From 8e8b60fe311a9dd0991ea3c50068128d376be65c Mon Sep 17 00:00:00 2001
From: yggverse <yggverse@project>
Date: Sun, 15 Jun 2025 18:40:11 +0300
Subject: [PATCH] cleanup irrelevant files on complete torrent download

---
 Cargo.toml     |  2 ++
 src/main.rs    |  6 +++++-
 src/storage.rs | 24 ++++++++++++++++++++++--
 3 files changed, 29 insertions(+), 3 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index 75ce413..14de7f6 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -15,11 +15,13 @@ anyhow = "1.0"
 clap = { version = "4.5", features = ["derive"] }
 hyper-util = "0.1"
 librqbit = {version = "9.0.0-beta.0", features = ["disable-upload"]}
+regex = "1.11"
 serde = { version = "1.0", features = ["derive"] }
 serde_json = "1.0"
 tokio = { version = "1.45", features = ["full"] }
 tracing-subscriber = "0.3"
 url = "2.5"
+walkdir = "2.5"
 [patch.crates-io]
 librqbit = { git = "https://github.com/ikatson/rqbit.git", package = "librqbit" }
 #librqbit = { version = "9.0.0-beta.0", path = "../../rqbit/crates/librqbit", package = "librqbit" }
\ No newline at end of file
diff --git a/src/main.rs b/src/main.rs
index 185e31b..fc02699 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -83,7 +83,7 @@ async fn main() -> Result<()> {
                                     list_only: arg.preload_regex.is_none(),
                                     // the destination folder to preload files match `only_files_regex`
                                     // * e.g. images for audio albums
-                                    output_folder: storage.output_folder(&i).ok(),
+                                    output_folder: storage.output_folder(&i, true).ok(),
                                     only_files_regex: arg.preload_regex.clone(),
                                     ..Default::default()
                                 }),
@@ -124,6 +124,10 @@ async fn main() -> Result<()> {
                                                         false,
                                                     )
                                                     .await?;
+                                                // cleanup irrelevant files (see rqbit#408)
+                                                if let Some(ref r) = arg.preload_regex {
+                                                    storage.purge_preload_regex(&i, r)?;
+                                                }
                                                 // ignore on the next crawl iterations for this session
                                                 index.insert(mt.info_hash().as_string());
                                             }
diff --git a/src/storage.rs b/src/storage.rs
index cc2c11d..93e2679 100644
--- a/src/storage.rs
+++ b/src/storage.rs
@@ -37,17 +37,37 @@ impl Storage {
         Ok(p)
     }
 
-    pub fn output_folder(&self, infohash: &str) -> Result<String> {
+    pub fn output_folder(&self, infohash: &str, create: bool) -> Result<String> {
         let mut p = PathBuf::new();
         p.push(&self.0);
         p.push(infohash);
         if p.is_file() {
             bail!("File destination is not directory!");
         }
-        fs::create_dir_all(&p)?;
+        if create {
+            fs::create_dir_all(&p)?;
+        }
+        if !p.is_dir() {
+            bail!("Destination directory not exists!")
+        }
         Ok(p.to_string_lossy().to_string())
     }
 
+    /// Recursively remove all files match `pattern` under `infohash` location
+    pub fn purge_preload_regex(&self, infohash: &str, pattern: &str) -> Result<()> {
+        let r = regex::Regex::new(pattern)?;
+        for e in walkdir::WalkDir::new(self.output_folder(infohash, false)?)
+            .into_iter()
+            .filter_map(Result::ok)
+        {
+            let p = e.path();
+            if p.is_file() && !r.is_match(p.to_str().unwrap()) {
+                std::fs::remove_file(p)?;
+            }
+        }
+        Ok(())
+    }
+
     pub fn path(&self) -> PathBuf {
         self.0.clone()
     }