Initial release of alma_blob_upload tool

.gitignore

@@ -0,0 +1,2 @@
+.idea
+.mypy_cache

README.md

@@ -1,3 +1,65 @@
 # almalinux-git-utils
 
-Utilities for working with the AlmaLinux OS Git server.
+Utilities for working with the AlmaLinux OS Git server.
+
+
+## alma_blob_upload
+
+The `alma_blob_upload` script uploads sources and BLOBs to the AlmaLinux
+sources cache.
+
+### Prerequirements
+
+Install the `python3-boto3` package:
+
+```shell
+# RPM-based distributions. On EL8 derivatives the package is available from EPEL.
+$ sudo dnf install python3 python3-boto3
+
+# Debian-based distributions
+$ sudo apt install python3-boto3
+```
+
+Create an AWS credentials file ~/.aws/credentials with the following content:
+
+```ini
+[default]
+aws_access_key_id = YOUR_ACCESS_KEY
+aws_secret_access_key = YOUR_SECRET_KEY
+```
+
+
+### Usage
+
+The utility supports two types of input: a CentOS git repository metadata file
+or a list of files to upload.
+
+For CentOS repositories workflow will be the following:
+
+1. Install the `get_sources.sh` script from the
+   [centos-git-common](https://git.centos.org/centos-git-common) repository.
+2. Clone a project and download its sources as described on the CentOS
+   [Wiki](https://wiki.centos.org/Sources).
+3. Run the `alma_blob_upload` tool (don't forget to replace `PROJECT_NAME` with
+   an actual project name):
+   ```shell
+   $ alma_blob_upload.py -i .PROJECT_NAME.metadata
+   ```
+
+Alternatively, you can upload a list of files in the following way:
+
+```shell
+$ alma_blob_upload.py -f SOURCES/FILE_1 SOURCES/FILE_N
+```
+
+The `alma_blob_upload` utility can also generate a CentOS-compatible metadata
+file:
+
+```
+$ alma_blob_upload.py -o .PROJECT_NAME.metadata -f SOURCES/FILE_1 SOURCES/FILE_N
+```
+
+
+## License
+
+Licensed under the GPLv3 license, see the LICENSE file for details.

