8 Commits

20 changed files with 1749 additions and 831 deletions

View File

@@ -223,6 +223,13 @@ def create_app(
app.extensions["access_logging"] = access_logging_service
app.extensions["site_registry"] = site_registry
from .s3_client import S3ProxyClient
api_base = app.config.get("API_BASE_URL") or "http://127.0.0.1:5000"
app.extensions["s3_proxy"] = S3ProxyClient(
api_base_url=api_base,
region=app.config.get("AWS_REGION", "us-east-1"),
)
operation_metrics_collector = None
if app.config.get("OPERATION_METRICS_ENABLED", False):
operation_metrics_collector = OperationMetricsCollector(

View File

@@ -36,11 +36,11 @@ class GzipMiddleware:
content_type = None
content_length = None
should_compress = False
is_streaming = False
passthrough = False
exc_info_holder = [None]
def custom_start_response(status: str, headers: List[Tuple[str, str]], exc_info=None):
nonlocal response_started, status_code, response_headers, content_type, content_length, should_compress, is_streaming
nonlocal response_started, status_code, response_headers, content_type, content_length, should_compress, passthrough
response_started = True
status_code = int(status.split(' ', 1)[0])
response_headers = list(headers)
@@ -51,23 +51,29 @@ class GzipMiddleware:
if name_lower == 'content-type':
content_type = value.split(';')[0].strip().lower()
elif name_lower == 'content-length':
try:
content_length = int(value)
except (ValueError, TypeError):
pass
elif name_lower == 'content-encoding':
should_compress = False
passthrough = True
return start_response(status, headers, exc_info)
elif name_lower == 'x-stream-response':
is_streaming = True
passthrough = True
return start_response(status, headers, exc_info)
if content_type and content_type in COMPRESSIBLE_MIMES:
if content_length is None or content_length >= self.min_size:
should_compress = True
else:
passthrough = True
return start_response(status, headers, exc_info)
return None
app_iter = self.app(environ, custom_start_response)
if is_streaming:
if passthrough:
return app_iter
response_body = b''.join(app_iter)

View File

@@ -309,6 +309,18 @@ class IamService:
if not self._is_allowed(principal, normalized, action):
raise IamError(f"Access denied for action '{action}' on bucket '{bucket_name}'")
def check_permissions(self, principal: Principal, bucket_name: str | None, actions: Iterable[str]) -> Dict[str, bool]:
self._maybe_reload()
bucket_name = (bucket_name or "*").lower() if bucket_name != "*" else (bucket_name or "*")
normalized_actions = {a: self._normalize_action(a) for a in actions}
results: Dict[str, bool] = {}
for original, canonical in normalized_actions.items():
if canonical not in ALLOWED_ACTIONS:
results[original] = False
else:
results[original] = self._is_allowed(principal, bucket_name, canonical)
return results
def buckets_for_principal(self, principal: Principal, buckets: Iterable[str]) -> List[str]:
return [bucket for bucket in buckets if self._is_allowed(principal, bucket, "list")]

View File

@@ -160,6 +160,7 @@ class KMSManager:
self.generate_data_key_max_bytes = generate_data_key_max_bytes
self._keys: Dict[str, KMSKey] = {}
self._master_key: bytes | None = None
self._master_aesgcm: AESGCM | None = None
self._loaded = False
@property
@@ -191,6 +192,7 @@ class KMSManager:
msvcrt.locking(lock_file.fileno(), msvcrt.LK_UNLCK, 1)
else:
fcntl.flock(lock_file.fileno(), fcntl.LOCK_UN)
self._master_aesgcm = AESGCM(self._master_key)
return self._master_key
def _load_keys(self) -> None:
@@ -231,18 +233,16 @@ class KMSManager:
_set_secure_file_permissions(self.keys_path)
def _encrypt_key_material(self, key_material: bytes) -> bytes:
"""Encrypt key material with the master key."""
aesgcm = AESGCM(self.master_key)
_ = self.master_key
nonce = secrets.token_bytes(12)
ciphertext = aesgcm.encrypt(nonce, key_material, None)
ciphertext = self._master_aesgcm.encrypt(nonce, key_material, None)
return nonce + ciphertext
def _decrypt_key_material(self, encrypted: bytes) -> bytes:
"""Decrypt key material with the master key."""
aesgcm = AESGCM(self.master_key)
_ = self.master_key
nonce = encrypted[:12]
ciphertext = encrypted[12:]
return aesgcm.decrypt(nonce, ciphertext, None)
return self._master_aesgcm.decrypt(nonce, ciphertext, None)
def create_key(self, description: str = "", key_id: str | None = None) -> KMSKey:
"""Create a new KMS key."""
@@ -404,22 +404,6 @@ class KMSManager:
plaintext, _ = self.decrypt(encrypted_key, context)
return plaintext
def get_provider(self, key_id: str | None = None) -> KMSEncryptionProvider:
"""Get an encryption provider for a specific key."""
self._load_keys()
if key_id is None:
if not self._keys:
key = self.create_key("Default KMS Key")
key_id = key.key_id
else:
key_id = next(iter(self._keys.keys()))
if key_id not in self._keys:
raise EncryptionError(f"Key not found: {key_id}")
return KMSEncryptionProvider(self, key_id)
def re_encrypt(self, ciphertext: bytes, destination_key_id: str,
source_context: Dict[str, str] | None = None,
destination_context: Dict[str, str] | None = None) -> bytes:

View File

@@ -176,11 +176,12 @@ class ReplicationFailureStore:
self.storage_root = storage_root
self.max_failures_per_bucket = max_failures_per_bucket
self._lock = threading.Lock()
self._cache: Dict[str, List[ReplicationFailure]] = {}
def _get_failures_path(self, bucket_name: str) -> Path:
return self.storage_root / ".myfsio.sys" / "buckets" / bucket_name / "replication_failures.json"
def load_failures(self, bucket_name: str) -> List[ReplicationFailure]:
def _load_from_disk(self, bucket_name: str) -> List[ReplicationFailure]:
path = self._get_failures_path(bucket_name)
if not path.exists():
return []
@@ -192,7 +193,7 @@ class ReplicationFailureStore:
logger.error(f"Failed to load replication failures for {bucket_name}: {e}")
return []
def save_failures(self, bucket_name: str, failures: List[ReplicationFailure]) -> None:
def _save_to_disk(self, bucket_name: str, failures: List[ReplicationFailure]) -> None:
path = self._get_failures_path(bucket_name)
path.parent.mkdir(parents=True, exist_ok=True)
data = {"failures": [f.to_dict() for f in failures[:self.max_failures_per_bucket]]}
@@ -202,6 +203,18 @@ class ReplicationFailureStore:
except OSError as e:
logger.error(f"Failed to save replication failures for {bucket_name}: {e}")
def load_failures(self, bucket_name: str) -> List[ReplicationFailure]:
if bucket_name in self._cache:
return list(self._cache[bucket_name])
failures = self._load_from_disk(bucket_name)
self._cache[bucket_name] = failures
return list(failures)
def save_failures(self, bucket_name: str, failures: List[ReplicationFailure]) -> None:
trimmed = failures[:self.max_failures_per_bucket]
self._cache[bucket_name] = trimmed
self._save_to_disk(bucket_name, trimmed)
def add_failure(self, bucket_name: str, failure: ReplicationFailure) -> None:
with self._lock:
failures = self.load_failures(bucket_name)
@@ -227,6 +240,7 @@ class ReplicationFailureStore:
def clear_failures(self, bucket_name: str) -> None:
with self._lock:
self._cache.pop(bucket_name, None)
path = self._get_failures_path(bucket_name)
if path.exists():
path.unlink()

