#!/bin/sh
set -e

# One-shot copy of all objects from the old MinIO service (deployed with
# compose.minio-migration.yml) into the RustFS object storage.
# Only the latest version of each object is copied, and only objects missing
# on RustFS: objects changed on RustFS since (e.g. edited documents) are never
# overwritten. Once a run completes, a marker object is written and further
# runs are refused, unless called with --force.

# Load secrets into environment
source /abra-entrypoint.sh -e

python - "$@" <<'PYEOF'
import os
import sys

import boto3
from botocore.exceptions import ClientError

bucket = os.environ.get("AWS_STORAGE_BUCKET_NAME", "drive-media-storage")
marker = "_minio-migration-done"
force = "--force" in sys.argv[1:]


def secret(name):
    with open(f"/run/secrets/{name}") as f:
        return f.read().strip()


def client(endpoint, access_key, secret_key):
    return boto3.client(
        "s3",
        endpoint_url=endpoint,
        aws_access_key_id=access_key,
        aws_secret_access_key=secret_key,
        region_name=os.environ.get("AWS_S3_REGION_NAME", "us-east-1"),
    )


def exists(s3, key):
    try:
        s3.head_object(Bucket=bucket, Key=key)
        return True
    except ClientError as error:
        if error.response["Error"]["Code"] in ("404", "NoSuchKey", "NotFound"):
            return False
        raise


# the old MinIO uses the minio_* secrets, RustFS the rustfs_* secrets
src = client(
    os.environ.get("MINIO_MIGRATION_ENDPOINT_URL", "http://minio:9000"),
    secret("minio_ru"),
    secret("minio_rp"),
)
dst = client(
    os.environ["AWS_S3_ENDPOINT_URL"],
    os.environ["AWS_S3_ACCESS_KEY_ID"],
    os.environ["AWS_S3_SECRET_ACCESS_KEY"],
)

if exists(dst, marker) and not force:
    sys.exit(
        "migrate-from-minio: the migration already completed, refusing to run again.\n"
        "Objects deleted on RustFS since then would be copied back from MinIO.\n"
        "Use 'migrate_from_minio --force' to run anyway (existing objects are never overwritten)."
    )

try:
    src.head_bucket(Bucket=bucket)
except Exception as error:
    sys.exit(f"migrate-from-minio: cannot reach bucket '{bucket}' on minio: {error}")

copied = skipped = 0
for page in src.get_paginator("list_objects_v2").paginate(Bucket=bucket):
    for obj in page.get("Contents", []):
        key = obj["Key"]
        # upload_fileobj is atomic: an object only exists once fully copied
        if exists(dst, key):
            skipped += 1
            continue
        source = src.get_object(Bucket=bucket, Key=key)
        extra = {"ContentType": source.get("ContentType", "binary/octet-stream")}
        if source.get("Metadata"):
            extra["Metadata"] = source["Metadata"]
        if source.get("ContentDisposition"):
            extra["ContentDisposition"] = source["ContentDisposition"]
        dst.upload_fileobj(source["Body"], bucket, key, ExtraArgs=extra)
        copied += 1
        if copied % 100 == 0:
            print(f"migrate-from-minio: {copied} objects copied...")

dst.put_object(Bucket=bucket, Key=marker, Body=b"")
print(f"migrate-from-minio: done, {copied} copied, {skipped} already present")
PYEOF