#
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements. See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership. The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on an
# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
"""This module contains a Google Cloud Storage operator."""
from __future__ import annotations
import warnings
from typing import TYPE_CHECKING, Sequence
from airflow.exceptions import AirflowException, AirflowProviderDeprecationWarning
from airflow.models import BaseOperator
from airflow.providers.google.cloud.hooks.gcs import GCSHook
if TYPE_CHECKING:
from airflow.utils.context import Context
[docs]class GCSToGCSOperator(BaseOperator):
"""
Copies objects from a bucket to another, with renaming if requested.
.. seealso::
For more information on how to use this operator, take a look at the guide:
:ref:`howto/operator:GCSToGCSOperator`
:param source_bucket: The source Google Cloud Storage bucket where the
object is. (templated)
:param source_object: The source name of the object to copy in the Google cloud
storage bucket. (templated)
You can use only one wildcard for objects (filenames) within your
bucket. The wildcard can appear inside the object name or at the
end of the object name. Appending a wildcard to the bucket name is
unsupported.
:param source_objects: A list of source name of the objects to copy in the Google cloud
storage bucket. (templated)
:param destination_bucket: The destination Google Cloud Storage bucket
where the object should be. If the destination_bucket is None, it defaults
to source_bucket. (templated)
:param destination_object: The destination name of the object in the
destination Google Cloud Storage bucket. (templated)
If a wildcard is supplied in the source_object argument, this is the
prefix that will be prepended to the final destination objects' paths.
Note that the source path's part before the wildcard will be removed;
if it needs to be retained it should be appended to destination_object.
For example, with prefix ``foo/*`` and destination_object ``blah/``, the
file ``foo/baz`` will be copied to ``blah/baz``; to retain the prefix write
the destination_object as e.g. ``blah/foo``, in which case the copied file
will be named ``blah/foo/baz``.
The same thing applies to source objects inside source_objects.
:param move_object: When move object is True, the object is moved instead
of copied to the new location. This is the equivalent of a mv command
as opposed to a cp command.
:param replace: Whether you want to replace existing destination files or not.
:param delimiter: (Deprecated) This is used to restrict the result to only the 'files' in a given
'folder'. If source_objects = ['foo/bah/'] and delimiter = '.avro', then only the 'files' in the
folder 'foo/bah/' with '.avro' delimiter will be copied to the destination object.
:param gcp_conn_id: (Optional) The connection ID used to connect to Google Cloud.
:param last_modified_time: When specified, the objects will be copied or moved,
only if they were modified after last_modified_time.
If tzinfo has not been set, UTC will be assumed.
:param maximum_modified_time: When specified, the objects will be copied or moved,
only if they were modified before maximum_modified_time.
If tzinfo has not been set, UTC will be assumed.
:param is_older_than: When specified, the objects will be copied if they are older
than the specified time in seconds.
:param impersonation_chain: Optional service account to impersonate using short-term
credentials, or chained list of accounts required to get the access_token
of the last account in the list, which will be impersonated in the request.
If set as a string, the account must grant the originating account
the Service Account Token Creator IAM role.
If set as a sequence, the identities from the list must grant
Service Account Token Creator IAM role to the directly preceding identity, with first
account from the list granting this role to the originating account (templated).
:param source_object_required: Whether you want to raise an exception when the source object
doesn't exist. It doesn't have any effect when the source objects are folders or patterns.
:param exact_match: When specified, only exact match of the source object (filename) will be
copied.