View File

@@ -3,6 +3,7 @@ from __future__ import annotations
import base64
import hashlib
import hmac
import json
import logging
import mimetypes
import re
@@ -999,6 +1000,7 @@ def _apply_object_headers(
etag: str,
) -> None:
if file_stat is not None:
if response.status_code != 206:
response.headers["Content-Length"] = str(file_stat.st_size)
response.headers["Last-Modified"] = http_date(file_stat.st_mtime)
response.headers["ETag"] = f'"{etag}"'
@@ -2779,7 +2781,7 @@ def object_handler(bucket_name: str, object_key: str):
try:
stat = path.stat()
file_size = stat.st_size
etag = storage._compute_etag(path)
etag = metadata.get("__etag__") or storage._compute_etag(path)
except PermissionError:
return _error_response("AccessDenied", "Permission denied accessing object", 403)
except OSError as exc:
@@ -2827,7 +2829,7 @@ def object_handler(bucket_name: str, object_key: str):
try:
stat = path.stat()
response = Response(status=200)
etag = storage._compute_etag(path)
etag = metadata.get("__etag__") or storage._compute_etag(path)
except PermissionError:
return _error_response("AccessDenied", "Permission denied accessing object", 403)
except OSError as exc:
@@ -2962,7 +2964,11 @@ def _bucket_policy_handler(bucket_name: str) -> Response:
store.delete_policy(bucket_name)
current_app.logger.info("Bucket policy removed", extra={"bucket": bucket_name})
return Response(status=204)
payload = request.get_json(silent=True)
raw_body = request.get_data(cache=False) or b""
try:
payload = json.loads(raw_body)
except (json.JSONDecodeError, ValueError):
return _error_response("MalformedPolicy", "Policy document must be JSON", 400)
if not payload:
return _error_response("MalformedPolicy", "Policy document must be JSON", 400)
try:

284
app/s3_client.py Normal file
View File

@@ -0,0 +1,284 @@
from __future__ import annotations
import json
import logging
import threading
import time
from typing import Any, Generator, Optional
import boto3
from botocore.config import Config
from botocore.exceptions import ClientError, EndpointConnectionError, ConnectionClosedError
from flask import current_app, session
logger = logging.getLogger(__name__)
UI_PROXY_USER_AGENT = "MyFSIO-UIProxy/1.0"
_BOTO_ERROR_MAP = {
"NoSuchBucket": 404,
"NoSuchKey": 404,
"NoSuchUpload": 404,
"BucketAlreadyExists": 409,
"BucketAlreadyOwnedByYou": 409,
"BucketNotEmpty": 409,
"AccessDenied": 403,
"InvalidAccessKeyId": 403,
"SignatureDoesNotMatch": 403,
"InvalidBucketName": 400,
"InvalidArgument": 400,
"MalformedXML": 400,
"EntityTooLarge": 400,
"QuotaExceeded": 403,
}
_UPLOAD_REGISTRY_MAX_AGE = 86400
_UPLOAD_REGISTRY_CLEANUP_INTERVAL = 3600
class UploadRegistry:
def __init__(self) -> None:
self._entries: dict[str, tuple[str, str, float]] = {}
self._lock = threading.Lock()
self._last_cleanup = time.monotonic()
def register(self, upload_id: str, bucket_name: str, object_key: str) -> None:
with self._lock:
self._entries[upload_id] = (bucket_name, object_key, time.monotonic())
self._maybe_cleanup()
def get_key(self, upload_id: str, bucket_name: str) -> Optional[str]:
with self._lock:
entry = self._entries.get(upload_id)
if entry is None:
return None
stored_bucket, key, created_at = entry
if stored_bucket != bucket_name:
return None
if time.monotonic() - created_at > _UPLOAD_REGISTRY_MAX_AGE:
del self._entries[upload_id]
return None
return key
def remove(self, upload_id: str) -> None:
with self._lock:
self._entries.pop(upload_id, None)
def _maybe_cleanup(self) -> None:
now = time.monotonic()
if now - self._last_cleanup < _UPLOAD_REGISTRY_CLEANUP_INTERVAL:
return
self._last_cleanup = now
cutoff = now - _UPLOAD_REGISTRY_MAX_AGE
stale = [uid for uid, (_, _, ts) in self._entries.items() if ts < cutoff]
for uid in stale:
del self._entries[uid]
class S3ProxyClient:
def __init__(self, api_base_url: str, region: str = "us-east-1") -> None:
if not api_base_url:
raise ValueError("api_base_url is required for S3ProxyClient")
self._api_base_url = api_base_url.rstrip("/")
self._region = region
self.upload_registry = UploadRegistry()
@property
def api_base_url(self) -> str:
return self._api_base_url
def get_client(self, access_key: str, secret_key: str) -> Any:
if not access_key or not secret_key:
raise ValueError("Both access_key and secret_key are required")
config = Config(
user_agent_extra=UI_PROXY_USER_AGENT,
connect_timeout=5,
read_timeout=30,
retries={"max_attempts": 0},
signature_version="s3v4",
s3={"addressing_style": "path"},
request_checksum_calculation="when_required",
response_checksum_validation="when_required",
)
return boto3.client(
"s3",
endpoint_url=self._api_base_url,
aws_access_key_id=access_key,
aws_secret_access_key=secret_key,
region_name=self._region,
config=config,
)
def _get_proxy() -> S3ProxyClient:
proxy = current_app.extensions.get("s3_proxy")
if proxy is None:
raise RuntimeError(
"S3 proxy not configured. Set API_BASE_URL or run both API and UI servers."
)
return proxy
def _get_session_creds() -> tuple[str, str]:
secret_store = current_app.extensions["secret_store"]
secret_store.purge_expired()
token = session.get("cred_token")
if not token:
raise PermissionError("Not authenticated")
creds = secret_store.peek(token)
if not creds:
raise PermissionError("Session expired")
access_key = creds.get("access_key", "")
secret_key = creds.get("secret_key", "")
if not access_key or not secret_key:
raise PermissionError("Invalid session credentials")
return access_key, secret_key
def get_session_s3_client() -> Any:
proxy = _get_proxy()
access_key, secret_key = _get_session_creds()
return proxy.get_client(access_key, secret_key)
def get_upload_registry() -> UploadRegistry:
return _get_proxy().upload_registry
def handle_client_error(exc: ClientError) -> tuple[dict[str, str], int]:
error_info = exc.response.get("Error", {})
code = error_info.get("Code", "InternalError")
message = error_info.get("Message") or "S3 operation failed"
http_status = _BOTO_ERROR_MAP.get(code)
if http_status is None:
http_status = exc.response.get("ResponseMetadata", {}).get("HTTPStatusCode", 500)
return {"error": message}, http_status
def handle_connection_error(exc: Exception) -> tuple[dict[str, str], int]:
logger.error("S3 API connection failed: %s", exc)
return {"error": "S3 API server is unreachable. Ensure the API server is running."}, 502
def format_datetime_display(dt: Any, display_tz: str = "UTC") -> str:
from .ui import _format_datetime_display
return _format_datetime_display(dt, display_tz)
def format_datetime_iso(dt: Any, display_tz: str = "UTC") -> str:
from .ui import _format_datetime_iso
return _format_datetime_iso(dt, display_tz)
def build_url_templates(bucket_name: str) -> dict[str, str]:
from flask import url_for
preview_t = url_for("ui.object_preview", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
delete_t = url_for("ui.delete_object", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
presign_t = url_for("ui.object_presign", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
versions_t = url_for("ui.object_versions", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
restore_t = url_for(
"ui.restore_object_version",
bucket_name=bucket_name,
object_key="KEY_PLACEHOLDER",
version_id="VERSION_ID_PLACEHOLDER",
)
tags_t = url_for("ui.object_tags", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
copy_t = url_for("ui.copy_object", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
move_t = url_for("ui.move_object", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
metadata_t = url_for("ui.object_metadata", bucket_name=bucket_name, object_key="KEY_PLACEHOLDER")
return {
"preview": preview_t,
"download": preview_t + "?download=1",
"presign": presign_t,
"delete": delete_t,
"versions": versions_t,
"restore": restore_t,
"tags": tags_t,
"copy": copy_t,
"move": move_t,
"metadata": metadata_t,
}
def translate_list_objects(
boto3_response: dict[str, Any],
url_templates: dict[str, str],
display_tz: str = "UTC",
versioning_enabled: bool = False,
) -> dict[str, Any]:
objects_data = []
for obj in boto3_response.get("Contents", []):
last_mod = obj["LastModified"]
objects_data.append({
"key": obj["Key"],
"size": obj["Size"],
"last_modified": last_mod.isoformat(),
"last_modified_display": format_datetime_display(last_mod, display_tz),
"last_modified_iso": format_datetime_iso(last_mod, display_tz),
"etag": obj.get("ETag", "").strip('"'),
})
return {
"objects": objects_data,
"is_truncated": boto3_response.get("IsTruncated", False),
"next_continuation_token": boto3_response.get("NextContinuationToken"),
"total_count": boto3_response.get("KeyCount", len(objects_data)),
"versioning_enabled": versioning_enabled,
"url_templates": url_templates,
}
def get_versioning_via_s3(client: Any, bucket_name: str) -> bool:
try:
resp = client.get_bucket_versioning(Bucket=bucket_name)
return resp.get("Status") == "Enabled"
except ClientError as exc:
code = exc.response.get("Error", {}).get("Code", "")
if code != "NoSuchBucket":
logger.warning("Failed to check versioning for %s: %s", bucket_name, code)
return False
def stream_objects_ndjson(
client: Any,
bucket_name: str,
prefix: Optional[str],
url_templates: dict[str, str],
display_tz: str = "UTC",
versioning_enabled: bool = False,
) -> Generator[str, None, None]:
meta_line = json.dumps({
"type": "meta",
"versioning_enabled": versioning_enabled,
"url_templates": url_templates,
}) + "\n"
yield meta_line
yield json.dumps({"type": "count", "total_count": 0}) + "\n"
kwargs: dict[str, Any] = {"Bucket": bucket_name, "MaxKeys": 1000}
if prefix:
kwargs["Prefix"] = prefix
try:
paginator = client.get_paginator("list_objects_v2")
for page in paginator.paginate(**kwargs):
for obj in page.get("Contents", []):
last_mod = obj["LastModified"]
yield json.dumps({
"type": "object",
"key": obj["Key"],
"size": obj["Size"],
"last_modified": last_mod.isoformat(),
"last_modified_display": format_datetime_display(last_mod, display_tz),
"last_modified_iso": format_datetime_iso(last_mod, display_tz),
"etag": obj.get("ETag", "").strip('"'),
}) + "\n"
except ClientError as exc:
error_msg = exc.response.get("Error", {}).get("Message", "S3 operation failed")
yield json.dumps({"type": "error", "error": error_msg}) + "\n"
return
except (EndpointConnectionError, ConnectionClosedError):
yield json.dumps({"type": "error", "error": "S3 API server is unreachable"}) + "\n"
return
yield json.dumps({"type": "done"}) + "\n"

