[commands/cache] make pip cache purge remove everything from http + wheels caches; make pip cache remove prune empty directories.

Author: duckinator

src/pip/_internal/commands/cache.py

@@ -1,3 +1,4 @@
+import itertools
 import os
 import textwrap
 from optparse import Values
@@ -184,7 +185,23 @@ def remove_cache_items(self, options: Values, args: List[Any]) -> None:
         for filename in files:
             os.unlink(filename)
             logger.verbose("Removed %s", filename)
+
+        http_dirs = filesystem.subdirs_with_no_files(self._cache_dir(options, "http"))
+        wheel_dirs = filesystem.subdirs_with_no_files(
+            self._cache_dir(options, "wheels")
+        )
+        for dirname in itertools.chain(http_dirs, wheel_dirs):
+            os.rmdir(dirname)
+            logger.verbose("Removed %s", dirname)
+
+        # selfcheck.json is no longer used by pip.
+        selfcheck_json = self._cache_dir(options, "selfcheck.json")
+        if os.path.isfile(selfcheck_json):
+            os.remove(selfcheck_json)
+            logger.verbose("Removed legacy selfcheck.json file")
+
         logger.info("Files removed: %s", len(files))
+        logger.info("Empty directories removed: %s", len(http_dirs) + len(wheel_dirs))
 
     def purge_cache(self, options: Values, args: List[Any]) -> None:
         if args:

src/pip/_internal/utils/filesystem.py

@@ -4,11 +4,13 @@
 import random
 import sys
 from contextlib import contextmanager
+from pathlib import Path
 from tempfile import NamedTemporaryFile
 from typing import Any, BinaryIO, Generator, List, Union, cast
 
 from pip._vendor.tenacity import retry, stop_after_delay, wait_fixed
 
+from pip._internal.exceptions import PipError
 from pip._internal.utils.compat import get_path_uid
 from pip._internal.utils.misc import format_size
 
@@ -151,3 +153,53 @@ def directory_size(path: str) -> Union[int, float]:
 
 def format_directory_size(path: str) -> str:
     return format_size(directory_size(path))
+
+
+def _leaf_subdirs(path: Path) -> Generator[Path, None, None]:
+    """Traverses the file tree, finding every empty directory."""
+
+    path_obj = Path(path)
+
+    if not path_obj.exists():
+        return
+
+    for item in path_obj.iterdir():
+        if not item.is_dir():
+            continue
+
+        subitems = item.iterdir()
+
+        # ASSUMPTION: Nothing in subitems will be None or False.
+        if not any(subitems):
+            yield item
+
+        if not any(subitem.is_file() for subitem in subitems):
+            yield from _leaf_subdirs(item)
+
+
+def _leaf_parents_without_files(path: Path, leaf: Path) -> Generator[str, None, None]:
+    """Yields +leaf+ and each parent directory below +path+, until one of
+    them includes a file (as opposed to directories or nothing)."""
+
+    if not str(leaf).startswith(str(path)):
+        # If +leaf+ is not a subdirectory of +path+, bail early to avoid
+        # an endless loop.
+        raise PipError("leaf is not a subdirectory of path")
+
+    path = Path(path)
+    leaf = Path(leaf)
+    while leaf != path:
+        if all(item.is_dir() for item in leaf.iterdir()):
+            yield str(leaf)
+        else:
+            break
+        leaf = leaf.parent
+
+
+def subdirs_with_no_files(path_str: str) -> Generator[str, None, None]:
+    """Yields every subdirectory of +path_str+ that has no files under it."""
+
+    path = Path(path_str)
+
+    for leaf in _leaf_subdirs(path):
+        yield from _leaf_parents_without_files(path, leaf)