:param match_glob: (Optional) filters objects based on the glob pattern given by the string (
e.g, ``'**/*/.json'``)
:Example:
The following Operator would copy a single file named
``sales/sales-2017/january.avro`` in the ``data`` bucket to the file named
``copied_sales/2017/january-backup.avro`` in the ``data_backup`` bucket ::
copy_single_file = GCSToGCSOperator(
task_id="copy_single_file",
source_bucket="data",
source_objects=["sales/sales-2017/january.avro"],
destination_bucket="data_backup",
destination_object="copied_sales/2017/january-backup.avro",
exact_match=True,
gcp_conn_id=google_cloud_conn_id,
)
The following Operator would copy all the Avro files from ``sales/sales-2017``
folder (i.e. all files with names starting with that prefix) in ``data`` bucket to the
``copied_sales/2017`` folder in the ``data_backup`` bucket. ::
copy_files = GCSToGCSOperator(
task_id='copy_files',
source_bucket='data',
source_objects=['sales/sales-2017'],
destination_bucket='data_backup',
destination_object='copied_sales/2017/',
match_glob='**/*.avro'
gcp_conn_id=google_cloud_conn_id
)
Or ::
copy_files = GCSToGCSOperator(
task_id='copy_files',
source_bucket='data',
source_object='sales/sales-2017/*.avro',
destination_bucket='data_backup',
destination_object='copied_sales/2017/',
gcp_conn_id=google_cloud_conn_id
)
The following Operator would move all the Avro files from ``sales/sales-2017``
folder (i.e. all files with names starting with that prefix) in ``data`` bucket to the
same folder in the ``data_backup`` bucket, deleting the original files in the
process. ::
move_files = GCSToGCSOperator(
task_id="move_files",
source_bucket="data",
source_object="sales/sales-2017/*.avro",
destination_bucket="data_backup",
move_object=True,
gcp_conn_id=google_cloud_conn_id,
)
The following Operator would move all the Avro files from ``sales/sales-2019``
and ``sales/sales-2020`` folder in ``data`` bucket to the same folder in the
``data_backup`` bucket, deleting the original files in the process. ::
move_files = GCSToGCSOperator(
task_id="move_files",
source_bucket="data",
source_objects=["sales/sales-2019/*.avro", "sales/sales-2020"],
destination_bucket="data_backup",
delimiter=".avro",
move_object=True,
gcp_conn_id=google_cloud_conn_id,
)
"""
[docs] template_fields: Sequence[str] = (
"source_bucket",
"source_object",
"source_objects",
"destination_bucket",
"destination_object",
"delimiter",
"impersonation_chain",
)
def __init__(
self,
*,
source_bucket,
source_object=None,
source_objects=None,
destination_bucket=None,
destination_object=None,
delimiter=None,
move_object=False,
replace=True,
gcp_conn_id="google_cloud_default",
last_modified_time=None,
maximum_modified_time=None,
is_older_than=None,
impersonation_chain: str | Sequence[str] | None = None,
source_object_required=False,
exact_match=False,
match_glob: str | None = None,
**kwargs,
):
super().__init__(**kwargs)
self.source_bucket = source_bucket
if source_object and WILDCARD in source_object:
warnings.warn(
"Usage of wildcard (*) in 'source_object' is deprecated, utilize 'match_glob' instead",
AirflowProviderDeprecationWarning,
stacklevel=2,
)
self.source_object = source_object
if source_objects and any(WILDCARD in obj for obj in source_objects):
warnings.warn(
"Usage of wildcard (*) in 'source_objects' is deprecated, utilize 'match_glob' instead",
AirflowProviderDeprecationWarning,
stacklevel=2,
)
self.source_objects = source_objects
self.destination_bucket = destination_bucket
self.destination_object = destination_object
if delimiter:
warnings.warn(
"Usage of 'delimiter' is deprecated, please use 'match_glob' instead",
AirflowProviderDeprecationWarning,
stacklevel=2,
)
self.delimiter = delimiter
self.move_object = move_object