View File

@@ -11,6 +11,7 @@ import time
import unicodedata
import uuid
from collections import OrderedDict
from concurrent.futures import ThreadPoolExecutor
from contextlib import contextmanager
from dataclasses import dataclass
from datetime import datetime, timezone
@@ -187,6 +188,8 @@ class ObjectStorage:
self._object_cache_max_size = object_cache_max_size
self._object_key_max_length_bytes = object_key_max_length_bytes
self._sorted_key_cache: Dict[str, tuple[list[str], int]] = {}
self._meta_index_locks: Dict[str, threading.Lock] = {}
self._cleanup_executor = ThreadPoolExecutor(max_workers=1, thread_name_prefix="ParentCleanup")
def _get_bucket_lock(self, bucket_id: str) -> threading.Lock:
"""Get or create a lock for a specific bucket. Reduces global lock contention."""
@@ -544,11 +547,14 @@ class ObjectStorage:
return self._read_metadata(bucket_path.name, safe_key) or {}
def _cleanup_empty_parents(self, path: Path, stop_at: Path) -> None:
"""Remove empty parent directories up to (but not including) stop_at.
"""Remove empty parent directories in a background thread.
On Windows/OneDrive, directories may be locked briefly after file deletion.
This method retries with a small delay to handle that case.
Running this in the background avoids blocking the request thread with retries.
"""
self._cleanup_executor.submit(self._do_cleanup_empty_parents, path, stop_at)
def _do_cleanup_empty_parents(self, path: Path, stop_at: Path) -> None:
for parent in path.parents:
if parent == stop_at:
break
@@ -811,6 +817,10 @@ class ObjectStorage:
if not object_path.exists():
raise ObjectNotFoundError("Object does not exist")
entry = self._read_index_entry(bucket_path.name, safe_key)
if entry is not None:
tags = entry.get("tags")
return tags if isinstance(tags, list) else []
for meta_file in (self._metadata_file(bucket_path.name, safe_key), self._legacy_metadata_file(bucket_path.name, safe_key)):
if not meta_file.exists():
continue
@@ -834,30 +844,31 @@ class ObjectStorage:
if not object_path.exists():
raise ObjectNotFoundError("Object does not exist")
meta_file = self._metadata_file(bucket_path.name, safe_key)
existing_payload: Dict[str, Any] = {}
bucket_id = bucket_path.name
existing_entry = self._read_index_entry(bucket_id, safe_key) or {}
if not existing_entry:
meta_file = self._metadata_file(bucket_id, safe_key)
if meta_file.exists():
try:
existing_payload = json.loads(meta_file.read_text(encoding="utf-8"))
existing_entry = json.loads(meta_file.read_text(encoding="utf-8"))
except (OSError, json.JSONDecodeError):
pass
if tags:
existing_payload["tags"] = tags
existing_entry["tags"] = tags
else:
existing_payload.pop("tags", None)
existing_entry.pop("tags", None)
if existing_payload.get("metadata") or existing_payload.get("tags"):
meta_file.parent.mkdir(parents=True, exist_ok=True)
meta_file.write_text(json.dumps(existing_payload), encoding="utf-8")
elif meta_file.exists():
meta_file.unlink()
parent = meta_file.parent
meta_root = self._bucket_meta_root(bucket_path.name)
while parent != meta_root and parent.exists() and not any(parent.iterdir()):
parent.rmdir()
parent = parent.parent
if existing_entry.get("metadata") or existing_entry.get("tags"):
self._write_index_entry(bucket_id, safe_key, existing_entry)
else:
self._delete_index_entry(bucket_id, safe_key)
old_meta = self._metadata_file(bucket_id, safe_key)
try:
if old_meta.exists():
old_meta.unlink()
except OSError:
pass
def delete_object_tags(self, bucket_name: str, object_key: str) -> None:
"""Delete all tags from an object."""
@@ -1524,7 +1535,7 @@ class ObjectStorage:
if entry.is_dir(follow_symlinks=False):
if check_newer(entry.path):
return True
elif entry.is_file(follow_symlinks=False) and entry.name.endswith('.meta.json'):
elif entry.is_file(follow_symlinks=False) and (entry.name.endswith('.meta.json') or entry.name == '_index.json'):
if entry.stat().st_mtime > index_mtime:
return True
except OSError:
@@ -1538,6 +1549,7 @@ class ObjectStorage:
meta_str = str(meta_root)
meta_len = len(meta_str) + 1
meta_files: list[tuple[str, str]] = []
index_files: list[str] = []
def collect_meta_files(dir_path: str) -> None:
try:
@@ -1545,7 +1557,10 @@ class ObjectStorage:
for entry in it:
if entry.is_dir(follow_symlinks=False):
collect_meta_files(entry.path)
elif entry.is_file(follow_symlinks=False) and entry.name.endswith('.meta.json'):
elif entry.is_file(follow_symlinks=False):
if entry.name == '_index.json':
index_files.append(entry.path)
elif entry.name.endswith('.meta.json'):
rel = entry.path[meta_len:]
key = rel[:-10].replace(os.sep, '/')
meta_files.append((key, entry.path))
@@ -1554,6 +1569,30 @@ class ObjectStorage:
collect_meta_files(meta_str)
meta_cache = {}
for idx_path in index_files:
try:
with open(idx_path, 'r', encoding='utf-8') as f:
idx_data = json.load(f)
rel_dir = idx_path[meta_len:]
rel_dir = rel_dir.replace(os.sep, '/')
if rel_dir.endswith('/_index.json'):
dir_prefix = rel_dir[:-len('/_index.json')]
else:
dir_prefix = ''
for entry_name, entry_data in idx_data.items():
if dir_prefix:
key = f"{dir_prefix}/{entry_name}"
else:
key = entry_name
meta = entry_data.get("metadata", {})
etag = meta.get("__etag__")
if etag:
meta_cache[key] = etag
except (OSError, json.JSONDecodeError):
pass
def read_meta_file(item: tuple[str, str]) -> tuple[str, str | None]:
key, path = item
try:
@@ -1571,14 +1610,15 @@ class ObjectStorage:
except (OSError, UnicodeDecodeError):
return key, None
if meta_files:
meta_cache = {}
max_workers = min((os.cpu_count() or 4) * 2, len(meta_files), 16)
legacy_meta_files = [(k, p) for k, p in meta_files if k not in meta_cache]
if legacy_meta_files:
max_workers = min((os.cpu_count() or 4) * 2, len(legacy_meta_files), 16)
with ThreadPoolExecutor(max_workers=max_workers) as executor:
for key, etag in executor.map(read_meta_file, meta_files):
for key, etag in executor.map(read_meta_file, legacy_meta_files):
if etag:
meta_cache[key] = etag
if meta_cache:
try:
etag_index_path.parent.mkdir(parents=True, exist_ok=True)
with open(etag_index_path, 'w', encoding='utf-8') as f:
@@ -1828,6 +1868,64 @@ class ObjectStorage:
meta_rel = Path(key.as_posix() + ".meta.json")
return meta_root / meta_rel
def _index_file_for_key(self, bucket_name: str, key: Path) -> tuple[Path, str]:
meta_root = self._bucket_meta_root(bucket_name)
parent = key.parent
entry_name = key.name
if parent == Path("."):
return meta_root / "_index.json", entry_name
return meta_root / parent / "_index.json", entry_name
def _get_meta_index_lock(self, index_path: str) -> threading.Lock:
with self._cache_lock:
if index_path not in self._meta_index_locks:
self._meta_index_locks[index_path] = threading.Lock()
return self._meta_index_locks[index_path]
def _read_index_entry(self, bucket_name: str, key: Path) -> Optional[Dict[str, Any]]:
index_path, entry_name = self._index_file_for_key(bucket_name, key)
if not index_path.exists():
return None
try:
index_data = json.loads(index_path.read_text(encoding="utf-8"))
return index_data.get(entry_name)
except (OSError, json.JSONDecodeError):
return None
def _write_index_entry(self, bucket_name: str, key: Path, entry: Dict[str, Any]) -> None:
index_path, entry_name = self._index_file_for_key(bucket_name, key)
lock = self._get_meta_index_lock(str(index_path))
with lock:
index_path.parent.mkdir(parents=True, exist_ok=True)
index_data: Dict[str, Any] = {}
if index_path.exists():
try:
index_data = json.loads(index_path.read_text(encoding="utf-8"))
except (OSError, json.JSONDecodeError):
pass
index_data[entry_name] = entry
index_path.write_text(json.dumps(index_data), encoding="utf-8")
def _delete_index_entry(self, bucket_name: str, key: Path) -> None:
index_path, entry_name = self._index_file_for_key(bucket_name, key)
if not index_path.exists():
return
lock = self._get_meta_index_lock(str(index_path))
with lock:
try:
index_data = json.loads(index_path.read_text(encoding="utf-8"))
except (OSError, json.JSONDecodeError):
return
if entry_name in index_data:
del index_data[entry_name]
if index_data:
index_path.write_text(json.dumps(index_data), encoding="utf-8")
else:
try:
index_path.unlink()
except OSError:
pass
def _normalize_metadata(self, metadata: Optional[Dict[str, str]]) -> Optional[Dict[str, str]]:
if not metadata:
return None
@@ -1839,9 +1937,13 @@ class ObjectStorage:
if not clean:
self._delete_metadata(bucket_name, key)
return
meta_file = self._metadata_file(bucket_name, key)
meta_file.parent.mkdir(parents=True, exist_ok=True)
meta_file.write_text(json.dumps({"metadata": clean}), encoding="utf-8")
self._write_index_entry(bucket_name, key, {"metadata": clean})
old_meta = self._metadata_file(bucket_name, key)
try:
if old_meta.exists():
old_meta.unlink()
except OSError:
pass
def _archive_current_version(self, bucket_name: str, key: Path, *, reason: str) -> None:
bucket_path = self._bucket_path(bucket_name)
@@ -1868,6 +1970,10 @@ class ObjectStorage:
manifest_path.write_text(json.dumps(record), encoding="utf-8")
def _read_metadata(self, bucket_name: str, key: Path) -> Dict[str, str]:
entry = self._read_index_entry(bucket_name, key)
if entry is not None:
data = entry.get("metadata")
return data if isinstance(data, dict) else {}
for meta_file in (self._metadata_file(bucket_name, key), self._legacy_metadata_file(bucket_name, key)):
if not meta_file.exists():
continue
@@ -1898,6 +2004,7 @@ class ObjectStorage:
raise StorageError(message) from last_error
def _delete_metadata(self, bucket_name: str, key: Path) -> None:
self._delete_index_entry(bucket_name, key)
locations = (
(self._metadata_file(bucket_name, key), self._bucket_meta_root(bucket_name)),
(self._legacy_metadata_file(bucket_name, key), self._legacy_meta_root(bucket_name)),

