Speed up fixing-ownership when leaving breeze on Linux (apache#45805)

When leaving breeze on Linux we are fixing ownership of potential
new generated files to be the same as the HOST user - because files
created in container will be owned by root.

This takes quite some time however, especially when you have
.venv or node_modules or other folders with large amount of files.

This change skips such files from being considered in fixing ownership
if the folders are already properly owned.

dev/breeze/src/airflow_breeze/utils/docker_command_utils.py

@@ -579,6 +579,8 @@ def fix_ownership_using_docker(quiet: bool = False):
         "-e",
         f"HOST_GROUP_ID={get_host_group_id()}",
         "-e",
+        f"VERBOSE={str(get_verbose()).lower()}",
+        "-e",
         f"DOCKER_IS_ROOTLESS={is_docker_rootless()}",
         "--rm",
         "-t",

scripts/in_container/run_fix_ownership.py

@@ -44,27 +44,41 @@ def change_ownership_of_files(path: Path) -> None:
         sys.exit(1)
     count_files = 0
     root_uid = pwd.getpwnam("root").pw_uid
-    for file in path.rglob("*"):
-        try:
-            if file.is_symlink() and file.lstat().st_uid == root_uid:
-                # Change ownership of symlink itself (by default stat/chown follow the symlinks)
-                os.chown(file, int(host_user_id), int(host_group_id), follow_symlinks=False)
-                count_files += 1
+    skip_folders = {".venv", "node_modules", ".mypy_cache", ".pytest_cache", ".ruff_cache"}
+    for root, dirs, files in os.walk(path):
+        original_length = len(dirs)
+        original_dirs = dirs.copy()
+        # skip known big folders if they are not owned by root
+        dirs[:] = [d for d in dirs if d not in skip_folders or (Path(root) / d).stat().st_uid == root_uid]
+        new_length = len(dirs)
+        if new_length != original_length:
+            if os.environ.get("VERBOSE", "false") == "true":
+                print(
+                    f"{root}: Skipped {original_length - new_length} "
+                    f"folders: {set(original_dirs) - set(dirs)}"
+                )
+        for name in files:
+            file = Path(root) / name
+            try:
+                if file.is_symlink() and file.lstat().st_uid == root_uid:
+                    # Change ownership of symlink itself (by default stat/chown follow the symlinks)
+                    os.chown(file, int(host_user_id), int(host_group_id), follow_symlinks=False)
+                    count_files += 1
+                    if os.environ.get("VERBOSE_COMMANDS", "false") == "true":
+                        print(f"Changed ownership of symlink {file}")
+                if file.stat().st_uid == root_uid:
+                    # And here change ownership of the file (or if it is a symlink - the file it points to)
+                    os.chown(file, int(host_user_id), int(host_group_id))
+                    count_files += 1
+                    if os.environ.get("VERBOSE_COMMANDS", "false") == "true":
+                        print(f"Changed ownership of {file.resolve()}")
+            except FileNotFoundError:
+                # This is OK - file might have been deleted in the meantime or linked in Host from
+                # another place
                 if os.environ.get("VERBOSE_COMMANDS", "false") == "true":
-                    print(f"Changed ownership of symlink {file}")
-            if file.stat().st_uid == root_uid:
-                # And here change ownership of the file (or if it is a symlink - the file it points to)
-                os.chown(file, int(host_user_id), int(host_group_id))
-                count_files += 1
-                if os.environ.get("VERBOSE_COMMANDS", "false") == "true":
-                    print(f"Changed ownership of {file.resolve()}")
-        except FileNotFoundError:
-            # This is OK - file might have been deleted in the meantime or linked in Host from
-            # another place
-            if os.environ.get("VERBOSE_COMMANDS", "false") == "true":
-                print(f"Could not change ownership of {file}")
-    if count_files:
-        print(f"Changed ownership of {count_files} files back to {host_user_id}:{host_group_id}.")
+                    print(f"Could not change ownership of {file}")
+        if count_files:
+            print(f"Changed ownership of {count_files} files back to {host_user_id}:{host_group_id}.")
 
 
 if __name__ == "__main__":