WebDAV sync support, access tokens, REST control endpoints (#10)

Implement complete WebDAV file serving compatible with various clients from Windows File Explorer to more specialized sync tools. The old control WebSocket has been updated to part-DAV, part REST API instead. Implemented user:pass BASIC auth. Added UI and backend for creating tokens that avoid the need to use actual username and password for requests from CLI or DAV.
This commit is contained in:
2026-04-26 04:22:52 +00:00
parent 9ac571463a
commit f8b2c9494a
31 changed files with 3377 additions and 529 deletions
+22 -13
View File
@@ -7,9 +7,13 @@ from sanic import Blueprint, json
from sanic.exceptions import BadRequest
from cista import __version__, auth, config, sso, watching
from cista.auth import (
create_token_handler,
delete_token_handler,
list_tokens_handler,
)
from cista.fileio import FileServer
from cista.protocol import ControlTypes, StatusMsg
from cista.util.apphelpers import asend, websocket_wrapper
from cista.util.apphelpers import websocket_wrapper
bp = Blueprint("api", url_prefix="/api")
fileserver = FileServer()
@@ -25,17 +29,6 @@ async def stop_fileserver(app):
await fileserver.stop()
@bp.websocket("control")
@websocket_wrapper
async def control(req, ws):
while True:
cmd = msgspec.json.decode(await ws.recv(), type=ControlTypes)
await asyncio.to_thread(cmd)
# Signal the watcher about affected paths
watching.notify_change(*cmd.affected_paths())
await asend(ws, StatusMsg(status="ack", req=cmd))
@bp.websocket("watch")
@websocket_wrapper
async def watch(req, ws):
@@ -144,3 +137,19 @@ async def update_name(request):
# Return the effective name (fallback to path.name if empty)
effective_name = name or config.config.path.name
return json({"message": "Server name updated", "name": effective_name})
# Token management endpoints (available in all modes; primary path in SSO mode)
@bp.get("tokens")
async def list_api_tokens(request):
return await list_tokens_handler(request)
@bp.post("tokens")
async def create_api_token(request):
return await create_token_handler(request)
@bp.delete("tokens/<token_id>")
async def delete_api_token(request, token_id):
return await delete_token_handler(request, token_id)
+78 -164
View File
@@ -1,19 +1,16 @@
import asyncio
import datetime
import mimetypes
import re
import time
from concurrent.futures import ThreadPoolExecutor
from multiprocessing import cpu_count
from pathlib import Path, PurePath, PurePosixPath
from stat import S_IFDIR, S_IFREG
from urllib.parse import unquote
from wsgiref.handlers import format_date_time
import sanic.helpers
from blake3 import blake3
from sanic import Blueprint, Sanic, empty, json, raw, redirect
from sanic.exceptions import BadRequest, Forbidden, NotFound
from sanic import Sanic, empty, raw, redirect
from sanic.exceptions import Forbidden, NotFound
from sanic.log import logger
from setproctitle import setproctitle
from stream_zip import ZIP_AUTO, stream_zip
@@ -21,18 +18,88 @@ from zstandard import ZstdCompressor
from cista import auth, config, preview, session, sso, watching
from cista.preview import shutdown_preview_workers, start_preview_workers
from cista.api import bp, fileserver
from cista.sanic_logging import configure_access_logging, configure_main_logging, format_access_log
from cista.api import bp
from cista import fileserver
from cista.sanic_logging import (
configure_access_logging,
configure_main_logging,
format_access_log,
)
from cista.sanic_logging import logger as access_logger
from cista.util.apphelpers import handle_sanic_exception
# Workaround until Sanic PR #2824 is merged
sanic.helpers._ENTITY_HEADERS = frozenset()
configure_access_logging()
app = Sanic("cista", strict_slashes=True)
app.router.ALLOWED_METHODS = (
*app.router.ALLOWED_METHODS,
"MKCOL",
"MOVE",
"COPY",
"PROPFIND",
)
configure_main_logging()
@app.on_request
async def use_session(req):
req.ctx._log_start = time.perf_counter()
req.ctx._auth_flow = ["session: start"]
auth.hydrate_request_auth_context(req, source="app.on_request")
# CSRF protection
if req.method == "GET" and req.headers.upgrade != "websocket":
