Organize repository files into hierarchical directory structure for improved scalability

- Update `FileBackend` to store files in a multi-level directory structure (`xx/yy/hash`) based on hash prefixes.
- Adjust method signatures and logic across `FileBackend`, `pkg6dev`, and `pkg6repo` to support mutable repository operations.
- Add test cases and scripts (`test_file_structure.rs`, `test_repo_operations.sh`) to validate the new file structure.
- Refactor catalog manager to use `RefCell` for interior mutability to enable lazy initialization.
- Clean up redundant and dead code, enhance comments, and replace outdated logic for file handling and metadata operations.

libips/src/repository/file_backend.rs

@@ -197,7 +197,8 @@ pub struct FileBackend {
     pub path: PathBuf,
     pub config: RepositoryConfig,
     /// Catalog manager for handling catalog operations
-    catalog_manager: Option<crate::repository::catalog::CatalogManager>,
+    /// Uses RefCell for interior mutability to allow mutation through immutable references
+    catalog_manager: Option<std::cell::RefCell<crate::repository::catalog::CatalogManager>>,
 }
 
 /// Format a SystemTime as an ISO 8601 timestamp string
@@ -415,8 +416,23 @@ impl Transaction {
 
         // Copy files to their final location
         for (source_path, hash) in self.files {
-            // Create the destination path in the files directory
+            // Create the destination path using the new directory structure
-            let dest_path = self.repo.join("file").join(&hash);
+            let dest_path = if hash.len() < 4 {
+                // Fallback for very short hashes (shouldn't happen with SHA256)
+                self.repo.join("file").join(&hash)
+            } else {
+                // Extract the first two and next two characters from the hash
+                let first_two = &hash[0..2];
+                let next_two = &hash[2..4];
+                
+                // Create the path: $REPO/file/XX/YY/XXYY...
+                self.repo.join("file").join(first_two).join(next_two).join(&hash)
+            };
+
+            // Create parent directories if they don't exist
+            if let Some(parent) = dest_path.parent() {
+                fs::create_dir_all(parent)?;
+            }
 
             // Copy the file if it doesn't already exist
             if !dest_path.exists() {
@@ -1211,7 +1227,7 @@ impl WritableRepository for FileBackend {
     }
 
     /// Rebuild repository metadata
-    fn rebuild(&self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()> {
+    fn rebuild(&mut self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()> {
         // Filter publishers if specified
         let publishers = if let Some(pub_name) = publisher {
             if !self.config.publishers.contains(&pub_name.to_string()) {
@@ -1228,7 +1244,7 @@ impl WritableRepository for FileBackend {
 
             if !no_catalog {
                 info!("Rebuilding catalog...");
-                // In a real implementation, we would rebuild the catalog
+                self.generate_catalog_parts(&pub_name, true)?;
             }
 
             if !no_index {
@@ -1241,7 +1257,7 @@ impl WritableRepository for FileBackend {
     }
 
     /// Refresh repository metadata
-    fn refresh(&self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()> {
+    fn refresh(&mut self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()> {
         // Filter publishers if specified
         let publishers = if let Some(pub_name) = publisher {
             if !self.config.publishers.contains(&pub_name.to_string()) {
@@ -1258,22 +1274,12 @@ impl WritableRepository for FileBackend {
 
             if !no_catalog {
                 info!("Refreshing catalog...");
-                // In a real implementation, we would refresh the catalog
+                self.generate_catalog_parts(&pub_name, true)?;
             }
 
             if !no_index {
                 info!("Refreshing search index...");
-
+                self.build_search_index(&pub_name)?;
-                // Check if the index exists
-                let index_path = self.path.join("index").join(&pub_name).join("search.json");
-                if !index_path.exists() {
-                    // If the index doesn't exist, build it
-                    self.build_search_index(&pub_name)?;
-                } else {
-                    // If the index exists, update it
-                    // For simplicity, we'll just rebuild it
-                    self.build_search_index(&pub_name)?;
-                }
             }
         }
 
