cloudpathlib.S3Path

Bases: CloudPath

Class for representing and operating on AWS S3 URIs, in the style of the Python standard library's pathlib module. Instances represent a path in S3 with filesystem path semantics, and convenient methods allow for basic operations like joining, reading, writing, iterating over contents, etc. This class almost entirely mimics the pathlib.Path interface, so most familiar properties and methods should be available and behave in the expected way.

The S3Client class handles authentication with AWS. If a client instance is not explicitly specified on S3Path instantiation, a default client is used. See S3Client's documentation for more details.

Source code in cloudpathlib/s3/s3path.py

@register_path_class("s3")
class S3Path(CloudPath):
    """Class for representing and operating on AWS S3 URIs, in the style of the Python standard
    library's [`pathlib` module](https://docs.python.org/3/library/pathlib.html). Instances
    represent a path in S3 with filesystem path semantics, and convenient methods allow for basic
    operations like joining, reading, writing, iterating over contents, etc. This class almost
    entirely mimics the [`pathlib.Path`](https://docs.python.org/3/library/pathlib.html#pathlib.Path)
    interface, so most familiar properties and methods should be available and behave in the
    expected way.

    The [`S3Client`](../s3client/) class handles authentication with AWS. If a client instance is
    not explicitly specified on `S3Path` instantiation, a default client is used. See `S3Client`'s
    documentation for more details.
    """

    cloud_prefix: str = "s3://"
    client: "S3Client"

    @property
    def drive(self) -> str:
        return self.bucket

    def mkdir(self, parents=False, exist_ok=False, mode: Optional[Any] = None):
        # not possible to make empty directory on s3
        pass

    def touch(self, exist_ok: bool = True, mode: Optional[Any] = None):
        if self.exists():
            if not exist_ok:
                raise FileExistsError(f"File exists: {self}")
            self.client._move_file(self, self)
        else:
            tf = TemporaryDirectory()
            p = Path(tf.name) / "empty"
            p.touch()

            self.client._upload_file(p, self)

            tf.cleanup()

    def stat(self, follow_symlinks=True):
        try:
            meta = self.client._get_metadata(self)
        except self.client.client.exceptions.NoSuchKey:
            raise NoStatError(
                f"No stats available for {self}; it may be a directory or not exist."
            )

        return os.stat_result(
            (
                None,  # mode
                None,  # ino
                self.cloud_prefix,  # dev,
                None,  # nlink,
                None,  # uid,
                None,  # gid,
                meta.get("size", 0),  # size,
                None,  # atime,
                meta.get("last_modified", 0).timestamp(),  # mtime,
                None,  # ctime,
            )
        )

    @property
    def bucket(self) -> str:
        return self._no_prefix.split("/", 1)[0]

    @property
    def key(self) -> str:
        key = self._no_prefix_no_drive

        # key should never have starting slash for
        # use with boto, etc.
        if key.startswith("/"):
            key = key[1:]

        return key

    @property
    def etag(self):
        return self.client._get_metadata(self).get("etag")

Attributes

anchor: str property

bucket: str property

client: S3Client instance-attribute

cloud_prefix: str = 's3://' class-attribute instance-attribute

drive: str property

etag property

fspath: str property

key: str property

name: str property

parent: Self property

parents: Tuple[Self, ...] property

parser: Self property

parts: Tuple[str, ...] property

stem: str property

suffix: str property

suffixes: List[str] property

Functions

__init__(cloud_path: Union[str, Self, CloudPath], *parts: str, client: Optional[Client] = None) -> None

Source code in cloudpathlib/cloudpath.py

def __init__(
    self,
    cloud_path: Union[str, Self, "CloudPath"],
    *parts: str,
    client: Optional["Client"] = None,
) -> None:
    # handle if local file gets opened. must be set at the top of the method in case any code
    # below raises an exception, this prevents __del__ from raising an AttributeError
    self._handle: Optional[IO] = None
    self._client: Optional["Client"] = None

    if parts:
        # ensure first part ends in "/"; (sometimes it is just prefix, sometimes a longer path)
        if not str(cloud_path).endswith("/"):
            cloud_path = str(cloud_path) + "/"

        cloud_path = str(cloud_path) + "/".join(p.strip("/") for p in parts)

    self.is_valid_cloudpath(cloud_path, raise_on_error=True)
    self._cloud_meta.validate_completeness()

