Files
govoplan-files/src/govoplan_files/backend/storage/archives.py

131 lines
5.3 KiB
Python

from __future__ import annotations
import mimetypes
import zipfile
from io import BytesIO
from os import PathLike
from pathlib import Path
from typing import Any, Iterable
from sqlalchemy.orm import Session
from govoplan_files.backend.db.models import FileAsset
from govoplan_files.backend.storage.common import FileConflictResolution, FileStorageError, UploadedStoredFile
from govoplan_files.backend.storage.backends import StorageBackendError, get_storage_backend
from govoplan_files.backend.storage.files import create_file_asset, current_versions_and_blobs
from govoplan_files.backend.storage.paths import filename_from_path, normalize_folder, normalize_logical_path
_ZIP_READ_CHUNK_SIZE = 1024 * 1024
def _read_zip_member(
archive: zipfile.ZipFile,
info: zipfile.ZipInfo,
*,
max_file_bytes: int,
max_total_bytes: int,
current_total: int,
) -> tuple[bytes, int]:
parts: list[bytes] = []
actual_size = 0
with archive.open(info) as source:
while True:
read_size = min(_ZIP_READ_CHUNK_SIZE, max_file_bytes + 1 - actual_size)
chunk = source.read(read_size)
if not chunk:
break
actual_size += len(chunk)
if actual_size > max_file_bytes:
raise FileStorageError(f"ZIP member {info.filename!r} exceeds per-file limit")
if current_total + actual_size > max_total_bytes:
raise FileStorageError("ZIP is too large after extraction")
parts.append(chunk)
return b"".join(parts), current_total + actual_size
def create_zip_file(session: Session, assets: Iterable[FileAsset], output_path: str | Path) -> None:
backend = get_storage_backend()
asset_list = list(assets)
version_blobs = current_versions_and_blobs(session, asset_list)
with zipfile.ZipFile(output_path, mode="w", compression=zipfile.ZIP_DEFLATED) as archive:
for asset in asset_list:
_version, blob = version_blobs[asset.id]
info = zipfile.ZipInfo(asset.display_path)
info.compress_type = zipfile.ZIP_DEFLATED
with archive.open(info, "w") as member:
try:
for chunk in backend.iter_bytes(blob.storage_key):
if chunk:
member.write(chunk)
except StorageBackendError as exc:
raise FileStorageError(str(exc)) from exc
def extract_zip_upload(
session: Session,
*,
tenant_id: str,
owner_type: str,
owner_id: str,
user_id: str,
zip_data: bytes | str | PathLike[str],
folder: str | None,
campaign_id: str | None,
conflict_strategy: str = "reject",
conflict_resolutions: Iterable[FileConflictResolution] | None = None,
metadata: dict[str, Any] | None = None,
is_admin: bool = False,
max_files: int = 1000,
max_file_bytes: int = 50 * 1024 * 1024,
max_total_bytes: int = 250 * 1024 * 1024,
) -> list[UploadedStoredFile]:
uploaded: list[UploadedStoredFile] = []
total = 0
base_folder = normalize_folder(folder)
try:
source = BytesIO(zip_data) if isinstance(zip_data, bytes) else zip_data
with zipfile.ZipFile(source) as archive:
infos = [info for info in archive.infolist() if not info.is_dir()]
if len(infos) > max_files:
raise FileStorageError(f"ZIP contains too many files (limit {max_files})")
for info in infos:
if info.flag_bits & 0x1:
raise FileStorageError("Encrypted ZIP uploads are not supported")
if info.file_size < 0:
raise FileStorageError("Invalid ZIP member")
if info.file_size > max_file_bytes:
raise FileStorageError(f"ZIP member {info.filename!r} exceeds per-file limit")
if total + info.file_size > max_total_bytes:
raise FileStorageError("ZIP is too large after extraction")
inner_path = normalize_logical_path(info.filename)
target_path = f"{base_folder}/{inner_path}" if base_folder else inner_path
data, total = _read_zip_member(
archive,
info,
max_file_bytes=max_file_bytes,
max_total_bytes=max_total_bytes,
current_total=total,
)
uploaded.append(
create_file_asset(
session,
tenant_id=tenant_id,
owner_type=owner_type,
owner_id=owner_id,
user_id=user_id,
filename=filename_from_path(inner_path),
data=data,
display_path=target_path,
content_type=mimetypes.guess_type(inner_path)[0] or "application/octet-stream",
metadata=metadata,
campaign_id=campaign_id,
conflict_strategy=conflict_strategy,
conflict_resolutions=conflict_resolutions,
is_admin=is_admin,
)
)
except zipfile.BadZipFile as exc:
raise FileStorageError("Invalid ZIP upload") from exc
return uploaded