@@ -1309,23 +1315,42 @@ impl FileBackend {
 
         Ok(())
     }
+    
+    /// Generate the file path for a given hash using the new directory structure
+    /// The path will be $REPO/file/XX/YY/XXYY... where XX and YY are the first four letters of the hash
+    fn generate_file_path(&self, hash: &str) -> PathBuf {
+        if hash.len() < 4 {
+            // Fallback for very short hashes (shouldn't happen with SHA256)
+            return self.path.join("file").join(hash);
+        }
+        
+        // Extract the first two and next two characters from the hash
+        let first_two = &hash[0..2];
+        let next_two = &hash[2..4];
+        
+        // Create the path: $REPO/file/XX/YY/XXYY...
+        self.path.join("file").join(first_two).join(next_two).join(hash)
+    }
 
     /// Get or initialize the catalog manager
+    /// 
+    /// This method returns a mutable reference to the catalog manager.
+    /// It uses interior mutability with RefCell to allow mutation through an immutable reference.
     pub fn get_catalog_manager(
         &mut self,
-    ) -> Result<&mut crate::repository::catalog::CatalogManager> {
+    ) -> Result<std::cell::RefMut<crate::repository::catalog::CatalogManager>> {
         if self.catalog_manager.is_none() {
             let catalog_dir = self.path.join("catalog");
-            self.catalog_manager = Some(crate::repository::catalog::CatalogManager::new(
+            let manager = crate::repository::catalog::CatalogManager::new(&catalog_dir)?;
-                &catalog_dir,
+            let refcell = std::cell::RefCell::new(manager);
-            )?);
+            self.catalog_manager = Some(refcell);
         }
 
-        Ok(self.catalog_manager.as_mut().unwrap())
+        // This is safe because we just checked that catalog_manager is Some
+        Ok(self.catalog_manager.as_ref().unwrap().borrow_mut())
     }
 
     /// URL encode a string for use in a filename
-    #[allow(dead_code)]
     fn url_encode(s: &str) -> String {
         let mut result = String::new();
         for c in s.chars() {
@@ -1342,7 +1367,6 @@ impl FileBackend {
     }
 
     /// Generate catalog parts for a publisher
-    #[allow(dead_code)]
     fn generate_catalog_parts(&mut self, publisher: &str, create_update_log: bool) -> Result<()> {
         info!("Generating catalog parts for publisher: {}", publisher);
 
@@ -1471,7 +1495,7 @@ impl FileBackend {
         }
 
         // Now get the catalog manager and create the catalog parts
-        let catalog_manager = self.get_catalog_manager()?;
+        let mut catalog_manager = self.get_catalog_manager()?;
 
         // Create and populate the base part
         let base_part_name = "catalog.base.C".to_string();
@@ -1796,7 +1820,7 @@ impl FileBackend {
 
         // Verify the file was stored
         let hash = Transaction::calculate_file_hash(&test_file_path)?;
-        let stored_file_path = self.path.join("file").join(&hash);
+        let stored_file_path = self.generate_file_path(&hash);
 
         if !stored_file_path.exists() {
             return Err(RepositoryError::Other("File was not stored correctly".to_string()));
@@ -1827,7 +1851,7 @@ impl FileBackend {
     }
 
     /// Publish files from a prototype directory
-    pub fn publish_files<P: AsRef<Path>>(&self, proto_dir: P, publisher: &str) -> Result<()> {
+    pub fn publish_files<P: AsRef<Path>>(&mut self, proto_dir: P, publisher: &str) -> Result<()> {
         let proto_dir = proto_dir.as_ref();
 
         // Check if the prototype directory exists
@@ -1901,8 +1925,13 @@ impl FileBackend {
         // Calculate the SHA256 hash of the file
         let hash = Transaction::calculate_file_hash(file_path)?;
 
-        // Create the destination path in the files directory
+        // Create the destination path using the new directory structure
-        let dest_path = self.path.join("file").join(&hash);
+        let dest_path = self.generate_file_path(&hash);
+
+        // Create parent directories if they don't exist
+        if let Some(parent) = dest_path.parent() {
+            fs::create_dir_all(parent)?;
+        }
 
         // Copy the file if it doesn't already exist
         if !dest_path.exists() {

libips/src/repository/mod.rs

@@ -330,10 +330,10 @@ pub trait WritableRepository {
     ) -> Result<()>;
 
     /// Rebuild repository metadata
-    fn rebuild(&self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()>;
+    fn rebuild(&mut self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()>;
 