    # versions of the raw string that provide useful methods
    self._str = str(cloud_path)
    self._url = urlparse(self._str)
    self._path = PurePosixPath(f"/{self._no_prefix}")

    # setup client
    if client is None:
        if isinstance(cloud_path, CloudPath):
            self._client = cloud_path.client
    else:
        self._client = client

    if client is not None and not isinstance(client, self._cloud_meta.client_class):
        raise ClientMismatchError(
            f"Client of type [{client.__class__}] is not valid for cloud path of type "
            f"[{self.__class__}]; must be instance of [{self._cloud_meta.client_class}], or "
            f"None to use default client for this cloud path class."
        )

    # track if local has been written to, if so it may need to be uploaded
    self._dirty = False

absolute() -> Self

Source code in cloudpathlib/cloudpath.py

def absolute(self) -> Self:
    return self

as_uri() -> str

Source code in cloudpathlib/cloudpath.py

def as_uri(self) -> str:
    return str(self)

as_url(presign: bool = False, expire_seconds: int = 60 * 60) -> str

Source code in cloudpathlib/cloudpath.py

def as_url(self, presign: bool = False, expire_seconds: int = 60 * 60) -> str:
    if presign:
        url = self.client._generate_presigned_url(self, expire_seconds=expire_seconds)
    else:
        url = self.client._get_public_url(self)
    return url

clear_cache()

Removes cache if it exists

Source code in cloudpathlib/cloudpath.py

def clear_cache(self):
    """Removes cache if it exists"""
    if self._local.exists():
        if self._local.is_file():
            self._local.unlink()
        else:
            shutil.rmtree(self._local)

copy(target: Union[str, os.PathLike, Self], follow_symlinks: bool = True, preserve_metadata: bool = False, force_overwrite_to_cloud: Optional[bool] = None) -> Union[Path, Self]

copy(
    target: Self,
    follow_symlinks=True,
    preserve_metadata=False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Self

copy(
    target: Path,
    follow_symlinks=True,
    preserve_metadata=False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Path

copy(
    target: str,
    follow_symlinks=True,
    preserve_metadata=False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Union[Path, CloudPath]

Copy self to target folder or file, if self is a file.

Source code in cloudpathlib/cloudpath.py

def copy(
    self,
    target: Union[str, os.PathLike, Self],
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Union[Path, Self]:
    """Copy self to target folder or file, if self is a file."""
    return self._copy(
        target,
        follow_symlinks=follow_symlinks,
        preserve_metadata=preserve_metadata,
        force_overwrite_to_cloud=force_overwrite_to_cloud,
        remove_src=False,
    )

copy_into(target_dir: Union[str, os.PathLike, Self], follow_symlinks: bool = True, preserve_metadata: bool = False, force_overwrite_to_cloud: Optional[bool] = None) -> Union[Path, Self]

copy_into(
    target_dir: Self,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Self

copy_into(
    target_dir: Path,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Path

copy_into(
    target_dir: str,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Union[Path, CloudPath]

Copy self into target directory, preserving the filename.

Source code in cloudpathlib/cloudpath.py

def copy_into(
    self,
    target_dir: Union[str, os.PathLike, Self],
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Union[Path, Self]:
    """Copy self into target directory, preserving the filename."""
    target_path = anypath.to_anypath(target_dir) / self.name

    result = self._copy(
        target_path,
        follow_symlinks=follow_symlinks,
        preserve_metadata=preserve_metadata,
        force_overwrite_to_cloud=force_overwrite_to_cloud,
        remove_src=False,
    )
    return cast(Union[Path, Self], result)

copytree(destination, force_overwrite_to_cloud=None, ignore=None)

copytree(
    destination: Self,
    force_overwrite_to_cloud: Optional[bool] = None,
    ignore: Optional[
        Callable[[str, Iterable[str]], Container[str]]
    ] = None,
) -> Self

copytree(
    destination: Path,
    force_overwrite_to_cloud: Optional[bool] = None,
    ignore: Optional[
        Callable[[str, Iterable[str]], Container[str]]
    ] = None,
) -> Path

copytree(
    destination: Union[str, os.PathLike, Self],
    force_overwrite_to_cloud: Optional[bool] = None,
    ignore: Optional[
        Callable[[str, Iterable[str]], Container[str]]
    ] = None,
) -> Union[Path, CloudPath]

Copy self to a directory, if self is a directory.