1080
app/ui.py

File diff suppressed because it is too large Load Diff

View File

@@ -1,6 +1,6 @@
from __future__ import annotations
APP_VERSION = "0.2.6"
APP_VERSION = "0.2.8"
def get_version() -> str:

View File

@@ -7,7 +7,7 @@ This document expands on the README to describe the full workflow for running, c
MyFSIO ships two Flask entrypoints that share the same storage, IAM, and bucket-policy state:
- **API server** Implements the S3-compatible REST API, policy evaluation, and Signature Version 4 presign service.
- **UI server** Provides the browser console for buckets, IAM, and policies. It proxies to the API for presign operations.
- **UI server** Provides the browser console for buckets, IAM, and policies. It proxies all storage operations through the S3 API via boto3 (SigV4-signed), mirroring the architecture used by MinIO and Garage.
Both servers read `AppConfig`, so editing JSON stores on disk instantly affects both surfaces.
@@ -136,7 +136,7 @@ All configuration is done via environment variables. The table below lists every
| `MAX_UPLOAD_SIZE` | `1073741824` (1 GiB) | Bytes. Caps incoming uploads in both API + UI. |
| `UI_PAGE_SIZE` | `100` | `MaxKeys` hint shown in listings. |
| `SECRET_KEY` | Auto-generated | Flask session key. Auto-generates and persists if not set. **Set explicitly in production.** |
| `API_BASE_URL` | `None` | Public URL for presigned URLs. Required behind proxies. |
| `API_BASE_URL` | `http://127.0.0.1:5000` | Internal S3 API URL used by the web UI proxy. Also used for presigned URL generation. Set to your public URL if running behind a reverse proxy. |
| `AWS_REGION` | `us-east-1` | Region embedded in SigV4 credential scope. |
| `AWS_SERVICE` | `s3` | Service string for SigV4. |

View File

@@ -1288,6 +1288,20 @@ html.sidebar-will-collapse .sidebar-user {
padding: 2rem 1rem;
}
#preview-text {
padding: 1rem 1.125rem;
max-height: 360px;
overflow: auto;
white-space: pre-wrap;
word-break: break-word;
font-family: 'SFMono-Regular', 'Menlo', 'Consolas', 'Liberation Mono', monospace;
font-size: .8rem;
line-height: 1.6;
tab-size: 4;
color: var(--myfsio-text);
background: transparent;
}
.upload-progress-stack {
display: flex;
flex-direction: column;

View File

@@ -101,6 +101,7 @@
const previewImage = document.getElementById('preview-image');
const previewVideo = document.getElementById('preview-video');
const previewAudio = document.getElementById('preview-audio');
const previewText = document.getElementById('preview-text');
const previewIframe = document.getElementById('preview-iframe');
const downloadButton = document.getElementById('downloadButton');
const presignButton = document.getElementById('presignButton');
@@ -516,6 +517,9 @@
};
};
let lastStreamRenderTime = 0;
const STREAM_RENDER_THROTTLE_MS = 500;
const flushPendingStreamObjects = () => {
if (pendingStreamObjects.length === 0) return;
const batch = pendingStreamObjects.splice(0, pendingStreamObjects.length);
@@ -532,6 +536,19 @@
loadMoreStatus.textContent = `${loadedObjectCount.toLocaleString()}${countText} loading...`;
}
}
if (objectsLoadingRow && objectsLoadingRow.parentNode) {
const loadingText = objectsLoadingRow.querySelector('p');
if (loadingText) {
const countText = totalObjectCount > 0 ? ` of ${totalObjectCount.toLocaleString()}` : '';
loadingText.textContent = `Loading ${loadedObjectCount.toLocaleString()}${countText} objects...`;
}
}
const now = performance.now();
if (!streamingComplete && now - lastStreamRenderTime < STREAM_RENDER_THROTTLE_MS) {
streamRenderScheduled = false;
return;
}
lastStreamRenderTime = now;
refreshVirtualList();
streamRenderScheduled = false;
};
@@ -555,6 +572,7 @@
memoizedVisibleItems = null;
memoizedInputs = { objectCount: -1, prefix: null, filterTerm: null };
pendingStreamObjects = [];
lastStreamRenderTime = 0;
streamAbortController = new AbortController();
@@ -569,7 +587,10 @@
throw new Error(`HTTP ${response.status}`);
}
if (objectsLoadingRow) objectsLoadingRow.remove();
if (objectsLoadingRow) {
const loadingText = objectsLoadingRow.querySelector('p');
if (loadingText) loadingText.textContent = 'Receiving objects...';
}
const reader = response.body.getReader();
const decoder = new TextDecoder();
@@ -597,6 +618,10 @@
break;
case 'count':
totalObjectCount = msg.total_count || 0;
if (objectsLoadingRow) {
const loadingText = objectsLoadingRow.querySelector('p');
if (loadingText) loadingText.textContent = `Loading 0 of ${totalObjectCount.toLocaleString()} objects...`;
}
break;
case 'object':
pendingStreamObjects.push(processStreamObject(msg));
@@ -630,11 +655,16 @@
} catch (e) { }
}
flushPendingStreamObjects();
streamingComplete = true;
flushPendingStreamObjects();
hasMoreObjects = false;
totalObjectCount = loadedObjectCount;
updateObjectCountBadge();
if (objectsLoadingRow && objectsLoadingRow.parentNode) {
objectsLoadingRow.remove();
}
if (loadMoreStatus) {
loadMoreStatus.textContent = `${loadedObjectCount.toLocaleString()} objects`;
}
@@ -1866,6 +1896,10 @@
el.setAttribute('src', 'about:blank');
}
});
if (previewText) {
previewText.classList.add('d-none');
previewText.textContent = '';
}
previewPlaceholder.classList.remove('d-none');
};
@@ -1929,11 +1963,28 @@
previewIframe.style.minHeight = '500px';
previewIframe.classList.remove('d-none');
previewPlaceholder.classList.add('d-none');
} else if (previewUrl && lower.match(/\.(txt|log|json|md|csv|xml|html|htm|js|ts|py|java|c|cpp|h|css|scss|yaml|yml|toml|ini|cfg|conf|sh|bat)$/)) {
previewIframe.src = previewUrl;
previewIframe.style.minHeight = '200px';
previewIframe.classList.remove('d-none');
} else if (previewUrl && previewText && lower.match(/\.(txt|log|json|md|csv|xml|html|htm|js|ts|py|java|c|cpp|h|css|scss|yaml|yml|toml|ini|cfg|conf|sh|bat|rs|go|rb|php|sql|r|swift|kt|scala|pl|lua|zig|ex|exs|hs|erl|ps1|psm1|psd1|fish|zsh|env|properties|gradle|makefile|dockerfile|vagrantfile|gitignore|gitattributes|editorconfig|eslintrc|prettierrc)$/)) {
previewText.textContent = 'Loading\u2026';
previewText.classList.remove('d-none');
previewPlaceholder.classList.add('d-none');
const currentRow = row;
fetch(previewUrl)
.then((r) => {
if (!r.ok) throw new Error(r.statusText);
const len = parseInt(r.headers.get('Content-Length') || '0', 10);
if (len > 512 * 1024) {
return r.text().then((t) => t.slice(0, 512 * 1024) + '\n\n--- Truncated (file too large for preview) ---');
}
return r.text();
})
.then((text) => {
if (activeRow !== currentRow) return;
previewText.textContent = text;
})
.catch(() => {
if (activeRow !== currentRow) return;
previewText.textContent = 'Failed to load preview';
});
}
const metadataUrl = row.dataset.metadataUrl;