     /// Refresh repository metadata
-    fn refresh(&self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()>;
+    fn refresh(&mut self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()>;
 
     /// Set the default publisher for the repository
     fn set_default_publisher(&mut self, publisher: &str) -> Result<()>;

libips/src/repository/rest_backend.rs

@@ -89,7 +89,7 @@ impl WritableRepository for RestBackend {
     }
 
     /// Rebuild repository metadata
-    fn rebuild(&self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()> {
+    fn rebuild(&mut self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()> {
         // This is a stub implementation
         // In a real implementation, we would make a REST API call to rebuild metadata
 
@@ -122,7 +122,7 @@ impl WritableRepository for RestBackend {
     }
 
     /// Refresh repository metadata
-    fn refresh(&self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()> {
+    fn refresh(&mut self, publisher: Option<&str>, no_catalog: bool, no_index: bool) -> Result<()> {
         // This is a stub implementation
         // In a real implementation, we would make a REST API call to refresh metadata
 

libips/src/repository/tests.rs

@@ -393,4 +393,39 @@ mod tests {
         // Clean up
         cleanup_test_dir(&test_dir);
     }
+    
+    #[test]
+    fn test_file_structure() {
+        // Create a test directory
+        let test_dir = create_test_dir("file_structure");
+        let repo_path = test_dir.join("repo");
+        
+        // Create a repository
+        let mut repo = FileBackend::create(&repo_path, RepositoryVersion::V4).unwrap();
+        
+        // Add a publisher
+        repo.add_publisher("test").unwrap();
+        
+        // Create a test file
+        let test_file_path = test_dir.join("test_file.txt");
+        fs::write(&test_file_path, "This is a test file").unwrap();
+        
+        // Store the file in the repository
+        let hash = repo.store_file(&test_file_path).unwrap();
+        
+        // Check if the file was stored in the correct directory structure
+        let first_two = &hash[0..2];
+        let next_two = &hash[2..4];
+        let expected_path = repo_path.join("file").join(first_two).join(next_two).join(&hash);
+        
+        // Verify that the file exists at the expected path
+        assert!(expected_path.exists(), "File was not stored at the expected path: {}", expected_path.display());
+        
+        // Verify that the file does NOT exist at the old path
+        let old_path = repo_path.join("file").join(&hash);
+        assert!(!old_path.exists(), "File was stored at the old path: {}", old_path.display());
+        
+        // Clean up
+        cleanup_test_dir(&test_dir);
+    }
 }

pkg6dev/src/main.rs

@@ -426,7 +426,7 @@ fn publish_package(
 
     // Open the repository
     info!("Opening repository at: {}", repo_path.display());
-    let repo = match FileBackend::open(repo_path) {
+    let mut repo = match FileBackend::open(repo_path) {
         Ok(repo) => repo,
         Err(_) => {
             info!("Repository does not exist, creating a new one...");

pkg6repo/src/main.rs

@@ -868,7 +868,7 @@ fn main() -> Result<()> {
             // Open the repository
             // In a real implementation with RestBackend, the key and cert parameters would be used for SSL authentication
             // For now, we're using FileBackend, which doesn't use these parameters
-            let repo = FileBackend::open(repo_uri_or_path)?;
+            let mut repo = FileBackend::open(repo_uri_or_path)?;
 
             // Get the publisher if specified
             let pub_option = if let Some(publishers) = publisher {
@@ -899,7 +899,7 @@ fn main() -> Result<()> {
             // Open the repository
             // In a real implementation with RestBackend, the key and cert parameters would be used for SSL authentication
             // For now, we're using FileBackend, which doesn't use these parameters
-            let repo = FileBackend::open(repo_uri_or_path)?;
+            let mut repo = FileBackend::open(repo_uri_or_path)?;
 