Source code in cloudpathlib/cloudpath.py

def copytree(self, destination, force_overwrite_to_cloud=None, ignore=None):
    """Copy self to a directory, if self is a directory."""
    if not self.is_dir():
        raise CloudPathNotADirectoryError(
            f"Origin path {self} must be a directory. To copy a single file use the method copy."
        )

    destination = anypath.to_anypath(destination)

    if destination.exists() and destination.is_file():
        raise CloudPathFileExistsError(
            f"Destination path {destination} of copytree must be a directory."
        )

    contents = list(self.iterdir())

    if ignore is not None:
        ignored_names = ignore(self._no_prefix_no_drive, [x.name for x in contents])
    else:
        ignored_names = set()

    destination.mkdir(parents=True, exist_ok=True)

    for subpath in contents:
        if subpath.name in ignored_names:
            continue
        if subpath.is_file():
            subpath.copy(
                destination / subpath.name, force_overwrite_to_cloud=force_overwrite_to_cloud
            )
        elif subpath.is_dir():
            subpath.copytree(
                destination / (subpath.name + ("" if subpath.name.endswith("/") else "/")),
                force_overwrite_to_cloud=force_overwrite_to_cloud,
                ignore=ignore,
            )

    return destination

download_to(destination: Union[str, os.PathLike]) -> Path

Source code in cloudpathlib/cloudpath.py

def download_to(self, destination: Union[str, os.PathLike]) -> Path:
    destination = Path(destination)

    if not self.exists():
        raise CloudPathNotExistsError(f"Cannot download because path does not exist: {self}")

    if self.is_file():
        if destination.is_dir():
            destination = destination / self.name
        return self.client._download_file(self, destination)
    else:
        destination.mkdir(exist_ok=True)
        for f in self.iterdir():
            rel = str(self)
            if not rel.endswith("/"):
                rel = rel + "/"

            rel_dest = str(f)[len(rel) :]
            f.download_to(destination / rel_dest)

        return destination

exists(follow_symlinks=True) -> bool

Source code in cloudpathlib/cloudpath.py

def exists(self, follow_symlinks=True) -> bool:
    return self.client._exists(self)

from_uri(uri: str) -> Self classmethod

Source code in cloudpathlib/cloudpath.py

@classmethod
def from_uri(cls, uri: str) -> Self:
    return cls(uri)

full_match(pattern: str, case_sensitive: Optional[bool] = None) -> bool

Source code in cloudpathlib/cloudpath.py

def full_match(self, pattern: str, case_sensitive: Optional[bool] = None) -> bool:
    if sys.version_info < (3, 13):
        raise NotImplementedError("full_match requires Python 3.13 or higher")

    # strip scheme from start of pattern before testing
    if pattern.startswith(self.anchor + self.drive):
        pattern = pattern[len(self.anchor + self.drive) :]
    elif pattern.startswith(self.anchor):
        # for http paths, keep leading slash
        pattern = pattern[len(self.anchor) - 1 :]

    # remove drive, which is kept on normal dispatch to pathlib
    return PurePosixPath(self._no_prefix_no_drive).full_match(  # type: ignore[attr-defined]
        pattern, case_sensitive=case_sensitive
    )

glob(pattern: Union[str, os.PathLike], case_sensitive: Optional[bool] = None, recurse_symlinks: bool = True) -> Generator[Self, None, None]

Source code in cloudpathlib/cloudpath.py

def glob(
    self,
    pattern: Union[str, os.PathLike],
    case_sensitive: Optional[bool] = None,
    recurse_symlinks: bool = True,
) -> Generator[Self, None, None]:
    pattern = self._glob_checks(pattern)

    pattern_parts = PurePosixPath(pattern).parts
    selector = _make_selector(
        tuple(pattern_parts), _posix_flavour, case_sensitive=case_sensitive
    )

    yield from self._glob(
        selector,
        "/" in pattern
        or "**"
        in pattern,  # recursive listing needed if explicit ** or any sub folder in pattern
    )

info() -> CloudPathInfo

Return a CloudPathInfo object for this path.