alma_blob_upload.py

@@ -0,0 +1,193 @@
+#!/usr/bin/env python3
+
+"""
+Uploads sources and BLOBs to the AlmaLinux sources cache.
+"""
+
+import argparse
+import hashlib
+import os.path
+import sys
+from typing import Iterator, List, TextIO, Tuple
+
+import boto3 # type: ignore
+from botocore.exceptions import ClientError # type: ignore
+
+
+def init_arg_parser() -> argparse.ArgumentParser:
+    """
+    Initializes a command line arguments parser.
+
+    Returns:
+        Command line arguments parser.
+    """
+    arg_parser = argparse.ArgumentParser(
+        prog="alma_blob_upload",
+        description="Uploads sources and BLOBs to the AlmaLinux sources cache"
+    )
+    group = arg_parser.add_mutually_exclusive_group(required=True)
+    group.add_argument('-f', '--file', nargs='+', help='file(s) to upload')
+    group.add_argument('-i', '--input-metadata', metavar='INPUT_FILE',
+                       help='input metadata file list to upload')
+    arg_parser.add_argument('-b', '--bucket', default="sources.almalinux.org",
+                            help='Amazon S3 bucket name. Default is '
+                                 'sources.almalinux.org')
+    arg_parser.add_argument('-o', '--output-metadata', metavar='OUTPUT_FILE',
+                            help='output metadata file path')
+    arg_parser.add_argument('-p', '--private', action='store_true',
+                            help='set uploaded file mode to private. All '
+                                 'uploaded files are public by default')
+    arg_parser.add_argument('--domain-name', default='sources.almalinux.org',
+                            help='AlmaLinux sources server domain name. '
+                                 'Default is sources.almalinux.org')
+    arg_parser.add_argument('-v', '--verbose', action='store_true',
+                            help='enable additional debug output')
+    return arg_parser
+
+
+def get_file_checksum(file_path: str, checksum_type: str = 'sha1',
+                      buff_size: int = 1048576) -> str:
+    """
+    Calculates a file checksum.
+
+    Args:
+        file_path: File path.
+        checksum_type: Checksum type.
+        buff_size: Number of bytes to read at once.
+
+    Returns:
+        File checksum.
+    """
+    hasher = hashlib.new(checksum_type)
+    with open(file_path, 'rb') as fd:
+        buff = fd.read(buff_size)
+        while len(buff):
+            hasher.update(buff)
+            buff = fd.read(buff_size)
+    return hasher.hexdigest()
+
+
+def normalize_path(path: str) -> str:
+    """
+    Returns an absolute path with all variables expanded.
+
+    Args:
+        path: Path to normalize.
+
+    Returns:
+        Normalized path.
+    """
+    return os.path.abspath(os.path.expanduser(os.path.expandvars(path)))
+
+
+def iter_metadata(metadata_path: str) -> Iterator[Tuple[str, str]]:
+    """
+    Iterates over records in a CentOS git repository-compatible metadata file.
+
+    Args:
+        metadata_path: Metadata file path.
+
+    Returns:
+        Iterator over files and their checksums.
+    """
+    with open(metadata_path, 'r') as fd:
+        for line in fd:
+            checksum, file_path = line.split()
+            file_path = normalize_path(file_path)
+            assert checksum == get_file_checksum(file_path)
+            yield file_path, checksum
+
+
+def iter_files(files: List[str]) -> Iterator[Tuple[str, str]]:
+    """
+    Iterates over a list of files and calculates checksums for them.
+
+    Args:
+        files: List of files.
+
+    Returns:
+        Iterator over files and their checksums.
+    """
+    for file_path in files:
+        file_path = normalize_path(file_path)
+        checksum = get_file_checksum(file_path)
+        yield file_path, checksum
+
+
+def is_file_exist(s3_client, bucket_name: str, checksum: str) -> bool:
+    """
+    Checks is a file with a given checksum is already uploaded.
+    Args:
+        s3_client: Amazon S3 client.
+        bucket_name: S3 bucket name.
+        checksum: File checksum.
+
+    Returns:
+        True if a file is already uploaded, False otherwise.
+    """
+    try:
+        s3_client.head_object(Bucket=bucket_name, Key=checksum)
+        return True
+    except ClientError:
+        return False
+
+
+def upload_file(s3_client, bucket_name: str, file_path: str,
+                checksum: str, private: bool):
+    """
+    Uploads a file to an Amazon S3 bucket.
+
+    Args:
+        s3_client: Amazon S3 client.
+        bucket_name: S3 bucket name.
+        file_path: File path.
+        checksum: File checksum.
+        private: False if file should be public, True otherwise.
+    """
+    acl = 'bucket-owner-full-control' if private else 'public-read'
+    s3_client.upload_file(file_path, bucket_name, checksum,
+                          ExtraArgs={'ACL': acl})
+
+
+def add_metadata_record(metadata_fd: TextIO, file_path: str, checksum: str):
+    """
+    Adds a source file record to a metadata file.
+
+    Args:
+        metadata_fd: Metadata file descriptor.
+        file_path: Source file path.
+        checksum: Source file checksum.
+    """
+    rel_path = os.path.relpath(file_path)
+    metadata_fd.write(f'{checksum} {rel_path}\n')
+
+
+def main(sys_args):
+    arg_parser = init_arg_parser()
+    args = arg_parser.parse_args(sys_args)
+    s3_client = boto3.client('s3')
+    if args.input_metadata:
+        iterator = iter_metadata(args.input_metadata)
+    else:
+        iterator = iter_files(args.file)
+    out_fd = None
+    if args.output_metadata:
+        out_fd = open(args.output_metadata, 'w')
+    try:
+        for file_path, checksum in iterator:
+            file_url = f'https://{args.domain_name}/{checksum}'
+            if not is_file_exist(s3_client, args.bucket, checksum):
+                upload_file(s3_client, args.bucket, file_path, checksum,
+                            args.private)
+                print(f'{file_path} uploaded: {file_url}')
+            else:
+                print(f'{file_path} exists: {file_url}')
+            if out_fd:
+                add_metadata_record(out_fd, file_path, checksum)
+    finally:
+        if out_fd:
+            out_fd.close()
+
+
+if __name__ == '__main__':
+    sys.exit(main(sys.argv[1:]))