self.replace = replace
self.gcp_conn_id = gcp_conn_id
self.last_modified_time = last_modified_time
self.maximum_modified_time = maximum_modified_time
self.is_older_than = is_older_than
self.impersonation_chain = impersonation_chain
self.source_object_required = source_object_required
self.exact_match = exact_match
self.match_glob = match_glob
self.resolved_source_objects: set[str] = set()
self.resolved_target_objects: set[str] = set()
[docs] def execute(self, context: Context):
hook = GCSHook(
gcp_conn_id=self.gcp_conn_id,
impersonation_chain=self.impersonation_chain,
)
if self.source_objects and self.source_object:
error_msg = (
f"You can either set source_object parameter or source_objects parameter but not both. "
f"Found source_object={self.source_object} and source_objects={self.source_objects}"
)
raise AirflowException(error_msg)
if not self.source_object and not self.source_objects:
error_msg = "You must set source_object parameter or source_objects parameter. None set"
raise AirflowException(error_msg)
if self.source_objects and not all(isinstance(item, str) for item in self.source_objects):
raise AirflowException("At least, one of the `objects` in the `source_objects` is not a string")
# If source_object is set, default it to source_objects
if self.source_object:
self.source_objects = [self.source_object]
if self.destination_bucket is None:
self.log.warning(
"destination_bucket is None. Defaulting it to source_bucket (%s)", self.source_bucket
)
self.destination_bucket = self.source_bucket
# An empty source_object means to copy all files
if len(self.source_objects) == 0:
self.source_objects = [""]
# Raise exception if empty string `''` is used twice in source_object, this is to avoid double copy
if self.source_objects.count("") > 1:
raise AirflowException("You can't have two empty strings inside source_object")
# Iterate over the source_objects and do the copy
for prefix in self.source_objects:
# Check if prefix contains wildcard
if WILDCARD in prefix:
self._copy_source_with_wildcard(hook=hook, prefix=prefix)
# Now search with prefix using provided delimiter if any
else:
self._copy_source_without_wildcard(hook=hook, prefix=prefix)
def _ignore_existing_files(self, hook, prefix, **kwargs):
# list all files in the Destination GCS bucket
# and only keep those files which are present in
# Source GCS bucket and not in Destination GCS bucket
delimiter = kwargs.get("delimiter")
match_glob = kwargs.get("match_glob")
objects = kwargs.get("objects")
if self.destination_object is None:
existing_objects = hook.list(
self.destination_bucket, prefix=prefix, delimiter=delimiter, match_glob=match_glob
)
else:
self.log.info("Replaced destination_object with source_object prefix.")
destination_objects = hook.list(
self.destination_bucket,
prefix=self.destination_object,
delimiter=delimiter,
match_glob=match_glob,
)
existing_objects = [
dest_object.replace(self.destination_object, prefix, 1) for dest_object in destination_objects
]
objects = set(objects) - set(existing_objects)
if objects:
self.log.info("%s files are going to be synced: %s.", len(objects), objects)
else:
self.log.info("There are no new files to sync. Have a nice day!")
return objects
def _copy_source_without_wildcard(self, hook, prefix):
"""
List all files in source_objects, copy files to destination_object, and rename each source file.
For source_objects with no wildcard, this operator would first list
all files in source_objects, using provided delimiter if any. Then copy
files from source_objects to destination_object and rename each source
file. Note that if the flag exact_match=False, then each item in the source_objects
(or source_object itself) will be considered as a prefix for the source objects search.