Source code in cloudpathlib/cloudpath.py

def info(self) -> "CloudPathInfo":
    """Return a CloudPathInfo object for this path."""
    return CloudPathInfo(self)

is_absolute() -> bool

Source code in cloudpathlib/cloudpath.py

def is_absolute(self) -> bool:
    return True

is_dir(follow_symlinks=True) -> bool

Source code in cloudpathlib/cloudpath.py

def is_dir(self, follow_symlinks=True) -> bool:
    return self.client._is_file_or_dir(self) == "dir"

is_file(follow_symlinks=True) -> bool

Source code in cloudpathlib/cloudpath.py

def is_file(self, follow_symlinks=True) -> bool:
    return self.client._is_file_or_dir(self) == "file"

is_junction()

Source code in cloudpathlib/cloudpath.py

def is_junction(self):
    return False  # only windows paths can be junctions, not cloudpaths

is_relative_to(other: Self) -> bool

Source code in cloudpathlib/cloudpath.py

def is_relative_to(self, other: Self) -> bool:
    try:
        self.relative_to(other)
        return True
    except ValueError:
        return False

is_valid_cloudpath(path: Union[str, CloudPath], raise_on_error: bool = False) -> Union[bool, TypeGuard[Self]] classmethod

is_valid_cloudpath(
    path: CloudPath, raise_on_error: bool = ...
) -> TypeGuard[Self]

is_valid_cloudpath(
    path: str, raise_on_error: bool = ...
) -> bool

Source code in cloudpathlib/cloudpath.py

@classmethod
def is_valid_cloudpath(
    cls, path: Union[str, "CloudPath"], raise_on_error: bool = False
) -> Union[bool, TypeGuard[Self]]:
    valid = str(path).lower().startswith(cls.cloud_prefix.lower())

    if raise_on_error and not valid:
        raise InvalidPrefixError(
            f"'{path}' is not a valid path since it does not start with '{cls.cloud_prefix}'"
        )

    return valid

iterdir() -> Generator[Self, None, None]

Source code in cloudpathlib/cloudpath.py

def iterdir(self) -> Generator[Self, None, None]:
    for f, _ in self.client._list_dir(self, recursive=False):
        if f != self:  # iterdir does not include itself in pathlib
            yield f

joinpath(*pathsegments: Union[str, os.PathLike]) -> Self

Source code in cloudpathlib/cloudpath.py

def joinpath(self, *pathsegments: Union[str, os.PathLike]) -> Self:
    return self._dispatch_to_path("joinpath", *pathsegments)

match(path_pattern: str, case_sensitive: Optional[bool] = None) -> bool

Source code in cloudpathlib/cloudpath.py

def match(self, path_pattern: str, case_sensitive: Optional[bool] = None) -> bool:
    # strip scheme from start of pattern before testing
    if path_pattern.startswith(self.anchor + self.drive + "/"):
        path_pattern = path_pattern[len(self.anchor + self.drive + "/") :]

    kwargs = dict(case_sensitive=case_sensitive)

    if sys.version_info < (3, 12):
        kwargs.pop("case_sensitive")

    return self._dispatch_to_path("match", path_pattern, **kwargs)

mkdir(parents=False, exist_ok=False, mode: Optional[Any] = None)

Source code in cloudpathlib/s3/s3path.py

def mkdir(self, parents=False, exist_ok=False, mode: Optional[Any] = None):
    # not possible to make empty directory on s3
    pass

move(target: Union[str, os.PathLike, Self], follow_symlinks: bool = True, preserve_metadata: bool = False, force_overwrite_to_cloud: Optional[bool] = None) -> Union[Path, Self]

move(
    target: Self,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Self

move(
    target: Path,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Path

move(
    target: str,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Union[Path, CloudPath]

Move self to target location, removing the source.

Source code in cloudpathlib/cloudpath.py

def move(
    self,
    target: Union[str, os.PathLike, Self],
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Union[Path, Self]:
    """Move self to target location, removing the source."""
    return self._copy(
        target,
        follow_symlinks=follow_symlinks,
        preserve_metadata=preserve_metadata,
        force_overwrite_to_cloud=force_overwrite_to_cloud,
        remove_src=True,
    )

move_into(target_dir: Union[str, os.PathLike, Self], follow_symlinks: bool = True, preserve_metadata: bool = False, force_overwrite_to_cloud: Optional[bool] = None) -> Union[Path, Self]

move_into(
    target_dir: Self,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Self

move_into(
    target_dir: Path,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Path

move_into(
    target_dir: str,
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Union[Path, CloudPath]

Move self into target directory, preserving the filename and removing the source.