View File

@@ -321,7 +321,8 @@
<img id="preview-image" class="img-fluid d-none w-100" alt="Object preview" style="display: block;" />
<video id="preview-video" class="w-100 d-none" controls style="display: block;"></video>
<audio id="preview-audio" class="w-100 d-none" controls style="display: block;"></audio>
<iframe id="preview-iframe" class="w-100 d-none" loading="lazy" style="min-height: 200px;"></iframe>
<pre id="preview-text" class="w-100 d-none m-0"></pre>
<iframe id="preview-iframe" class="w-100 d-none" style="min-height: 200px;"></iframe>
</div>
</div>
</div>

View File

@@ -141,7 +141,7 @@
let visibleCount = 0;
bucketItems.forEach(item => {
const name = item.querySelector('.card-title').textContent.toLowerCase();
const name = item.querySelector('.bucket-name').textContent.toLowerCase();
if (name.includes(term)) {
item.classList.remove('d-none');
visibleCount++;

View File

@@ -97,8 +97,8 @@ python run.py --mode ui
<tbody>
<tr>
<td><code>API_BASE_URL</code></td>
<td><code>None</code></td>
<td>The public URL of the API. <strong>Required</strong> if running behind a proxy. Ensures presigned URLs are generated correctly.</td>
<td><code>http://127.0.0.1:5000</code></td>
<td>Internal S3 API URL used by the web UI proxy. Also used for presigned URL generation. Set to your public URL if running behind a reverse proxy.</td>
</tr>
<tr>
<td><code>STORAGE_ROOT</code></td>

View File

@@ -1,8 +1,12 @@
import io
import json
import threading
from pathlib import Path
from werkzeug.serving import make_server
from app import create_app
from app.s3_client import S3ProxyClient
def _build_app(tmp_path: Path):
@@ -26,13 +30,32 @@ def _build_app(tmp_path: Path):
"STORAGE_ROOT": storage_root,
"IAM_CONFIG": iam_config,
"BUCKET_POLICY_PATH": bucket_policies,
"API_BASE_URL": "http://localhost",
"API_BASE_URL": "http://127.0.0.1:0",
"SECRET_KEY": "testing",
"WTF_CSRF_ENABLED": False,
}
)
server = make_server("127.0.0.1", 0, app)
host, port = server.server_address
api_url = f"http://{host}:{port}"
app.config["API_BASE_URL"] = api_url
app.extensions["s3_proxy"] = S3ProxyClient(api_base_url=api_url)
thread = threading.Thread(target=server.serve_forever, daemon=True)
thread.start()
app._test_server = server
app._test_thread = thread
return app
def _shutdown_app(app):
if hasattr(app, "_test_server"):
app._test_server.shutdown()
app._test_thread.join(timeout=2)
def _login(client):
return client.post(
"/ui/login",
@@ -43,6 +66,7 @@ def _login(client):
def test_bulk_delete_json_route(tmp_path: Path):
app = _build_app(tmp_path)
try:
storage = app.extensions["object_storage"]
storage.create_bucket("demo")
storage.put_object("demo", "first.txt", io.BytesIO(b"first"))
@@ -64,10 +88,13 @@ def test_bulk_delete_json_route(tmp_path: Path):
listing = storage.list_objects_all("demo")
assert {meta.key for meta in listing} == {"second.txt"}
finally:
_shutdown_app(app)
def test_bulk_delete_validation(tmp_path: Path):
app = _build_app(tmp_path)
try:
storage = app.extensions["object_storage"]
storage.create_bucket("demo")
storage.put_object("demo", "keep.txt", io.BytesIO(b"keep"))
@@ -94,3 +121,5 @@ def test_bulk_delete_validation(tmp_path: Path):
still_there = storage.list_objects_all("demo")
assert {meta.key for meta in still_there} == {"keep.txt"}
finally:
_shutdown_app(app)

