django/django/core/files/storage.py

import os
import pathlib
from datetime import datetime
from urllib.parse import urljoin

from django.conf import settings
from django.core.exceptions import SuspiciousFileOperation
from django.core.files import File, locks
from django.core.files.move import file_move_safe
from django.core.files.utils import validate_file_name
from django.core.signals import setting_changed
from django.utils import timezone
from django.utils._os import safe_join
from django.utils.crypto import get_random_string
from django.utils.deconstruct import deconstructible
from django.utils.encoding import filepath_to_uri
from django.utils.functional import LazyObject, cached_property
from django.utils.module_loading import import_string
from django.utils.text import get_valid_filename

__all__ = (
    "Storage",
    "FileSystemStorage",
    "DefaultStorage",
    "default_storage",
    "get_storage_class",
)


class Storage:
    """
    A base storage class, providing some default behaviors that all other
    storage systems can inherit or override, as necessary.
    """

    # The following methods represent a public interface to private methods.
    # These shouldn't be overridden by subclasses unless absolutely necessary.

    def open(self, name, mode="rb"):
        """Retrieve the specified file from storage."""
        return self._open(name, mode)

    def save(self, name, content, max_length=None):
        """
        Save new content to the file specified by name. The content should be
        a proper File object or any Python file-like object, ready to be read
        from the beginning.
        """
        # Get the proper name for the file, as it will actually be saved.
        if name is None:
            name = content.name

        if not hasattr(content, "chunks"):
            content = File(content, name)

        name = self.get_available_name(name, max_length=max_length)
        name = self._save(name, content)
        # Ensure that the name returned from the storage system is still valid.
        validate_file_name(name, allow_relative_path=True)
        return name

    # These methods are part of the public API, with default implementations.

    def get_valid_name(self, name):
        """
        Return a filename, based on the provided filename, that's suitable for
        use in the target storage system.
        """
        return get_valid_filename(name)

    def get_alternative_name(self, file_root, file_ext):
        """
        Return an alternative filename, by adding an underscore and a random 7
        character alphanumeric string (before the file extension, if one
        exists) to the filename.
        """
        return "%s_%s%s" % (file_root, get_random_string(7), file_ext)

    def get_available_name(self, name, max_length=None):
        """
        Return a filename that's free on the target storage system and
        available for new content to be written to.
        """
        name = str(name).replace("\\", "/")
        dir_name, file_name = os.path.split(name)
        if ".." in pathlib.PurePath(dir_name).parts:
            raise SuspiciousFileOperation(
                "Detected path traversal attempt in '%s'" % dir_name
            )
        validate_file_name(file_name)
        file_root, file_ext = os.path.splitext(file_name)
        # If the filename already exists, generate an alternative filename
        # until it doesn't exist.
        # Truncate original name if required, so the new filename does not
        # exceed the max_length.
        while self.exists(name) or (max_length and len(name) > max_length):
            # file_ext includes the dot.
            name = os.path.join(
                dir_name, self.get_alternative_name(file_root, file_ext)
            )
            if max_length is None:
                continue
            # Truncate file_root if max_length exceeded.
            truncation = len(name) - max_length
            if truncation > 0:
                file_root = file_root[:-truncation]
                # Entire file_root was truncated in attempt to find an
                # available filename.
                if not file_root:
                    raise SuspiciousFileOperation(
                        'Storage can not find an available filename for "%s". '
                        "Please make sure that the corresponding file field "
                        'allows sufficient "max_length".' % name
                    )
                name = os.path.join(
                    dir_name, self.get_alternative_name(file_root, file_ext)
                )
        return name

    def generate_filename(self, filename):
        """
        Validate the filename by calling get_valid_name() and return a filename
        to be passed to the save() method.
        """
        filename = str(filename).replace("\\", "/")
        # `filename` may include a path as returned by FileField.upload_to.
        dirname, filename = os.path.split(filename)
        if ".." in pathlib.PurePath(dirname).parts:
            raise SuspiciousFileOperation(
                "Detected path traversal attempt in '%s'" % dirname
            )
        return os.path.normpath(os.path.join(dirname, self.get_valid_name(filename)))

    def path(self, name):
        """
        Return a local filesystem path where the file can be retrieved using
        Python's built-in open() function. Storage systems that can't be
        accessed using open() should *not* implement this method.
        """
        raise NotImplementedError("This backend doesn't support absolute paths.")