Example 1:
The following Operator would copy all the files from ``a/`` folder
(i.e a/a.csv, a/b.csv, a/c.csv) in ``data`` bucket to the ``b/`` folder in
the ``data_backup`` bucket (b/a.csv, b/b.csv, b/c.csv) ::
copy_files = GCSToGCSOperator(
task_id="copy_files_without_wildcard",
source_bucket="data",
source_objects=["a/"],
destination_bucket="data_backup",
destination_object="b/",
gcp_conn_id=google_cloud_conn_id,
)
Example 2:
The following Operator would copy all avro files from ``a/`` folder
(i.e a/a.avro, a/b.avro, a/c.avro) in ``data`` bucket to the ``b/`` folder in
the ``data_backup`` bucket (b/a.avro, b/b.avro, b/c.avro) ::
copy_files = GCSToGCSOperator(
task_id="copy_files_without_wildcard",
source_bucket="data",
source_objects=["a/"],
destination_bucket="data_backup",
destination_object="b/",
delimiter=".avro",
gcp_conn_id=google_cloud_conn_id,
)
Example 3:
The following Operator would copy files (a/file_1.txt, a/file_2.csv, a/file_3.avro)
in ``data`` bucket to the ``b/`` folder in
the ``data_backup`` bucket (b/file_1.txt, b/file_2.csv, b/file_3.avro) ::
copy_files = GCSToGCSOperator(
task_id="copy_files_without_wildcard",
source_bucket="data",
source_objects=["a/file_1.txt", "a/file_2.csv", "a/file_3.avro"],
destination_bucket="data_backup",
destination_object="b/",
gcp_conn_id=google_cloud_conn_id,
)
Example 4:
The following Operator would copy files corresponding to the prefix 'a/foo.txt'
(a/foo.txt, a/foo.txt.abc, a/foo.txt/subfolder/file.txt) in ``data`` bucket to
the ``b/`` folder in the ``data_backup`` bucket
(b/foo.txt, b/foo.txt.abc, b/foo.txt/subfolder/file.txt) ::
copy_files = GCSToGCSOperator(
task_id="copy_files_without_wildcard",
source_bucket="data",
source_object="a/foo.txt",
destination_bucket="data_backup",
destination_object="b/",
gcp_conn_id=google_cloud_conn_id,
)
"""
objects = hook.list(
self.source_bucket, prefix=prefix, delimiter=self.delimiter, match_glob=self.match_glob
)
objects = [obj for obj in objects if self._check_exact_match(obj, prefix)]
if not self.replace:
# If we are not replacing, ignore files already existing in source buckets
objects = self._ignore_existing_files(
hook, prefix, objects=objects, delimiter=self.delimiter, match_glob=self.match_glob
)
# If objects is empty, and we have prefix, let's check if prefix is a blob
# and copy directly
if len(objects) == 0 and prefix:
if hook.exists(self.source_bucket, prefix):
self._copy_single_object(
hook=hook, source_object=prefix, destination_object=self.destination_object
)
elif self.source_object_required:
msg = f"{prefix} does not exist in bucket {self.source_bucket}"
self.log.warning(msg)
raise AirflowException(msg)
if len(objects) == 1 and objects[0][-1] != "/":
self._copy_file(hook=hook, source_object=objects[0])
elif len(objects):
self._copy_multiple_objects(hook=hook, source_objects=objects, prefix=prefix)
def _copy_file(self, hook, source_object):
destination_object = self.destination_object or source_object
if self.destination_object and self.destination_object[-1] == "/":
file_name = source_object.split("/")[-1]
destination_object += file_name
self._copy_single_object(
hook=hook, source_object=source_object, destination_object=destination_object
)
def _copy_multiple_objects(self, hook, source_objects, prefix):
# Check whether the prefix is a root directory for all the rest of objects.
_pref = prefix.rstrip("/")
is_directory = prefix.endswith("/") or all(
obj.replace(_pref, "", 1).startswith("/") for obj in source_objects
)
if is_directory:
base_path = prefix.rstrip("/") + "/"
else:
base_path = prefix[0 : prefix.rfind("/") + 1] if "/" in prefix else ""
for source_obj in source_objects:
if not self._check_exact_match(source_obj, prefix):
continue
if self.destination_object is None:
destination_object = source_obj
else:
file_name_postfix = source_obj.replace(base_path, "", 1)
destination_object = self.destination_object.rstrip("/") + "/" + file_name_postfix
self._copy_single_object(
hook=hook, source_object=source_obj, destination_object=destination_object
)
def _check_exact_match(self, source_object: str, prefix: str) -> bool:
"""Checks whether source_object's name matches the prefix according to the exact_match flag."""
if self.exact_match and (source_object != prefix or not source_object.endswith(prefix)):
return False
return True
def _copy_source_with_wildcard(self, hook, prefix):
total_wildcards = prefix.count(WILDCARD)
if total_wildcards > 1:
error_msg = (
"Only one wildcard '*' is allowed in source_object parameter. "
f"Found {total_wildcards} in {prefix}."
