feat: file deletion implementation

crates/daemons/crond/Cargo.toml

@@ -1,5 +1,5 @@
 [package]
-name = "crond"
+name = "revolt-crond"
 version = "0.7.19"
 license = "AGPL-3.0-or-later"
 authors = ["Paul Makles <me@insrt.uk>"]
@@ -9,6 +9,14 @@ description = "Revolt Daemon Service: Timed data clean up tasks"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
+# Utility
+log = "0.4"
+
+# Async
+tokio = { version = "1" }
+
 # Core
 revolt-database = { version = "0.7.19", path = "../../core/database" }
+revolt-result = { version = "0.7.19", path = "../../core/result" }
+revolt-config = { version = "0.7.19", path = "../../core/config" }
 revolt-files = { version = "0.7.19", path = "../../core/files" }

crates/daemons/crond/src/main.rs

@@ -1,3 +1,19 @@
-fn main() {
-    println!("Hello, world!");
+use revolt_config::configure;
+use revolt_database::DatabaseInfo;
+use revolt_result::Result;
+use tasks::{file_deletion, prune_dangling_files};
+use tokio::try_join;
+
+pub mod tasks;
+
+#[tokio::main]
+async fn main() -> Result<()> {
+    configure!(crond);
+
+    let db = DatabaseInfo::Auto.connect().await.expect("database");
+    try_join!(
+        file_deletion::task(db.clone()),
+        prune_dangling_files::task(db)
+    )
+    .map(|_| ())
 }

crates/daemons/crond/src/tasks/file_deletion.rs

@@ -0,0 +1,39 @@
+use std::time::Duration;
+
+use log::info;
+use revolt_database::Database;
+use revolt_files::delete_from_s3;
+use revolt_result::Result;
+use tokio::time::sleep;
+
+pub async fn task(db: Database) -> Result<()> {
+    loop {
+        let files = db.fetch_deleted_attachments().await?;
+
+        for file in files {
+            let count = db
+                .count_file_hash_references(file.hash.as_ref().expect("no `hash` present"))
+                .await?;
+
+            // No other files reference this file on disk anymore
+            if count <= 1 {
+                let file_hash = db
+                    .fetch_attachment_hash(file.hash.as_ref().expect("no `hash` present"))
+                    .await?;
+
+                // Delete from S3
+                delete_from_s3(&file_hash.bucket_id, &file_hash.path).await?;
+
+                // Delete the hash
+                db.delete_attachment_hash(&file_hash.id).await?;
+                info!("Deleted file hash {}", file_hash.id);
+            }
+
+            // Delete the file
+            db.delete_attachment(&file.id).await?;
+            info!("Deleted file {}", file.id);
+        }
+
+        sleep(Duration::from_secs(60)).await;
+    }
+}

crates/daemons/crond/src/tasks/mod.rs

@@ -0,0 +1,2 @@
+pub mod file_deletion;
+pub mod prune_dangling_files;

crates/daemons/crond/src/tasks/prune_dangling_files.rs

@@ -0,0 +1,36 @@
+use std::time::Duration;
+
+use revolt_database::{iso8601_timestamp::Timestamp, Database};
+use revolt_result::Result;
+use tokio::time::sleep;
+
+use log::info;
+
+pub async fn task(db: Database) -> Result<()> {
+    loop {
+        // This could just be a single database query
+        // ... but timestamps are inconsistently serialised
+        // ... sometimes they are dates/numbers, hard to query
+        // ... in the future, we could use Postgres instead! :D
+        // ...
+        // ... on the plus side, it's still only 2 queries
+
+        let files = db.fetch_dangling_files().await?;
+        let file_ids: Vec<String> = files
+            .into_iter()
+            .filter(|file| {
+                file.uploaded_at.is_some_and(|uploaded_at| {
+                    Timestamp::now_utc().duration_since(uploaded_at) > Duration::from_secs(60 * 60)
+                })
+            })
+            .map(|file| file.id)
+            .collect();
+
+        if !file_ids.is_empty() {
+            db.mark_attachments_as_deleted(&file_ids).await?;
+            info!("Marked {} dangling files for deletion", file_ids.len());
+        }
+
+        sleep(Duration::from_secs(60)).await;
+    }
+}