Copies all objects from source s3 path to target s3 path, filters results by last modified date

copy_objects_by_last_modified.py

from datetime import datetime
from typing import Optional, Callable

import awswrangler
import boto3
import pytz


def copy_bucket_content(
    source_s3_uri: str,
    target_s3_uri: str,
    last_modified_begin: datetime = datetime.now(pytz.UTC),
    filter_object_path: Optional[Callable[[str], bool]] = None,
    aws_profile_name: Optional[str] = None,
):
    if aws_profile_name is not None:
        boto3.setup_default_session(profile_name=aws_profile_name)

    object_list = awswrangler.s3.list_objects(
        source_s3_uri,
        last_modified_begin=last_modified_begin,
    )

    if filter_object_path is not None:
        object_list = [filter_object_path(o) for o in object_list]

    awswrangler.s3.copy_objects(
        object_list,
        source_path=source_s3_uri,
        target_path=target_s3_uri,
    )

if __name__ == '__main__':
    copy_bucket_content(
        source_s3_uri="",
        target_s3_uri="",
        last_modified_begin=datetime(year=2022, month=4, day=1, tzinfo=pytz.UTC)
    )