mass file lister as an attempt to tackle #14507

modules/util.py

@@ -66,3 +66,73 @@ def truncate_path(target_path, base_path=cwd):
     except ValueError:
         pass
     return abs_target
+
+
+class MassFileListerCachedDir:
+    """A class that caches file metadata for a specific directory."""
+
+    def __init__(self, dirname):
+        self.files = None
+        self.files_cased = None
+        self.dirname = dirname
+
+        stats = ((x.name, x.stat(follow_symlinks=False)) for x in os.scandir(self.dirname))
+        files = [(n, s.st_mtime, s.st_ctime) for n, s in stats]
+        self.files = {x[0].lower(): x for x in files}
+        self.files_cased = {x[0]: x for x in files}
+
+
+class MassFileLister:
+    """A class that provides a way to check for the existence and mtime/ctile of files without doing more than one stat call per file."""
+
+    def __init__(self):
+        self.cached_dirs = {}
+
+    def find(self, path):
+        """
+        Find the metadata for a file at the given path.
+
+        Returns:
+            tuple or None: A tuple of (name, mtime, ctime) if the file exists, or None if it does not.
+        """
+
+        dirname, filename = os.path.split(path)
+
+        cached_dir = self.cached_dirs.get(dirname)
+        if cached_dir is None:
+            cached_dir = MassFileListerCachedDir(dirname)
+            self.cached_dirs[dirname] = cached_dir
+
+        stats = cached_dir.files_cased.get(filename)
+        if stats is not None:
+            return stats
+
+        stats = cached_dir.files.get(filename.lower())
+        if stats is None:
+            return None
+
+        try:
+            os_stats = os.stat(path, follow_symlinks=False)
+            return filename, os_stats.st_mtime, os_stats.st_ctime
+        except Exception:
+            return None
+
+    def exists(self, path):
+        """Check if a file exists at the given path."""
+
+        return self.find(path) is not None
+
+    def mctime(self, path):
+        """
+        Get the modification and creation times for a file at the given path.
+
+        Returns:
+            tuple: A tuple of (mtime, ctime) if the file exists, or (0, 0) if it does not.
+        """
+
+        stats = self.find(path)
+        return (0, 0) if stats is None else stats[1:3]
+
+    def reset(self):
+        """Clear the cache of all directories."""
+        self.cached_dirs.clear()