    # The following methods form the public API for storage systems, but with
    # no default implementations. Subclasses must implement *all* of these.

    def delete(self, name):
        """
        Delete the specified file from the storage system.
        """
        raise NotImplementedError(
            "subclasses of Storage must provide a delete() method"
        )

    def exists(self, name):
        """
        Return True if a file referenced by the given name already exists in the
        storage system, or False if the name is available for a new file.
        """
        raise NotImplementedError(
            "subclasses of Storage must provide an exists() method"
        )

    def listdir(self, path):
        """
        List the contents of the specified path. Return a 2-tuple of lists:
        the first item being directories, the second item being files.
        """
        raise NotImplementedError(
            "subclasses of Storage must provide a listdir() method"
        )

    def size(self, name):
        """
        Return the total size, in bytes, of the file specified by name.
        """
        raise NotImplementedError("subclasses of Storage must provide a size() method")

    def url(self, name):
        """
        Return an absolute URL where the file's contents can be accessed
        directly by a web browser.
        """
        raise NotImplementedError("subclasses of Storage must provide a url() method")

    def get_accessed_time(self, name):
        """
        Return the last accessed time (as a datetime) of the file specified by
        name. The datetime will be timezone-aware if USE_TZ=True.
        """
        raise NotImplementedError(
            "subclasses of Storage must provide a get_accessed_time() method"
        )

    def get_created_time(self, name):
        """
        Return the creation time (as a datetime) of the file specified by name.
        The datetime will be timezone-aware if USE_TZ=True.
        """
        raise NotImplementedError(
            "subclasses of Storage must provide a get_created_time() method"
        )

    def get_modified_time(self, name):
        """
        Return the last modified time (as a datetime) of the file specified by
        name. The datetime will be timezone-aware if USE_TZ=True.
        """
        raise NotImplementedError(
            "subclasses of Storage must provide a get_modified_time() method"
        )


@deconstructible
class FileSystemStorage(Storage):
    """
    Standard filesystem storage
    """

    # The combination of O_CREAT and O_EXCL makes os.open() raise OSError if
    # the file already exists before it's opened.
    OS_OPEN_FLAGS = os.O_WRONLY | os.O_CREAT | os.O_EXCL | getattr(os, "O_BINARY", 0)

    def __init__(
        self,
        location=None,
        base_url=None,
        file_permissions_mode=None,
        directory_permissions_mode=None,
    ):
        self._location = location
        self._base_url = base_url
        self._file_permissions_mode = file_permissions_mode
        self._directory_permissions_mode = directory_permissions_mode
        setting_changed.connect(self._clear_cached_properties)

    def _clear_cached_properties(self, setting, **kwargs):
        """Reset setting based property values."""
        if setting == "MEDIA_ROOT":
            self.__dict__.pop("base_location", None)
            self.__dict__.pop("location", None)
        elif setting == "MEDIA_URL":
            self.__dict__.pop("base_url", None)
        elif setting == "FILE_UPLOAD_PERMISSIONS":
            self.__dict__.pop("file_permissions_mode", None)
        elif setting == "FILE_UPLOAD_DIRECTORY_PERMISSIONS":
            self.__dict__.pop("directory_permissions_mode", None)

    def _value_or_setting(self, value, setting):
        return setting if value is None else value

    @cached_property
    def base_location(self):
        return self._value_or_setting(self._location, settings.MEDIA_ROOT)

    @cached_property
    def location(self):
        return os.path.abspath(self.base_location)

    @cached_property
    def base_url(self):
        if self._base_url is not None and not self._base_url.endswith("/"):
            self._base_url += "/"
        return self._value_or_setting(self._base_url, settings.MEDIA_URL)

    @cached_property
    def file_permissions_mode(self):
        return self._value_or_setting(
            self._file_permissions_mode, settings.FILE_UPLOAD_PERMISSIONS
        )

    @cached_property
    def directory_permissions_mode(self):
        return self._value_or_setting(
            self._directory_permissions_mode, settings.FILE_UPLOAD_DIRECTORY_PERMISSIONS
        )

    def _open(self, name, mode="rb"):
        return File(open(self.path(name), mode))

    def _save(self, name, content):
        full_path = self.path(name)