return # Ordinary GET requests are fine
# Check that origin matches host, for browsers which should all send Origin.
# Curl doesn't send any Origin header, so we allow it anyway.
origin = req.headers.origin
if origin and origin.split("//", 1)[1] != req.host:
raise Forbidden("Invalid origin: Cross-Site requests not permitted")
@app.on_response
async def log_access(req, res):
"""Log HTTP access in a clean single-line format."""
if req.headers.get("upgrade", "").lower() == "websocket":
return res
start = getattr(req.ctx, "_log_start", None)
duration_ms = (time.perf_counter() - start) * 1000 if start is not None else 0.0
client = req.client_ip or "-"
host = req.host or "-"
path = req.path
if req.query_string:
qs = req.query_string
if isinstance(qs, bytes):
qs = qs.decode(errors="replace")
path = f"{path}?{qs}"
extra = getattr(req.ctx, "_log_extra", None)
line = format_access_log(
client, res.status, req.method, host, path, duration_ms, extra=extra
)
access_logger.info(line)
return res
@app.on_response
async def forward_sso_cookies(req, res):
"""Forward Set-Cookie headers from SSO validation to client."""
if cookies := getattr(req.ctx, "sso_cookies", None):
for cookie in cookies:
res.headers.add("set-cookie", cookie)
@app.on_response
async def persist_auth_session(req, res):
"""Persist a session cookie after successful Authorization-based auth."""
username = getattr(req.ctx, "_create_session_username", None)
if not username or res.status >= 400:
return
existing = getattr(req.ctx, "session", None)
if isinstance(existing, dict) and existing.get("username") == username:
return
session.create(res, username, secure=req.scheme == "https")
# Register either SSO proxy or built-in auth routes based on PASKIA_BACKEND_URL
if sso.paskia_enabled():
app.blueprint(sso.bp) # SSO proxy for /auth/* routes
@@ -40,6 +107,7 @@ else:
app.blueprint(auth.bp) # Built-in auth routes
app.blueprint(preview.bp)
app.blueprint(bp)
app.blueprint(fileserver.bp)
app.exception(Exception)(handle_sanic_exception)
@@ -70,161 +138,7 @@ async def main_stop(app):
logger.debug("Cista worker threads all finished")
@app.on_request
async def use_session(req):
req.ctx._log_start = time.perf_counter()
req.ctx.session = session.get(req)
try:
req.ctx.username = req.ctx.session["username"] # type: ignore
req.ctx.user = config.config.users[req.ctx.username]
except (AttributeError, KeyError, TypeError):
req.ctx.username = None
req.ctx.user = None
# CSRF protection
if req.method == "GET" and req.headers.upgrade != "websocket":
return # Ordinary GET requests are fine
# Check that origin matches host, for browsers which should all send Origin.
# Curl doesn't send any Origin header, so we allow it anyway.
origin = req.headers.origin
if origin and origin.split("//", 1)[1] != req.host:
raise Forbidden("Invalid origin: Cross-Site requests not permitted")
@app.on_response
async def log_access(req, res):
"""Log HTTP access in a clean single-line format."""
if req.headers.get("upgrade", "").lower() == "websocket":
return res
start = getattr(req.ctx, "_log_start", None)
duration_ms = (time.perf_counter() - start) * 1000 if start is not None else 0.0
client = req.client_ip or "-"
host = req.host or "-"
path = req.path
if req.query_string:
qs = req.query_string
if isinstance(qs, bytes):
qs = qs.decode(errors="replace")
path = f"{path}?{qs}"
extra = getattr(req.ctx, "_log_extra", None)
line = format_access_log(client, res.status, req.method, host, path, duration_ms, extra=extra)
access_logger.info(line)
return res
@app.on_response
async def forward_sso_cookies(req, res):
"""Forward Set-Cookie headers from SSO validation to client."""