             // Get the publisher if specified
             let pub_option = if let Some(publishers) = publisher {

test_file_structure.rs

@@ -0,0 +1,57 @@
+use std::fs;
+use std::path::Path;
+use libips::repository::{FileBackend, WritableRepository, ReadableRepository, RepositoryVersion};
+
+fn main() -> Result<(), Box<dyn std::error::Error>> {
+    // Create a temporary directory for the test
+    let test_dir = Path::new("/tmp/pkg6_file_structure_test");
+    if test_dir.exists() {
+        fs::remove_dir_all(test_dir)?;
+    }
+    fs::create_dir_all(test_dir)?;
+    
+    println!("Created test directory: {}", test_dir.display());
+    
+    // Create a new repository
+    let mut repo = FileBackend::create(test_dir, RepositoryVersion::V1)?;
+    
+    // Add a publisher
+    repo.add_publisher("test")?;
+    
+    println!("Created repository with publisher 'test'");
+    
+    // Create a test file
+    let test_file_path = test_dir.join("test_file.txt");
+    fs::write(&test_file_path, "This is a test file")?;
+    
+    println!("Created test file: {}", test_file_path.display());
+    
+    // Store the file in the repository
+    let hash = repo.store_file(&test_file_path)?;
+    
+    println!("Stored file with hash: {}", hash);
+    
+    // Check if the file was stored in the correct directory structure
+    let first_two = &hash[0..2];
+    let next_two = &hash[2..4];
+    let expected_path = test_dir.join("file").join(first_two).join(next_two).join(&hash);
+    
+    if expected_path.exists() {
+        println!("SUCCESS: File was stored at the correct path: {}", expected_path.display());
+    } else {
+        println!("ERROR: File was not stored at the expected path: {}", expected_path.display());
+        
+        // Check if the file was stored in the old location
+        let old_path = test_dir.join("file").join(&hash);
+        if old_path.exists() {
+            println!("File was stored at the old path: {}", old_path.display());
+        } else {
+            println!("File was not stored at the old path either");
+        }
+    }
+    
+    // Clean up
+    fs::remove_dir_all(test_dir)?;
+    
+    Ok(())
+}

test_repo_operations.sh

@@ -0,0 +1,77 @@
+#!/bin/bash
+
+# Test script to verify that repository operations work with the new directory structure
+
+# Create a temporary directory for the test
+TEST_DIR="/tmp/pkg6_repo_operations_test"
+rm -rf "$TEST_DIR"
+mkdir -p "$TEST_DIR"
+
+echo "Created test directory: $TEST_DIR"
+
+# Create a repository
+echo "Creating repository..."
+cargo run --bin pkg6repo -- create "$TEST_DIR/repo"
+
+# Add a publisher
+echo "Adding publisher..."
+cargo run --bin pkg6repo -- add-publisher -r "$TEST_DIR/repo" test
+
+# Create a test file
+echo "Creating test file..."
+echo "This is a test file" > "$TEST_DIR/test_file.txt"
+
+# Create a simple manifest
+echo "Creating manifest..."
+cat > "$TEST_DIR/test.p5m" << EOF
+{
+  "attributes": [
+    {
+      "key": "pkg.fmri",
+      "values": [
+        "pkg://test/example@1.0,5.11-0:20250727T123000Z"
+      ],
+      "properties": {}
+    }
+  ],
+  "files": [
+    {
+      "path": "usr/share/doc/example/test.txt",
+      "mode": "0644",
+      "owner": "root",
+      "group": "root"
+    }
+  ]
+}
+EOF
+
+# Create a prototype directory
+echo "Creating prototype directory..."
+mkdir -p "$TEST_DIR/prototype/usr/share/doc/example"
+cp "$TEST_DIR/test_file.txt" "$TEST_DIR/prototype/usr/share/doc/example/test.txt"
+
+# Publish the package
+echo "Publishing package..."
+cargo run --bin pkg6repo -- publish -r "$TEST_DIR/repo" -p test -m "$TEST_DIR/test.p5m" "$TEST_DIR/prototype"
+
+# Check if the file was stored in the correct directory structure
+echo "Checking file structure..."
+find "$TEST_DIR/repo/file" -type f | while read -r file; do
+    hash=$(basename "$file")
+    first_two=${hash:0:2}
+    next_two=${hash:2:2}
+    expected_path="$TEST_DIR/repo/file/$first_two/$next_two/$hash"
+    
+    if [ "$file" = "$expected_path" ]; then
+        echo "SUCCESS: File was stored at the correct path: $file"
+    else
+        echo "ERROR: File was stored at an unexpected path: $file"
+        echo "Expected: $expected_path"
+    fi
+done
+
+# Clean up
+echo "Cleaning up..."
+rm -rf "$TEST_DIR"
+
+echo "Test completed."