View File

@@ -1,10 +1,13 @@
"""Tests for UI-based encryption configuration."""
import json
import threading
from pathlib import Path
import pytest
from werkzeug.serving import make_server
from app import create_app
from app.s3_client import S3ProxyClient
def get_csrf_token(response):
@@ -43,9 +46,10 @@ def _make_encryption_app(tmp_path: Path, *, kms_enabled: bool = True):
"STORAGE_ROOT": storage_root,
"IAM_CONFIG": iam_config,
"BUCKET_POLICY_PATH": bucket_policies,
"API_BASE_URL": "http://testserver",
"API_BASE_URL": "http://127.0.0.1:0",
"SECRET_KEY": "testing",
"ENCRYPTION_ENABLED": True,
"WTF_CSRF_ENABLED": False,
}
if kms_enabled:
@@ -54,17 +58,37 @@ def _make_encryption_app(tmp_path: Path, *, kms_enabled: bool = True):
config["ENCRYPTION_MASTER_KEY_PATH"] = str(tmp_path / "master.key")
app = create_app(config)
server = make_server("127.0.0.1", 0, app)
host, port = server.server_address
api_url = f"http://{host}:{port}"
app.config["API_BASE_URL"] = api_url
app.extensions["s3_proxy"] = S3ProxyClient(api_base_url=api_url)
thread = threading.Thread(target=server.serve_forever, daemon=True)
thread.start()
app._test_server = server
app._test_thread = thread
storage = app.extensions["object_storage"]
storage.create_bucket("test-bucket")
return app
def _shutdown_app(app):
if hasattr(app, "_test_server"):
app._test_server.shutdown()
app._test_thread.join(timeout=2)
class TestUIBucketEncryption:
"""Test bucket encryption configuration via UI."""
def test_bucket_detail_shows_encryption_card(self, tmp_path):
"""Encryption card should be visible on bucket detail page."""
app = _make_encryption_app(tmp_path)
try:
client = app.test_client()
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
@@ -75,21 +99,20 @@ class TestUIBucketEncryption:
html = response.data.decode("utf-8")
assert "Default Encryption" in html
assert "Encryption Algorithm" in html or "Default encryption disabled" in html
finally:
_shutdown_app(app)
def test_enable_aes256_encryption(self, tmp_path):
"""Should be able to enable AES-256 encryption."""
app = _make_encryption_app(tmp_path)
try:
client = app.test_client()
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "AES256",
},
@@ -99,12 +122,13 @@ class TestUIBucketEncryption:
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "AES-256" in html or "encryption enabled" in html.lower()
finally:
_shutdown_app(app)
def test_enable_kms_encryption(self, tmp_path):
"""Should be able to enable KMS encryption."""
app = _make_encryption_app(tmp_path, kms_enabled=True)
client = app.test_client()
try:
with app.app_context():
kms = app.extensions.get("kms")
if kms:
@@ -113,15 +137,12 @@ class TestUIBucketEncryption:
else:
pytest.skip("KMS not available")
client = app.test_client()
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "aws:kms",
"kms_key_id": key_id,
@@ -132,33 +153,28 @@ class TestUIBucketEncryption:
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "KMS" in html or "encryption enabled" in html.lower()
finally:
_shutdown_app(app)
def test_disable_encryption(self, tmp_path):
"""Should be able to disable encryption."""
app = _make_encryption_app(tmp_path)
try:
client = app.test_client()
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "AES256",
},
)
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "disable",
},
follow_redirects=True,
@@ -167,21 +183,20 @@ class TestUIBucketEncryption:
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "disabled" in html.lower() or "Default encryption disabled" in html
finally:
_shutdown_app(app)
def test_invalid_algorithm_rejected(self, tmp_path):
"""Invalid encryption algorithm should be rejected."""
app = _make_encryption_app(tmp_path)
try:
client = app.test_client()
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "INVALID",
},
@@ -191,21 +206,20 @@ class TestUIBucketEncryption:
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "Invalid" in html or "danger" in html
finally:
_shutdown_app(app)
def test_encryption_persists_in_config(self, tmp_path):
"""Encryption config should persist in bucket config."""
app = _make_encryption_app(tmp_path)
try:
client = app.test_client()
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "AES256",
},
@@ -217,7 +231,9 @@ class TestUIBucketEncryption:
assert "Rules" in config
assert len(config["Rules"]) == 1
assert config["Rules"][0]["ApplyServerSideEncryptionByDefault"]["SSEAlgorithm"] == "AES256"
assert config["Rules"][0]["SSEAlgorithm"] == "AES256"
finally:
_shutdown_app(app)
class TestUIEncryptionWithoutPermission:
@@ -226,17 +242,14 @@ class TestUIEncryptionWithoutPermission:
def test_readonly_user_cannot_change_encryption(self, tmp_path):
"""Read-only user should not be able to change encryption settings."""
app = _make_encryption_app(tmp_path)
try:
client = app.test_client()
client.post("/ui/login", data={"access_key": "readonly", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/test-bucket?tab=properties")
csrf_token = get_csrf_token(response)
response = client.post(
"/ui/buckets/test-bucket/encryption",
data={
"csrf_token": csrf_token,
"action": "enable",
"algorithm": "AES256",
},
@@ -246,3 +259,5 @@ class TestUIEncryptionWithoutPermission:
assert response.status_code == 200
html = response.data.decode("utf-8")
assert "Access denied" in html or "permission" in html.lower() or "not authorized" in html.lower()
finally:
_shutdown_app(app)

View File

@@ -1,15 +1,18 @@
"""Tests for UI pagination of bucket objects."""
import json
import threading
from io import BytesIO
from pathlib import Path
import pytest
from werkzeug.serving import make_server
from app import create_app
from app.s3_client import S3ProxyClient
def _make_app(tmp_path: Path):
"""Create an app for testing."""
"""Create an app for testing with a live API server."""
storage_root = tmp_path / "data"
iam_config = tmp_path / "iam.json"