)
raise AirflowException(error_msg)
self.log.info("Delimiter ignored because wildcard is in prefix")
prefix_, delimiter = prefix.split(WILDCARD, 1)
objects = hook.list(self.source_bucket, prefix=prefix_, delimiter=delimiter)
# TODO: After deprecating delimiter and wildcards in source objects,
# remove previous line and uncomment the following:
# match_glob = f"**/*{delimiter}" if delimiter else None
# objects = hook.list(self.source_bucket, prefix=prefix_, match_glob=match_glob)
if not self.replace:
# If we are not replacing, list all files in the Destination GCS bucket
# and only keep those files which are present in
# Source GCS bucket and not in Destination GCS bucket
objects = self._ignore_existing_files(hook, prefix_, delimiter=delimiter, objects=objects)
# TODO: After deprecating delimiter and wildcards in source objects,
# remove previous line and uncomment the following:
# objects = self._ignore_existing_files(hook, prefix_, match_glob=match_glob, objects=objects)
for source_object in objects:
if self.destination_object is None:
destination_object = source_object
else:
destination_object = source_object.replace(prefix_, self.destination_object, 1)
self._copy_single_object(
hook=hook, source_object=source_object, destination_object=destination_object
)
def _copy_single_object(self, hook, source_object, destination_object):
if self.is_older_than:
# Here we check if the given object is older than the given time
# If given, last_modified_time and maximum_modified_time is ignored
if hook.is_older_than(self.source_bucket, source_object, self.is_older_than):
self.log.info("Object is older than %s seconds ago", self.is_older_than)
else:
self.log.debug("Object is not older than %s seconds ago", self.is_older_than)
return
elif self.last_modified_time and self.maximum_modified_time:
# check to see if object was modified between last_modified_time and
# maximum_modified_time
if hook.is_updated_between(
self.source_bucket, source_object, self.last_modified_time, self.maximum_modified_time
):
self.log.info(
"Object has been modified between %s and %s",
self.last_modified_time,
self.maximum_modified_time,
)
else:
self.log.debug(
"Object was not modified between %s and %s",
self.last_modified_time,
self.maximum_modified_time,
)
return
elif self.last_modified_time is not None:
# Check to see if object was modified after last_modified_time
if hook.is_updated_after(self.source_bucket, source_object, self.last_modified_time):
self.log.info("Object has been modified after %s ", self.last_modified_time)
else:
self.log.debug("Object was not modified after %s ", self.last_modified_time)
return
elif self.maximum_modified_time is not None:
# Check to see if object was modified before maximum_modified_time
if hook.is_updated_before(self.source_bucket, source_object, self.maximum_modified_time):
self.log.info("Object has been modified before %s ", self.maximum_modified_time)
else:
self.log.debug("Object was not modified before %s ", self.maximum_modified_time)
return
self.log.info(
"Executing copy of gs://%s/%s to gs://%s/%s",
self.source_bucket,
source_object,
self.destination_bucket,
destination_object,
)
self.resolved_source_objects.add(source_object)
if not destination_object:
self.resolved_target_objects.add(source_object)
else:
self.resolved_target_objects.add(destination_object)
hook.rewrite(self.source_bucket, source_object, self.destination_bucket, destination_object)
if self.move_object:
hook.delete(self.source_bucket, source_object)
[docs] def get_openlineage_facets_on_complete(self, task_instance):
"""
Implementing _on_complete because execute method does preprocessing on internals.
This means we won't have to normalize self.source_object and self.source_objects,
destination bucket and so on.
"""
from openlineage.client.run import Dataset
from airflow.providers.openlineage.extractors import OperatorLineage
return OperatorLineage(
inputs=[
Dataset(namespace=f"gs://{self.source_bucket}", name=source)
for source in sorted(self.resolved_source_objects)
],
outputs=[
Dataset(namespace=f"gs://{self.destination_bucket}", name=target)
for target in sorted(self.resolved_target_objects)
],
)