        # Create any intermediate directories that do not exist.
        directory = os.path.dirname(full_path)
        try:
            if self.directory_permissions_mode is not None:
                # Set the umask because os.makedirs() doesn't apply the "mode"
                # argument to intermediate-level directories.
                old_umask = os.umask(0o777 & ~self.directory_permissions_mode)
                try:
                    os.makedirs(
                        directory, self.directory_permissions_mode, exist_ok=True
                    )
                finally:
                    os.umask(old_umask)
            else:
                os.makedirs(directory, exist_ok=True)
        except FileExistsError:
            raise FileExistsError("%s exists and is not a directory." % directory)

        # There's a potential race condition between get_available_name and
        # saving the file; it's possible that two threads might return the
        # same name, at which point all sorts of fun happens. So we need to
        # try to create the file, but if it already exists we have to go back
        # to get_available_name() and try again.

        while True:
            try:
                # This file has a file path that we can move.
                if hasattr(content, "temporary_file_path"):
                    file_move_safe(content.temporary_file_path(), full_path)

                # This is a normal uploadedfile that we can stream.
                else:
                    # The current umask value is masked out by os.open!
                    fd = os.open(full_path, self.OS_OPEN_FLAGS, 0o666)
                    _file = None
                    try:
                        locks.lock(fd, locks.LOCK_EX)
                        for chunk in content.chunks():
                            if _file is None:
                                mode = "wb" if isinstance(chunk, bytes) else "wt"
                                _file = os.fdopen(fd, mode)
                            _file.write(chunk)
                    finally:
                        locks.unlock(fd)
                        if _file is not None:
                            _file.close()
                        else:
                            os.close(fd)
            except FileExistsError:
                # A new name is needed if the file exists.
                name = self.get_available_name(name)
                full_path = self.path(name)
            else:
                # OK, the file save worked. Break out of the loop.
                break

        if self.file_permissions_mode is not None:
            os.chmod(full_path, self.file_permissions_mode)

        # Ensure the saved path is always relative to the storage root.
        name = os.path.relpath(full_path, self.location)
        # Store filenames with forward slashes, even on Windows.
        return str(name).replace("\\", "/")

    def delete(self, name):
        if not name:
            raise ValueError("The name must be given to delete().")
        name = self.path(name)
        # If the file or directory exists, delete it from the filesystem.
        try:
            if os.path.isdir(name):
                os.rmdir(name)
            else:
                os.remove(name)
        except FileNotFoundError:
            # FileNotFoundError is raised if the file or directory was removed
            # concurrently.
            pass

    def exists(self, name):
        return os.path.lexists(self.path(name))

    def listdir(self, path):
        path = self.path(path)
        directories, files = [], []
        with os.scandir(path) as entries:
            for entry in entries:
                if entry.is_dir():
                    directories.append(entry.name)
                else:
                    files.append(entry.name)
        return directories, files

    def path(self, name):
        return safe_join(self.location, name)

    def size(self, name):
        return os.path.getsize(self.path(name))

    def url(self, name):
        if self.base_url is None:
            raise ValueError("This file is not accessible via a URL.")
        url = filepath_to_uri(name)
        if url is not None:
            url = url.lstrip("/")
        return urljoin(self.base_url, url)

    def _datetime_from_timestamp(self, ts):
        """
        If timezone support is enabled, make an aware datetime object in UTC;
        otherwise make a naive one in the local timezone.
        """
        tz = timezone.utc if settings.USE_TZ else None
        return datetime.fromtimestamp(ts, tz=tz)

    def get_accessed_time(self, name):
        return self._datetime_from_timestamp(os.path.getatime(self.path(name)))

    def get_created_time(self, name):
        return self._datetime_from_timestamp(os.path.getctime(self.path(name)))

    def get_modified_time(self, name):
        return self._datetime_from_timestamp(os.path.getmtime(self.path(name)))


def get_storage_class(import_path=None):
    return import_string(import_path or settings.DEFAULT_FILE_STORAGE)


class DefaultStorage(LazyObject):
    def _setup(self):
        self._wrapped = get_storage_class()()


default_storage = DefaultStorage()