Add a cached fs abstraction (#141)

crates/common/src/cached_fs/mod.rs

@@ -0,0 +1,49 @@
+//! This module implements a cached file system allowing for results to be stored in-memory rather
+//! rather being queried from the file system again.
+
+use std::fs;
+use std::io::{Error, Result};
+use std::path::{Path, PathBuf};
+
+use moka::sync::Cache;
+use once_cell::sync::Lazy;
+
+pub fn read(path: impl AsRef<Path>) -> Result<Vec<u8>> {
+    static READ_CACHE: Lazy<Cache<PathBuf, Vec<u8>>> = Lazy::new(|| Cache::new(10_000));
+
+    let path = path.as_ref().canonicalize()?;
+    match READ_CACHE.get(path.as_path()) {
+        Some(content) => Ok(content),
+        None => {
+            let content = fs::read(path.as_path())?;
+            READ_CACHE.insert(path, content.clone());
+            Ok(content)
+        }
+    }
+}
+
+pub fn read_to_string(path: impl AsRef<Path>) -> Result<String> {
+    let content = read(path)?;
+    String::from_utf8(content).map_err(|_| {
+        Error::new(
+            std::io::ErrorKind::InvalidData,
+            "The contents of the file are not valid UTF8",
+        )
+    })
+}
+
+pub fn read_dir(path: impl AsRef<Path>) -> Result<Box<dyn Iterator<Item = Result<PathBuf>>>> {
+    static READ_DIR_CACHE: Lazy<Cache<PathBuf, Vec<PathBuf>>> = Lazy::new(|| Cache::new(10_000));
+
+    let path = path.as_ref().canonicalize()?;
+    match READ_DIR_CACHE.get(path.as_path()) {
+        Some(entries) => Ok(Box::new(entries.into_iter().map(Ok)) as Box<_>),
+        None => {
+            let entries = fs::read_dir(path.as_path())?
+                .flat_map(|maybe_entry| maybe_entry.map(|entry| entry.path()))
+                .collect();
+            READ_DIR_CACHE.insert(path.clone(), entries);
+            Ok(read_dir(path).unwrap())
+        }
+    }
+}

crates/common/src/iterators/files_with_extension_iterator.rs

@@ -19,6 +19,11 @@ pub struct FilesWithExtensionIterator {
     /// this vector then they will be returned when the [`Iterator::next`] method is called. If not
     /// then we visit one of the next directories to visit.
     files_matching_allowed_extensions: Vec<PathBuf>,
+
+    /// This option controls if the the cached file system should be used or not. This could be
+    /// better for certain cases where the entries in the directories do not change and therefore
+    /// caching can be used.
+    use_cached_fs: bool,
 }
 
 impl FilesWithExtensionIterator {
@@ -27,6 +32,7 @@ impl FilesWithExtensionIterator {
             allowed_extensions: Default::default(),
             directories_to_search: vec![root_directory.as_ref().to_path_buf()],
             files_matching_allowed_extensions: Default::default(),
+            use_cached_fs: Default::default(),
         }
     }
 
@@ -37,6 +43,11 @@ impl FilesWithExtensionIterator {
         self.allowed_extensions.insert(allowed_extension.into());
         self
     }
+
+    pub fn with_use_cached_fs(mut self, use_cached_fs: bool) -> Self {
+        self.use_cached_fs = use_cached_fs;
+        self
+    }
 }
 
 impl Iterator for FilesWithExtensionIterator {
@@ -49,16 +60,19 @@ impl Iterator for FilesWithExtensionIterator {
 
         let directory_to_search = self.directories_to_search.pop()?;
 
-        // Read all of the entries in the directory. If we failed to read this dir's entires then we
-        // elect to just ignore it and look in the next directory, we do that by calling the next
-        // method again on the iterator, which is an intentional decision that we made here instead
-        // of panicking.
-        let Ok(dir_entries) = std::fs::read_dir(directory_to_search) else {
-            return self.next();
+        let iterator = if self.use_cached_fs {
+            let Ok(dir_entries) = crate::cached_fs::read_dir(directory_to_search.as_path()) else {
+                return self.next();
+            };
+            Box::new(dir_entries) as Box<dyn Iterator<Item = std::io::Result<PathBuf>>>
+        } else {
+            let Ok(dir_entries) = std::fs::read_dir(directory_to_search) else {
+                return self.next();
+            };
+            Box::new(dir_entries.map(|maybe_entry| maybe_entry.map(|entry| entry.path()))) as Box<_>
         };
 
-        for entry in dir_entries.flatten() {
-            let entry_path = entry.path();
+        for entry_path in iterator.flatten() {
             if entry_path.is_dir() {
                 self.directories_to_search.push(entry_path)
             } else if entry_path.is_file()

crates/common/src/lib.rs

@@ -1,6 +1,7 @@
 //! This crate provides common concepts, functionality, types, macros, and more that other crates in
 //! the workspace can benefit from.
 
+pub mod cached_fs;
 pub mod fs;
 pub mod futures;
 pub mod iterators;