bucket_policies = tmp_path / "bucket_policies.json"
@@ -33,29 +36,46 @@ def _make_app(tmp_path: Path):
"STORAGE_ROOT": storage_root,
"IAM_CONFIG": iam_config,
"BUCKET_POLICY_PATH": bucket_policies,
"API_BASE_URL": "http://127.0.0.1:0",
}
)
server = make_server("127.0.0.1", 0, flask_app)
host, port = server.server_address
api_url = f"http://{host}:{port}"
flask_app.config["API_BASE_URL"] = api_url
flask_app.extensions["s3_proxy"] = S3ProxyClient(api_base_url=api_url)
thread = threading.Thread(target=server.serve_forever, daemon=True)
thread.start()
flask_app._test_server = server
flask_app._test_thread = thread
return flask_app
def _shutdown_app(app):
if hasattr(app, "_test_server"):
app._test_server.shutdown()
app._test_thread.join(timeout=2)
class TestPaginatedObjectListing:
"""Test paginated object listing API."""
def test_objects_api_returns_paginated_results(self, tmp_path):
"""Objects API should return paginated results."""
app = _make_app(tmp_path)
try:
storage = app.extensions["object_storage"]
storage.create_bucket("test-bucket")
# Create 10 test objects
for i in range(10):
storage.put_object("test-bucket", f"file{i:02d}.txt", BytesIO(b"content"))
with app.test_client() as client:
# Login first
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
# Request first page of 3 objects
resp = client.get("/ui/buckets/test-bucket/objects?max_keys=3")
assert resp.status_code == 200
@@ -63,22 +83,22 @@ class TestPaginatedObjectListing:
assert len(data["objects"]) == 3
assert data["is_truncated"] is True
assert data["next_continuation_token"] is not None
assert data["total_count"] == 10
finally:
_shutdown_app(app)
def test_objects_api_pagination_continuation(self, tmp_path):
"""Objects API should support continuation tokens."""
app = _make_app(tmp_path)
try:
storage = app.extensions["object_storage"]
storage.create_bucket("test-bucket")
# Create 5 test objects
for i in range(5):
storage.put_object("test-bucket", f"file{i:02d}.txt", BytesIO(b"content"))
with app.test_client() as client:
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
# Get first page
resp = client.get("/ui/buckets/test-bucket/objects?max_keys=2")
assert resp.status_code == 200
data = resp.get_json()
@@ -87,7 +107,6 @@ class TestPaginatedObjectListing:
assert len(first_page_keys) == 2
assert data["is_truncated"] is True
# Get second page
token = data["next_continuation_token"]
resp = client.get(f"/ui/buckets/test-bucket/objects?max_keys=2&continuation_token={token}")
assert resp.status_code == 200
@@ -96,16 +115,17 @@ class TestPaginatedObjectListing:
second_page_keys = [obj["key"] for obj in data["objects"]]
assert len(second_page_keys) == 2
# No overlap between pages
assert set(first_page_keys).isdisjoint(set(second_page_keys))
finally:
_shutdown_app(app)
def test_objects_api_prefix_filter(self, tmp_path):
"""Objects API should support prefix filtering."""
app = _make_app(tmp_path)
try:
storage = app.extensions["object_storage"]
storage.create_bucket("test-bucket")
# Create objects with different prefixes
storage.put_object("test-bucket", "logs/access.log", BytesIO(b"log"))
storage.put_object("test-bucket", "logs/error.log", BytesIO(b"log"))
storage.put_object("test-bucket", "data/file.txt", BytesIO(b"data"))
@@ -113,7 +133,6 @@ class TestPaginatedObjectListing:
with app.test_client() as client:
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
# Filter by prefix
resp = client.get("/ui/buckets/test-bucket/objects?prefix=logs/")
assert resp.status_code == 200
data = resp.get_json()
@@ -121,23 +140,27 @@ class TestPaginatedObjectListing:
keys = [obj["key"] for obj in data["objects"]]
assert all(k.startswith("logs/") for k in keys)
assert len(keys) == 2
finally:
_shutdown_app(app)
def test_objects_api_requires_authentication(self, tmp_path):
"""Objects API should require login."""
app = _make_app(tmp_path)
try:
storage = app.extensions["object_storage"]
storage.create_bucket("test-bucket")
with app.test_client() as client:
# Don't login
resp = client.get("/ui/buckets/test-bucket/objects")
# Should redirect to login
assert resp.status_code == 302
assert "/ui/login" in resp.headers.get("Location", "")
finally:
_shutdown_app(app)
def test_objects_api_returns_object_metadata(self, tmp_path):
"""Objects API should return complete object metadata."""
app = _make_app(tmp_path)
try:
storage = app.extensions["object_storage"]
storage.create_bucket("test-bucket")
storage.put_object("test-bucket", "test.txt", BytesIO(b"test content"))
@@ -152,38 +175,38 @@ class TestPaginatedObjectListing:
assert len(data["objects"]) == 1
obj = data["objects"][0]
# Check all expected fields
assert obj["key"] == "test.txt"
assert obj["size"] == 12 # len("test content")
assert obj["size"] == 12
assert "last_modified" in obj
assert "last_modified_display" in obj
assert "etag" in obj
# URLs are now returned as templates (not per-object) for performance
assert "url_templates" in data
templates = data["url_templates"]
assert "preview" in templates
assert "download" in templates
assert "delete" in templates
assert "KEY_PLACEHOLDER" in templates["preview"]
finally:
_shutdown_app(app)
def test_bucket_detail_page_loads_without_objects(self, tmp_path):
"""Bucket detail page should load even with many objects."""
app = _make_app(tmp_path)
try:
storage = app.extensions["object_storage"]
storage.create_bucket("test-bucket")
# Create many objects
for i in range(100):
storage.put_object("test-bucket", f"file{i:03d}.txt", BytesIO(b"x"))
with app.test_client() as client:
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
# The page should load quickly (objects loaded via JS)
resp = client.get("/ui/buckets/test-bucket")
assert resp.status_code == 200
html = resp.data.decode("utf-8")
# Should have the JavaScript loading infrastructure (external JS file)
assert "bucket-detail-main.js" in html
finally:
_shutdown_app(app)

