FabledCurator/backend/app/services/backup_destination.py

"""The off-site backup destination: an S3 bucket (Fabled Archives, or any S3
service) that the database dumps and the media blobs are sent to.

The address, bucket, access key and region are plain settings. The secret key
is sealed at rest with the credential key (`/images/secrets`), the same key
that seals account logins, and no API ever returns it: the settings answer
only whether one is set. Because the credential key is kept out of every
backup, a database dump carries the secret only as ciphertext nobody can open
without the original server.
"""
from __future__ import annotations

from dataclasses import dataclass
from pathlib import Path
from urllib.parse import urlsplit

from .backup_s3 import S3Error, S3Target, list_objects
from .credential_crypto import CredentialCrypto, InvalidCredentialBlob

KEY_PATH = Path("/images/secrets/credential_key.b64")
DEFAULT_REGION = "us-east-1"
_TEST_LISTING = 5


def _crypto() -> CredentialCrypto:
    return CredentialCrypto(KEY_PATH)


def seal(secret: str) -> str:
    return _crypto().encrypt(secret).decode("ascii")


def unseal(sealed: str) -> str:
    return _crypto().decrypt(sealed.encode("ascii"))


@dataclass(frozen=True)
class DestinationFields:
    endpoint: str | None
    bucket: str | None
    access_key: str | None
    region: str


def clean_fields(body: dict) -> tuple[DestinationFields | None, str | None]:
    """The destination fields from a request body, trimmed, with empty ones as
    None; or (None, why) when the address can't be an S3 endpoint."""
    def text(name: str) -> str | None:
        v = body.get(name)
        if v is None:
            return None
        if not isinstance(v, str):
            raise TypeError(name)
        return v.strip() or None

    try:
        endpoint = text("endpoint")
        bucket = text("bucket")
        access_key = text("access_key")
        region = text("region") or DEFAULT_REGION
    except TypeError as exc:
        return None, f"{exc.args[0]} must be text"
    if endpoint is not None:
        u = urlsplit(endpoint)
        if u.scheme not in ("http", "https") or not u.hostname:
            return None, ("the address must start with http:// or https:// and name "
                          "a host, e.g. http://archives.example:8081")
        if u.query or u.fragment:
            return None, "the address can't carry a query or a fragment"
        endpoint = endpoint.rstrip("/")
    if bucket is not None and "/" in bucket:
        return None, "the bucket is a name, not a path: put any path in the address"
    return DestinationFields(endpoint, bucket, access_key, region), None


def target_for(fields: DestinationFields, secret: str | None) -> S3Target | None:
    """The S3 target, or None while any part of it is missing."""
    if not (fields.endpoint and fields.bucket and fields.access_key and secret):
        return None
    return S3Target(fields.endpoint, fields.bucket, fields.access_key, secret,
                    region=fields.region)


def saved_target(settings) -> S3Target | None:
    """The destination saved in ImportSettings, or None when none is set up.
    Raises S3Error when the secret can't be opened (the credential key
    changed since it was saved)."""
    sealed = settings.backup_s3_secret_sealed
    if not sealed:
        return None
    try:
        secret = unseal(sealed)
    except InvalidCredentialBlob:
        raise S3Error(
            "The saved secret key can't be opened on this server (the credential "
            "key changed since it was saved). Enter the secret key again."
        ) from None
    return target_for(fields_of(settings), secret)


def fields_of(settings) -> DestinationFields:
    return DestinationFields(
        settings.backup_s3_endpoint, settings.backup_s3_bucket,
        settings.backup_s3_access_key, settings.backup_s3_region or DEFAULT_REGION,
    )


def describe(settings) -> dict:
    """The destination as the settings screen shows it: never the secret."""
    f = fields_of(settings)
    return {
        "endpoint": f.endpoint or "",
        "bucket": f.bucket or "",
        "access_key": f.access_key or "",
        "region": f.region,
        "secret_set": bool(settings.backup_s3_secret_sealed),
        "upload_limit_mbps": settings.backup_upload_limit_mbps,
        "configured": bool(
            f.endpoint and f.bucket and f.access_key and settings.backup_s3_secret_sealed
        ),
    }


def check_target(target: S3Target) -> dict:
    """Reach the bucket by listing it. A listing proves the address, the
    bucket and both keys at once, and writes nothing, because an Archives key
    is allowed to add objects but a test object could never be removed."""
    try:
        page = list_objects(target, limit=_TEST_LISTING)
    except S3Error as exc:
        return {"ok": False, "message": str(exc)}
    if not page.objects:
        said = f"Reached {target.bucket} at {target.endpoint}. The bucket is empty."
    else:
        more = "+" if page.next_token else ""
        said = (f"Reached {target.bucket} at {target.endpoint}. It holds "
                f"{len(page.objects)}{more} object(s), e.g. {page.objects[0].key}.")
    return {"ok": True, "message": said}