if cookies := getattr(req.ctx, "sso_cookies", None):
for cookie in cookies:
res.headers.add("set-cookie", cookie)
@app.before_server_start
def http_fileserver(app):
bp = Blueprint("fileserver")
@bp.on_request
async def verify_fileserver(request):
"""Verify access to file server routes."""
await auth.verify(request)
@bp.put("/files/<name:path>")
async def upload_file_chunk(request, *args, **kwargs):
body = request.body
header = request.headers.get("content-range")
if header:
start, end, total = _parse_content_range(header, len(body))
else:
start = 0
end = len(body)
total = end
raw_name = kwargs.get("name")
if raw_name is None and args:
raw_name = args[0]
if not isinstance(raw_name, str) or not raw_name:
prefix = "/files/"
if not request.path.startswith(prefix):
raise BadRequest("Invalid upload path")
raw_name = request.path[len(prefix) :]
rel_name = unquote(raw_name)
upload_info = await asyncio.to_thread(
fileserver.upload_info,
rel_name,
start,
body,
total,
)
extras = []
chunk_len = end - start
whole_file = start == 0 and end == total
if not whole_file:
start_mib = _to_mib_int(start)
chunk_mib = _to_mib_int(chunk_len)
# Keep range logs compact for fixed-size upload blocks.
if chunk_mib == 16:
extras.append(f"{start_mib}MiB")
else:
extras.append(f"{start_mib}+{chunk_mib}MiB")
if upload_info.get("created"):
extras.append(f"created {_to_mib_int(total)}MiB")
size_before = upload_info.get("size_before")
size_after = upload_info.get("size_after")
if (
size_before is not None
and size_after is not None
and size_before != size_after
):
extras.append("resized")
request.ctx._log_extra = " ".join(extras) if extras else None
path = PurePosixPath(rel_name)
watching.notify_change(path, *path.parents)
return json(
{
"status": "ack",
"req": {
"name": rel_name,
"size": total,
"start": start,
"end": end,
},
}
)
bp.static(
"/files/",
config.config.path,
use_content_range=True,
stream_large_files=True,
directory_view=True,
)
app.blueprint(bp)
www = {}
_CONTENT_RANGE_RE = re.compile(r"^bytes (\d+)-(\d+)/(\d+)$")
def _parse_content_range(header: str, body_len: int) -> tuple[int, int, int]:
m = _CONTENT_RANGE_RE.fullmatch(header.strip())
if m is None:
raise BadRequest("Invalid Content-Range format")
start, end_inclusive, total = (int(v) for v in m.groups())
if total <= 0:
raise BadRequest("Invalid Content-Range total size")
if start > end_inclusive:
raise BadRequest("Invalid Content-Range range")
if end_inclusive >= total:
raise BadRequest("Content-Range exceeds total size")
expected_len = end_inclusive - start + 1
if expected_len != body_len:
raise BadRequest(
f"Content length mismatch for range: expected {expected_len}, got {body_len}"
)
return start, end_inclusive + 1, total
def _to_mib_int(value_bytes: int) -> int:
return round(value_bytes / (1 << 20))
def _load_wwwroot(www):
+917 -5
View File
File diff suppressed because it is too large Load Diff
+35
View File
@@ -22,6 +22,7 @@ class Config(msgspec.Struct):
name: str = ""
users: dict[str, User] = {}
links: dict[str, Link] = {}
tokens: dict[str, Token] = {}
# Typing: arguments for config-modifying functions
@@ -43,6 +44,14 @@ class Link(msgspec.Struct, omit_defaults=True):
expires: int = 0
class Token(msgspec.Struct, omit_defaults=True):
key: str = "" # plain text secret (shown once on creation)
username: str = "" # set in built-in mode
sso_user_id: str = "" # set in SSO mode
name: str = ""
created: int = 0 # noqa: N815
# Global variables - initialized during application startup
config: Config
conffile: Path
@@ -204,3 +213,29 @@ def del_user(conf: Config, name: str) -> Config:
settings = msgspec.to_builtins(conf, enc_hook=enc_hook)
settings["users"].pop(name)
return msgspec.convert(settings, Config, dec_hook=dec_hook)
@modifies_config
def update_token(conf: Config, token_id: str, changes: dict) -> Config:
"""Create or update a token."""