Source code in cloudpathlib/cloudpath.py

def move_into(
    self,
    target_dir: Union[str, os.PathLike, Self],
    follow_symlinks: bool = True,
    preserve_metadata: bool = False,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Union[Path, Self]:
    """Move self into target directory, preserving the filename and removing the source."""
    target_path = anypath.to_anypath(target_dir) / self.name

    result = self._copy(
        target_path,
        follow_symlinks=follow_symlinks,
        preserve_metadata=preserve_metadata,
        force_overwrite_to_cloud=force_overwrite_to_cloud,
        remove_src=True,
    )
    return cast(Union[Path, Self], result)

open(mode: str = 'r', buffering: int = -1, encoding: Optional[str] = None, errors: Optional[str] = None, newline: Optional[str] = None, force_overwrite_from_cloud: Optional[bool] = None, force_overwrite_to_cloud: Optional[bool] = None, buffer_size: Optional[int] = None) -> IO[Any]

open(
    mode: OpenTextMode = "r",
    buffering: int = -1,
    encoding: Optional[str] = None,
    errors: Optional[str] = None,
    newline: Optional[str] = None,
    force_overwrite_from_cloud: Optional[bool] = None,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> TextIOWrapper

open(
    mode: OpenBinaryMode,
    buffering: Literal[0],
    encoding: None = None,
    errors: None = None,
    newline: None = None,
    force_overwrite_from_cloud: Optional[bool] = None,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> FileIO

open(
    mode: OpenBinaryModeUpdating,
    buffering: Literal[-1, 1] = -1,
    encoding: None = None,
    errors: None = None,
    newline: None = None,
    force_overwrite_from_cloud: Optional[bool] = None,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> BufferedRandom

open(
    mode: OpenBinaryModeWriting,
    buffering: Literal[-1, 1] = -1,
    encoding: None = None,
    errors: None = None,
    newline: None = None,
    force_overwrite_from_cloud: Optional[bool] = None,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> BufferedWriter

open(
    mode: OpenBinaryModeReading,
    buffering: Literal[-1, 1] = -1,
    encoding: None = None,
    errors: None = None,
    newline: None = None,
    force_overwrite_from_cloud: Optional[bool] = None,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> BufferedReader

open(
    mode: OpenBinaryMode,
    buffering: int = -1,
    encoding: None = None,
    errors: None = None,
    newline: None = None,
    force_overwrite_from_cloud: Optional[bool] = None,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> BinaryIO

open(
    mode: str,
    buffering: int = -1,
    encoding: Optional[str] = None,
    errors: Optional[str] = None,
    newline: Optional[str] = None,
    force_overwrite_from_cloud: Optional[bool] = None,
    force_overwrite_to_cloud: Optional[bool] = None,
) -> IO[Any]

Source code in cloudpathlib/cloudpath.py

def open(
    self,
    mode: str = "r",
    buffering: int = -1,
    encoding: Optional[str] = None,
    errors: Optional[str] = None,
    newline: Optional[str] = None,
    force_overwrite_from_cloud: Optional[bool] = None,  # extra kwarg not in pathlib
    force_overwrite_to_cloud: Optional[bool] = None,  # extra kwarg not in pathlib
    buffer_size: Optional[int] = None,  # extra kwarg for streaming mode
) -> "IO[Any]":
    # if trying to call open on a directory that exists
    exists_on_cloud = self.exists()

    if exists_on_cloud and not self.is_file():
        raise CloudPathIsADirectoryError(
            f"Cannot open directory, only files. Tried to open ({self})"
        )

    if not exists_on_cloud and any(m in mode for m in ("r", "a")):
        raise CloudPathFileNotFoundError(
            f"File opened for read or append, but it does not exist on cloud: {self}"
        )

    if mode == "x" and self.exists():
        raise CloudPathFileExistsError(f"Cannot open existing file ({self}) for creation.")

    # Use streaming I/O if file_cache_mode is streaming
    if self.client.file_cache_mode == FileCacheMode.streaming:
        # Import here to keep it localized to streaming functionality
        from .cloud_io import CloudBufferedIO, CloudTextIO

        # Get the raw IO class from the cloud implementation
        raw_io_class = self._cloud_meta.raw_io_class
        if raw_io_class is None:
            raise NotImplementedError(
                f"Streaming I/O is not implemented for {self._cloud_meta.name}"
            )

        # Calculate buffer size from buffering or buffer_size parameter
        if buffer_size is None:
            if buffering == 0:
                # Unbuffered binary mode
                buffer_size = 1  # Minimal buffering
                if "b" not in mode:
                    mode += "b"  # Force binary mode for unbuffered
            elif buffering > 0:
                buffer_size = buffering
            else:
                buffer_size = 64 * 1024  # Default 64 KiB

