Repository URL to install this package:
|
Version:
5.1.0 ▾
|
import errno
import logging
import os
import shutil
import stat
import sys
import threading
from collections.abc import Collection, Iterator
from concurrent import futures
from contextlib import contextmanager, suppress
from secrets import token_urlsafe
from typing import TYPE_CHECKING, Any, Optional, Union
from fsspec.callbacks import DEFAULT_CALLBACK
from dvc_objects.executors import ThreadPoolExecutor
from . import system
from .callbacks import wrap_file
if TYPE_CHECKING:
from fsspec import Callback
from .base import AnyFSPath, FileSystem
logger = logging.getLogger(__name__)
umask = os.umask(0)
os.umask(umask)
LOCAL_CHUNK_SIZE = 2**20 # 1 MB
COPY_PBAR_MIN_SIZE = 2**30 # 1 GB
def is_exec(mode: int) -> bool:
return bool(mode & (stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH))
def relpath(path: "AnyFSPath", start: "AnyFSPath" = os.curdir) -> "AnyFSPath":
path = os.fspath(path)
start = os.path.abspath(os.fspath(start))
# Windows path on different drive than curdir doesn't have relpath
if os.name == "nt":
# Since python 3.8 os.realpath resolves network shares to their UNC
# path. So, to be certain that relative paths correctly captured,
# we need to resolve to UNC path first. We resolve only the drive
# name so that we don't follow any 'real' symlinks on the path
def resolve_network_drive_windows(path_to_resolve):
drive, tail = os.path.splitdrive(path_to_resolve)
return os.path.join(os.path.realpath(drive), tail)
path = resolve_network_drive_windows(os.path.abspath(path))
start = resolve_network_drive_windows(start)
if not os.path.commonprefix([start, path]):
return path
return os.path.relpath(path, start)
def move(src: "AnyFSPath", dst: "AnyFSPath") -> None:
"""Atomically move src to dst and chmod it with mode.
Moving is performed in two stages to make the whole operation atomic in
case src and dst are on different filesystems and actual physical copying
of data is happening.
"""
dst = os.path.abspath(dst)
tmp = tmp_fname(dst)
if os.path.islink(src):
shutil.copy(src, tmp)
_unlink(src, _chmod)
else:
shutil.move(src, tmp)
shutil.move(tmp, dst)
def _chmod(func, p, excinfo):
perm = os.lstat(p).st_mode
perm |= stat.S_IWRITE
try:
os.chmod(p, perm)
except OSError as exc:
# broken symlink or file is not owned by us
if exc.errno not in [errno.ENOENT, errno.EPERM]:
raise
func(p)
def _unlink(path: "AnyFSPath", onerror):
try:
os.unlink(path)
except OSError:
onerror(os.unlink, path, sys.exc_info())
def remove(path: "AnyFSPath") -> None:
logger.debug("Removing '%s'", path)
try:
if os.path.isdir(path):
shutil.rmtree(path, onerror=_chmod)
else:
_unlink(path, _chmod)
except OSError as exc:
if exc.errno != errno.ENOENT:
raise
def makedirs(path, exist_ok: bool = False, mode: Optional[int] = None) -> None:
if mode is None:
os.makedirs(path, exist_ok=exist_ok)
return
# Modified version of os.makedirs() with support for extended mode
# (e.g. S_ISGID)
head, tail = os.path.split(path)
if not tail:
head, tail = os.path.split(head)
if head and tail and not os.path.exists(head):
try:
makedirs(head, exist_ok=exist_ok, mode=mode)
except FileExistsError:
# Defeats race condition when another thread created the path
pass
cdir = os.curdir
if isinstance(tail, bytes):
cdir = bytes(os.curdir, "ASCII") # type: ignore[assignment]
if tail == cdir: # xxx/newdir/. exists if xxx/newdir exists
return
try:
os.mkdir(path, mode)
except OSError:
# Cannot rely on checking for EEXIST, since the operating system
# could give priority to other errors like EACCES or EROFS
if not exist_ok or not os.path.isdir(path):
raise
try:
os.chmod(path, mode)
except OSError:
logger.debug( # type: ignore[attr-defined]
"failed to chmod '%o' '%s'", mode, path, exc_info=True
)
def copyfile(
