zulip/tools/upload-release

#!/usr/bin/env python3
import argparse
import hashlib
import os
import re
import sys
import tempfile
from typing import IO, cast

sys.path.insert(0, os.path.join(os.path.dirname(__file__), ".."))
from scripts.lib.setup_path import setup_path

setup_path()

import boto3.session
from mypy_boto3_s3 import S3Client
from natsort import natsorted


def sha256_contents(f: IO[bytes]) -> str:
    sha256_hash = hashlib.sha256()
    for byte_block in iter(lambda: f.read(4096), b""):
        sha256_hash.update(byte_block)
    return sha256_hash.hexdigest()


parser = argparse.ArgumentParser(description="Upload a Zulip release")
parser.add_argument(
    dest="filename",
    help="Tarball to upload",
    metavar="FILE",
)
args = parser.parse_args()
if not os.path.exists(args.filename):
    parser.error(f"File does not exist: {args.filename}")
new_basename = os.path.basename(args.filename)
if not new_basename.startswith("zulip-server-") or not new_basename.endswith(".tar.gz"):
    parser.error("File does not match zulip-server-*.tar.gz")

session = boto3.session.Session()
client: S3Client = session.client("s3")
bucket = session.resource("s3", region_name="us-east-1").Bucket("zulip-download")

file_hashes: dict[str, str] = {}
with open(args.filename, "rb") as new_file:
    print(f"Hashing {new_basename}..")
    file_hashes[new_basename] = sha256_contents(new_file)

print("Fetching existing hashes..")
for obj_summary in bucket.objects.filter(Prefix="server/zulip-server-"):
    head = client.head_object(Bucket=bucket.name, Key=obj_summary.key)
    assert obj_summary.key.startswith("server/")
    filename = obj_summary.key.removeprefix("server/")
    if filename in file_hashes:
        print(f"  {filename} was already uploaded, skipping existing hash")
        continue
    metadata = head["Metadata"]
    if "sha256sum" not in metadata:
        print(f"  {filename} does not have SHA256 metadata!")
        with tempfile.TemporaryFile() as obj_contents:
            print("    Downloading..")
            bucket.download_fileobj(Key=obj_summary.key, Fileobj=obj_contents)
            obj_contents.seek(0)
            print("    Hashing..")
            metadata["sha256sum"] = sha256_contents(obj_contents)
            print(f"    Got {metadata['sha256sum']}")

        print("    Updating..")
        obj_summary.copy_from(
            ContentType=head["ContentType"],
            CopySource=f"{bucket.name}/{obj_summary.key}",
            Metadata=metadata,
            MetadataDirective="REPLACE",
        )

    file_hashes[filename] = metadata["sha256sum"]

ordered_filenames = cast(list[str], natsorted(file_hashes.keys(), reverse=True))
assert ordered_filenames[0] == "zulip-server-latest.tar.gz"

print(f"Uploading {new_basename}..")
extra = {
    "Metadata": {"sha256sum": file_hashes[new_basename]},
    "ACL": "public-read",
    "ContentType": "application/gzip",
}
bucket.upload_file(
    Filename=args.filename,
    Key=f"server/{new_basename}",
    ExtraArgs=extra,
)

if (
    next(name for name in ordered_filenames if re.match(r"^zulip-server-(\d+\.\d+)\.tar.gz$", name))
    == new_basename
):
    print("This looks like the most recent full release; copying to zulip-server-latest.tar.gz..")
    bucket.copy(
        CopySource={"Bucket": bucket.name, "Key": f"server/{new_basename}"},
        Key="server/zulip-server-latest.tar.gz",
        ExtraArgs=extra,
    )
    file_hashes["zulip-server-latest.tar.gz"] = file_hashes[new_basename]
else:
    print("Not copying to zulip-server-latest.tar.gz!")

print("Updating SHA256SUMS.txt..")
contents = ""
for ordered_filename in ordered_filenames:
    # natsorted type annotation is insufficiently generic
    assert isinstance(ordered_filename, str)
    contents += f"{file_hashes[ordered_filename]}  {ordered_filename}\n"
bucket.put_object(
    ACL="public-read",
    Body=contents.encode(),
    ContentType="text/plain",
    Key="server/SHA256SUMS.txt",
)