqlx::query(
-            "SELECT id, title, track_number, disc_number, duration_seconds
-             FROM furumusic__track
-             WHERE release_id = $1 AND is_hidden = false
-             ORDER BY disc_number NULLS FIRST, track_number NULLS LAST, title",
+            "SELECT t.id, t.title, t.track_number, t.disc_number, t.duration_seconds,
+                    c.content_id
+             FROM furumusic__track t
+             JOIN furumusic__media_file m ON m.id = t.audio_file_id
+             LEFT JOIN furumusic__federation_content_id_cache c
+                    ON c.media_file_id = m.id AND c.sha256_hash = m.sha256_hash
+             WHERE t.release_id = $1 AND t.is_hidden = false
+             ORDER BY t.disc_number NULLS FIRST, t.track_number NULLS LAST, t.title",
         )
         .bind(release_id)
         .fetch_all(pool)
         .await?;
+        let mut tracks = Vec::with_capacity(track_rows.len());
+        for row in track_rows {
+            let track_id: i64 = row.get(0);
+            let duration: f64 = row.get(4);
+            tracks.push(CatalogTrack {
+                title: row.get(1),
+                track_number: row.get(2),
+                disc_number: row.get(3),
+                duration_seconds: (duration > 0.0).then_some(duration),
+                content_id: row.get(5),
+                item_id: item_id_of(own, track_id),
+            });
+        }
         releases.push(CatalogRelease {
             title: release_row.get(1),
             release_type: release_row.get(2),
             year: release_row.get(3),
-            tracks: track_rows
-                .into_iter()
-                .map(|row| {
-                    let track_id: i64 = row.get(0);
-                    let duration: f64 = row.get(4);
-                    CatalogTrack {
-                        title: row.get(1),
-                        track_number: row.get(2),
-                        disc_number: row.get(3),
-                        duration_seconds: (duration > 0.0).then_some(duration),
-                        item_id: item_id_of(own, track_id),
-                    }
-                })
-                .collect(),
+            tracks,
         });
     }
 

src/federation/storage.rs

@@ -44,6 +44,14 @@ const SCHEMA: &[&str] = &[
         ticket        TEXT NOT NULL,
         last_seen_ms  BIGINT NOT NULL
     )",
+    "CREATE TABLE IF NOT EXISTS furumusic__federation_content_id_cache (
+        media_file_id  BIGINT PRIMARY KEY,
+        sha256_hash    TEXT NOT NULL,
+        content_id     TEXT NOT NULL,
+        updated_at     TEXT NOT NULL
+    )",
+    "CREATE INDEX IF NOT EXISTS idx_furumusic_federation_content_id_cache_content_id
+        ON furumusic__federation_content_id_cache(content_id)",
 ];
 
 #[derive(Debug, Clone)]