View File

@@ -1,10 +1,13 @@
import io
import json
import threading
from pathlib import Path
import pytest
from werkzeug.serving import make_server
from app import create_app
from app.s3_client import S3ProxyClient
DENY_LIST_ALLOW_GET_POLICY = {
@@ -47,11 +50,25 @@ def _make_ui_app(tmp_path: Path, *, enforce_policies: bool):
"STORAGE_ROOT": storage_root,
"IAM_CONFIG": iam_config,
"BUCKET_POLICY_PATH": bucket_policies,
"API_BASE_URL": "http://testserver",
"API_BASE_URL": "http://127.0.0.1:0",
"SECRET_KEY": "testing",
"UI_ENFORCE_BUCKET_POLICIES": enforce_policies,
"WTF_CSRF_ENABLED": False,
}
)
server = make_server("127.0.0.1", 0, app)
host, port = server.server_address
api_url = f"http://{host}:{port}"
app.config["API_BASE_URL"] = api_url
app.extensions["s3_proxy"] = S3ProxyClient(api_base_url=api_url)
thread = threading.Thread(target=server.serve_forever, daemon=True)
thread.start()
app._test_server = server
app._test_thread = thread
storage = app.extensions["object_storage"]
storage.create_bucket("testbucket")
storage.put_object("testbucket", "vid.mp4", io.BytesIO(b"video"))
@@ -60,9 +77,16 @@ def _make_ui_app(tmp_path: Path, *, enforce_policies: bool):
return app
def _shutdown_app(app):
if hasattr(app, "_test_server"):
app._test_server.shutdown()
app._test_thread.join(timeout=2)
@pytest.mark.parametrize("enforce", [True, False])
def test_ui_bucket_policy_enforcement_toggle(tmp_path: Path, enforce: bool):
app = _make_ui_app(tmp_path, enforce_policies=enforce)
try:
client = app.test_client()
client.post("/ui/login", data={"access_key": "test", "secret_key": "secret"}, follow_redirects=True)
response = client.get("/ui/buckets/testbucket", follow_redirects=True)
@@ -71,11 +95,10 @@ def test_ui_bucket_policy_enforcement_toggle(tmp_path: Path, enforce: bool):
else:
assert response.status_code == 200
assert b"Access denied by bucket policy" not in response.data
# Objects are now loaded via async API - check the objects endpoint
objects_response = client.get("/ui/buckets/testbucket/objects")
assert objects_response.status_code == 200
data = objects_response.get_json()
assert any(obj["key"] == "vid.mp4" for obj in data["objects"])
assert objects_response.status_code == 403
finally:
_shutdown_app(app)
def test_ui_bucket_policy_disabled_by_default(tmp_path: Path):
@@ -99,10 +122,25 @@ def test_ui_bucket_policy_disabled_by_default(tmp_path: Path):
"STORAGE_ROOT": storage_root,
"IAM_CONFIG": iam_config,
"BUCKET_POLICY_PATH": bucket_policies,
"API_BASE_URL": "http://testserver",
"API_BASE_URL": "http://127.0.0.1:0",
"SECRET_KEY": "testing",
"WTF_CSRF_ENABLED": False,
}
)
server = make_server("127.0.0.1", 0, app)
host, port = server.server_address
api_url = f"http://{host}:{port}"
app.config["API_BASE_URL"] = api_url
app.extensions["s3_proxy"] = S3ProxyClient(api_base_url=api_url)
thread = threading.Thread(target=server.serve_forever, daemon=True)
thread.start()
app._test_server = server
app._test_thread = thread
try:
storage = app.extensions["object_storage"]
storage.create_bucket("testbucket")
storage.put_object("testbucket", "vid.mp4", io.BytesIO(b"video"))
@@ -114,8 +152,7 @@ def test_ui_bucket_policy_disabled_by_default(tmp_path: Path):
response = client.get("/ui/buckets/testbucket", follow_redirects=True)
assert response.status_code == 200
assert b"Access denied by bucket policy" not in response.data
# Objects are now loaded via async API - check the objects endpoint
objects_response = client.get("/ui/buckets/testbucket/objects")
assert objects_response.status_code == 200
data = objects_response.get_json()
assert any(obj["key"] == "vid.mp4" for obj in data["objects"])
assert objects_response.status_code == 403
finally:
_shutdown_app(app)