src: "AnyFSPath",
dest: "AnyFSPath",
callback: "Callback" = DEFAULT_CALLBACK,
) -> None:
"""Copy file with progress bar"""
if os.path.isdir(dest):
dest = os.path.join(dest, os.path.basename(src))
try:
system.reflink(src, dest)
# NOTE: reflink may or may not clone src permissions
os.chmod(dest, 0o666 & ~umask)
return
except OSError:
pass
total = os.path.getsize(src)
if total < COPY_PBAR_MIN_SIZE:
shutil.copyfile(src, dest)
return
callback.set_size(total)
with open(src, "rb") as fsrc, open(dest, "wb+") as fdest:
wrapped = wrap_file(fsrc, callback)
shutil.copyfileobj(wrapped, fdest, length=LOCAL_CHUNK_SIZE)
def tmp_fname(prefix: str = "") -> str:
"""Temporary name for a partial download"""
return f"{prefix}.{token_urlsafe(16)}.tmp"
@contextmanager
def as_atomic(
fs: "FileSystem", to_info: "AnyFSPath", create_parents: bool = False
) -> Iterator["AnyFSPath"]:
parent = fs.parent(to_info)
if create_parents:
fs.makedirs(parent, exist_ok=True)
tmp_info = fs.join(parent, tmp_fname())
try:
yield tmp_info
except BaseException:
# Handle stuff like KeyboardInterrupt
# as well as other errors that might
# arise during file transfer.
with suppress(FileNotFoundError):
fs.remove(tmp_info)
raise
else:
fs.move(tmp_info, to_info)
def exists(
fs: "FileSystem",
file_paths: Union["AnyFSPath", Collection["AnyFSPath"]],
callback: "Callback" = DEFAULT_CALLBACK,
batch_size: Optional[int] = None,
) -> dict[str, bool]:
"""Return batched fs.exists results.
Runs batched fs.exists() calls in parallel with fs.ls() until all paths
have been checked.
"""
if not file_paths:
return {}
paths = {file_paths} if isinstance(file_paths, str) else set(file_paths)
if len(paths) == 1:
path = paths.pop()
return {path: fs.exists(path)}
paths_lock = threading.Lock()
results: dict[str, bool] = {}
results_lock = threading.Lock()
callback.set_size(len(paths))
jobs = batch_size or fs.jobs
exists_jobs = jobs - 1 if jobs > 1 else 1
executor = ThreadPoolExecutor(max_workers=2, cancel_on_error=True)
logger.debug("Querying status for '%d' files", len(paths))
exist_fut = executor.submit(
_exist_query,
fs,
paths,
paths_lock,
results,
results_lock,
exists_jobs,
callback,
)
list_fut = executor.submit(
_list_query,
fs,
paths,
paths_lock,
results,
results_lock,
callback,
)
_done, not_done = futures.wait(
[exist_fut, list_fut], return_when=futures.FIRST_COMPLETED
)
for fut in not_done:
fut.cancel()
# NOTE: if we started a long running lsdir it will continue to run in
# the background until the task completes
executor.shutdown(wait=False)
return results
def _exist_query(
fs: "FileSystem",
paths: set["AnyFSPath"],
paths_lock: threading.Lock,
results: dict[str, bool],
results_lock: threading.Lock,
batch_size: int,
callback: "Callback",
):
while True:
with paths_lock:
if not paths:
return
batch = [paths.pop() for _ in range(batch_size) if paths]
for i, result in enumerate(fs.exists(batch, batch_size=batch_size)):
with results_lock:
path = batch[i]
if path not in results:
results[path] = result
callback.relative_update()
def _list_query(
fs: "FileSystem",
paths: set["AnyFSPath"],
paths_lock: threading.Lock,
results: dict[str, bool],
results_lock: threading.Lock,
callback: "Callback",
):
with paths_lock:
parents = {fs.parent(path) for path in paths}
for parent in parents:
with paths_lock:
if not paths:
return
kwargs: dict[str, Any] = {}
if fs.version_aware:
kwargs["versions"] = True
contents = fs.ls(parent, **kwargs)
with paths_lock:
exist_paths = set()
for path in contents:
if path in paths:
paths.remove(path)
exist_paths.add(path)
with results_lock:
for path in exist_paths:
if path not in results:
results[path] = True
callback.relative_update()
with paths_lock, results_lock:
while paths:
path = paths.pop()
if path not in results:
results[path] = False
callback.relative_update()