        # Return appropriate streaming I/O object
        if "b" in mode:
            return CloudBufferedIO(  # type: ignore[return-value]
                raw_io_class=raw_io_class,
                client=self.client,
                cloud_path=self,
                mode=mode,
                buffer_size=buffer_size,
            )
        else:
            return CloudTextIO(  # type: ignore[return-value]
                raw_io_class=raw_io_class,
                client=self.client,
                cloud_path=self,
                mode=mode,
                encoding=encoding,
                errors=errors,
                newline=newline,
                buffer_size=buffer_size,
            )

    # Standard cached mode
    self._refresh_cache(force_overwrite_from_cloud=force_overwrite_from_cloud)

    # create any directories that may be needed if the file is new
    if not self._local.exists():
        self._local.parent.mkdir(parents=True, exist_ok=True)
        original_mtime = 0.0
    else:
        original_mtime = self._local.stat().st_mtime

    buffer = self._local.open(
        mode=mode,
        buffering=buffering,
        encoding=encoding,
        errors=errors,
        newline=newline,
    )

    # write modes need special on closing the buffer
    if any(m in mode for m in ("w", "+", "x", "a")):
        # dirty, handle, patch close
        wrapped_close = buffer.close

        # since we are pretending this is a cloud file, upload it to the cloud
        # when the buffer is closed
        def _patched_close_upload(*args, **kwargs) -> None:
            wrapped_close(*args, **kwargs)

            # we should be idempotent and not upload again if
            # we already ran our close method patch
            if not self._dirty:
                return

            # original mtime should match what was in the cloud; because of system clocks or rounding
            # by the cloud provider, the new version in our cache is "older" than the original version;
            # explicitly set the new modified time to be after the original modified time.
            if self._local.stat().st_mtime < original_mtime:
                new_mtime = original_mtime + 1
                os.utime(self._local, times=(new_mtime, new_mtime))

            self._upload_local_to_cloud(force_overwrite_to_cloud=force_overwrite_to_cloud)
            self._dirty = False

        buffer.close = _patched_close_upload  # type: ignore

        # keep reference in case we need to close when __del__ is called on this object
        self._handle = buffer

        # opened for write, so mark dirty
        self._dirty = True

    # if we don't want any cache around, remove the cache
    # as soon as the file is closed
    if self.client.file_cache_mode == FileCacheMode.close_file:
        # this may be _patched_close_upload, in which case we need to
        # make sure to call that first so the file gets uploaded
        wrapped_close_for_cache = buffer.close

        def _patched_close_empty_cache(*args, **kwargs):
            wrapped_close_for_cache(*args, **kwargs)

            # remove local file as last step on closing
            self.clear_cache()

        buffer.close = _patched_close_empty_cache  # type: ignore

    return buffer

read_bytes() -> bytes

Source code in cloudpathlib/cloudpath.py

def read_bytes(self) -> bytes:
    with self.open(mode="rb") as f:
        return f.read()

read_text(encoding: Optional[str] = None, errors: Optional[str] = None, newline: Optional[str] = None) -> str

Source code in cloudpathlib/cloudpath.py

def read_text(
    self,
    encoding: Optional[str] = None,
    errors: Optional[str] = None,
    newline: Optional[str] = None,
) -> str:
    with self.open(mode="r", encoding=encoding, errors=errors, newline=newline) as f:
        return f.read()

relative_to(other: Self, walk_up: bool = False) -> PurePosixPath

Source code in cloudpathlib/cloudpath.py

def relative_to(self, other: Self, walk_up: bool = False) -> PurePosixPath:
    # We don't dispatch regularly since this never returns a cloud path (since it is relative, and cloud paths are
    # absolute)
    if not isinstance(other, CloudPath):
        raise ValueError(f"{self} is a cloud path, but {other} is not")
    if self.anchor != other.anchor:
        raise ValueError(
            f"{self} is a {self.anchor} path, but {other} is a {other.anchor} path"
        )

    kwargs = dict(walk_up=walk_up)

    if sys.version_info < (3, 12):
        kwargs.pop("walk_up")

    return self._path.relative_to(other._path, **kwargs)  # type: ignore[call-arg]

rename(target: Self) -> Self

Source code in cloudpathlib/cloudpath.py

def rename(self, target: Self) -> Self:
    # for cloud services replace == rename since we don't just rename,
    # we actually move files
    return self.replace(target)

replace(target: Self) -> Self

Source code in cloudpathlib/cloudpath.py

def replace(self, target: Self) -> Self:
    if type(self) is not type(target):
        raise TypeError(
            f"The target based to rename must be an instantiated class of type: {type(self)}"
        )

    if self.is_dir():
        raise CloudPathIsADirectoryError(
            f"Path {self} is a directory; rename/replace the files recursively."
        )

    if target == self:
        # Request is to replace/rename this with the same path - nothing to do
        return self

    if target.exists():
        target.unlink()

    self.client._move_file(self, target)
    return target

resolve(strict: bool = False) -> Self

Source code in cloudpathlib/cloudpath.py

def resolve(self, strict: bool = False) -> Self:
    return self

rglob(pattern: Union[str, os.PathLike], case_sensitive: Optional[bool] = None, recurse_symlinks: bool = True) -> Generator[Self, None, None]

Source code in cloudpathlib/cloudpath.py

def rglob(
    self,
    pattern: Union[str, os.PathLike],
    case_sensitive: Optional[bool] = None,
    recurse_symlinks: bool = True,
) -> Generator[Self, None, None]:
    pattern = self._glob_checks(pattern)

    pattern_parts = PurePosixPath(pattern).parts
    selector = _make_selector(
        ("**",) + tuple(pattern_parts), _posix_flavour, case_sensitive=case_sensitive
    )

    yield from self._glob(selector, True)

rmdir() -> None

Source code in cloudpathlib/cloudpath.py

def rmdir(self) -> None:
    if self.is_file():
        raise CloudPathNotADirectoryError(
            f"Path {self} is a file; call unlink instead of rmdir."
        )
    try:
        next(self.iterdir())
        raise DirectoryNotEmptyError(
            f"Directory not empty: '{self}'. Use rmtree to delete recursively."
        )
    except StopIteration:
        pass
    self.client._remove(self)

rmtree() -> None

Delete an entire directory tree.

Source code in cloudpathlib/cloudpath.py

def rmtree(self) -> None:
    """Delete an entire directory tree."""
    if self.is_file():
        raise CloudPathNotADirectoryError(
            f"Path {self} is a file; call unlink instead of rmtree."
        )
    self.client._remove(self)

samefile(other_path: Union[str, os.PathLike]) -> bool

Source code in cloudpathlib/cloudpath.py

def samefile(self, other_path: Union[str, os.PathLike]) -> bool:
    # all cloud paths are absolute and the paths are used for hash
    return self == other_path

stat(follow_symlinks=True)

Source code in cloudpathlib/s3/s3path.py

def stat(self, follow_symlinks=True):
    try:
        meta = self.client._get_metadata(self)
    except self.client.client.exceptions.NoSuchKey:
        raise NoStatError(
            f"No stats available for {self}; it may be a directory or not exist."
        )

    return os.stat_result(
        (
            None,  # mode
            None,  # ino
            self.cloud_prefix,  # dev,
            None,  # nlink,
            None,  # uid,
            None,  # gid,
            meta.get("size", 0),  # size,
            None,  # atime,
            meta.get("last_modified", 0).timestamp(),  # mtime,
            None,  # ctime,
        )
    )

touch(exist_ok: bool = True, mode: Optional[Any] = None)

Source code in cloudpathlib/s3/s3path.py

def touch(self, exist_ok: bool = True, mode: Optional[Any] = None):
    if self.exists():
        if not exist_ok:
            raise FileExistsError(f"File exists: {self}")
        self.client._move_file(self, self)
    else:
        tf = TemporaryDirectory()
        p = Path(tf.name) / "empty"
        p.touch()

        self.client._upload_file(p, self)

        tf.cleanup()

unlink(missing_ok: bool = True) -> None

Source code in cloudpathlib/cloudpath.py

def unlink(self, missing_ok: bool = True) -> None:
    # Note: missing_ok defaults to False in pathlib, but changing the default now would be a breaking change.
    if self.is_dir():
        raise CloudPathIsADirectoryError(
            f"Path {self} is a directory; call rmdir instead of unlink."
        )
    self.client._remove(self, missing_ok)

upload_from(source: Union[str, os.PathLike], force_overwrite_to_cloud: Optional[bool] = None) -> Self

Upload a file or directory to the cloud path.

Source code in cloudpathlib/cloudpath.py

def upload_from(
    self,
    source: Union[str, os.PathLike],
    force_overwrite_to_cloud: Optional[bool] = None,
) -> Self:
    """Upload a file or directory to the cloud path."""
    source = Path(source)

    if source.is_dir():
        for p in source.iterdir():
            (self / p.name).upload_from(p, force_overwrite_to_cloud=force_overwrite_to_cloud)

        return self

    else:
        if self.exists() and self.is_dir():
            dst = self / source.name
        else:
            dst = self

        dst._upload_file_to_cloud(source, force_overwrite_to_cloud=force_overwrite_to_cloud)

        return dst

validate(v: str) -> Self classmethod

Used as a Pydantic validator. See https://docs.pydantic.dev/2.0/usage/types/custom/

Source code in cloudpathlib/cloudpath.py

@classmethod
def validate(cls, v: str) -> Self:
    """Used as a Pydantic validator. See
    https://docs.pydantic.dev/2.0/usage/types/custom/"""
    return cls(v)

walk(top_down: bool = True, on_error: Optional[Callable] = None, follow_symlinks: bool = False) -> Generator[Tuple[Self, List[str], List[str]], None, None]

Source code in cloudpathlib/cloudpath.py

def walk(
    self,
    top_down: bool = True,
    on_error: Optional[Callable] = None,
    follow_symlinks: bool = False,
) -> Generator[Tuple[Self, List[str], List[str]], None, None]:
    try:
        file_tree = self._build_subtree(recursive=True)  # walking is always recursive
        yield from self._walk_results_from_tree(self, file_tree, top_down=top_down)

    except Exception as e:
        if on_error is not None:
            on_error(e)
        else:
            raise

with_name(name: str) -> Self

Source code in cloudpathlib/cloudpath.py

def with_name(self, name: str) -> Self:
    return self._dispatch_to_path("with_name", name)

with_segments(*pathsegments) -> Self

Create a new CloudPath with the same client out of the given segments. The first segment will be interpreted as the bucket/container name.

Source code in cloudpathlib/cloudpath.py

def with_segments(self, *pathsegments) -> Self:
    """Create a new CloudPath with the same client out of the given segments.
    The first segment will be interpreted as the bucket/container name.
    """
    return self._new_cloudpath("/".join(pathsegments))

with_stem(stem: str) -> Self

Source code in cloudpathlib/cloudpath.py

def with_stem(self, stem: str) -> Self:
    try:
        return self._dispatch_to_path("with_stem", stem)
    except AttributeError:
        # with_stem was only added in python 3.9, so we fallback for compatibility
        return self.with_name(stem + self.suffix)

with_suffix(suffix: str) -> Self

Source code in cloudpathlib/cloudpath.py

def with_suffix(self, suffix: str) -> Self:
    return self._dispatch_to_path("with_suffix", suffix)

write_bytes(data: bytes) -> int

Open the file in bytes mode, write to it, and close the file.

NOTE: vendored from pathlib since we override open https://github.com/python/cpython/blob/3.8/Lib/pathlib.py#L1235-L1242

Source code in cloudpathlib/cloudpath.py

def write_bytes(self, data: bytes) -> int:
    """Open the file in bytes mode, write to it, and close the file.

    NOTE: vendored from pathlib since we override open
    https://github.com/python/cpython/blob/3.8/Lib/pathlib.py#L1235-L1242
    """
    # type-check for the buffer interface before truncating the file
    view = memoryview(data)
    with self.open(mode="wb") as f:
        return f.write(view)

write_text(data: str, encoding: Optional[str] = None, errors: Optional[str] = None, newline: Optional[str] = None) -> int

Open the file in text mode, write to it, and close the file.

NOTE: vendored from pathlib since we override open https://github.com/python/cpython/blob/3.10/Lib/pathlib.py#L1146-L1155

Source code in cloudpathlib/cloudpath.py

def write_text(
    self,
    data: str,
    encoding: Optional[str] = None,
    errors: Optional[str] = None,
    newline: Optional[str] = None,
) -> int:
    """Open the file in text mode, write to it, and close the file.

    NOTE: vendored from pathlib since we override open
    https://github.com/python/cpython/blob/3.10/Lib/pathlib.py#L1146-L1155
    """
    if not isinstance(data, str):
        raise TypeError("data must be str, not %s" % data.__class__.__name__)

    with self.open(mode="w", encoding=encoding, errors=errors, newline=newline) as f:
        return f.write(data)