try:
t = msgspec.convert(
msgspec.to_builtins(conf.tokens[token_id], enc_hook=enc_hook),
Token,
dec_hook=dec_hook,
)
except KeyError:
t = Token()
tdict = msgspec.to_builtins(t, enc_hook=enc_hook)
tdict.update(changes)
settings = msgspec.to_builtins(conf, enc_hook=enc_hook)
settings["tokens"][token_id] = msgspec.convert(tdict, Token, dec_hook=dec_hook)
return msgspec.convert(settings, Config, dec_hook=dec_hook)
@modifies_config
def del_token(conf: Config, token_id: str) -> Config:
"""Delete a token by its stable id."""
settings = msgspec.to_builtins(conf, enc_hook=enc_hook)
settings["tokens"].pop(token_id, None)
return msgspec.convert(settings, Config, dec_hook=dec_hook)
+605
View File
@@ -0,0 +1,605 @@
import asyncio
import mimetypes
import os
import re
import shutil
import xml.etree.ElementTree as ET
from datetime import datetime, timezone
from pathlib import Path, PurePosixPath
from urllib.parse import quote as url_quote, unquote, urlparse
from wsgiref.handlers import format_date_time
from sanic import Blueprint, HTTPResponse, empty, json
from sanic.exceptions import BadRequest, NotFound
from cista import auth, config, watching
from cista.api import fileserver
from cista.util import filename
bp = Blueprint("fileserver", url_prefix="/files")
_CONTENT_RANGE_RE = re.compile(r"^bytes (\d+)-(\d+)/(\d+)$")
_RANGE_RE = re.compile(r"^bytes=(\d*)-(\d*)$")
_FILE_CHUNK_SIZE = 1 << 20
_DAV_NS = "DAV:"
ET.register_namespace("D", _DAV_NS)
def _dav_tag(name: str) -> str:
return f"{{{_DAV_NS}}}{name}"
@bp.on_request
async def verify_fileserver(request):
"""Verify access to file server routes."""
await auth.verify(request)
@bp.put("/<name:path>")
async def upload_file_chunk(request, name):
body = request.body
header = request.headers.get("content-range")
if header:
start, end, total = _parse_content_range(header, len(body))
else:
start = 0
end = len(body)
total = end
rel, _ = _safe_relpath(name)
rel_name = rel.as_posix()
upload_info = await asyncio.to_thread(
fileserver.upload_info,
rel_name,
start,
body,
total,
)
extras = []
chunk_len = end - start
whole_file = start == 0 and end == total
if not whole_file:
start_mib = _to_mib_int(start)
chunk_mib = _to_mib_int(chunk_len)
# Keep range logs compact for fixed-size upload blocks.
if chunk_mib == 16:
extras.append(f"{start_mib}MiB")
else:
extras.append(f"{start_mib}+{chunk_mib}MiB")
if upload_info.get("created"):
extras.append(f"created {_to_mib_int(total)}MiB")
size_before = upload_info.get("size_before")
size_after = upload_info.get("size_after")
if size_before is not None and size_after is not None and size_before != size_after:
extras.append("resized")
request.ctx._log_extra = " ".join(extras) if extras else None
watching.notify_change(rel, *rel.parents)
return json(
{
"status": "ack",
"req": {
"name": rel_name,
"size": total,
"start": start,
"end": end,
},
}
)
@bp.delete("/<name:path>")
async def delete_file(request, name):
rel, path = _safe_relpath(name)
if not rel.parts:
raise BadRequest("Refusing to delete root folder")
def _delete():
if not path.exists():
raise NotFound(f"File not found: {name}")
if path.is_dir():
shutil.rmtree(path)
else:
path.unlink()
await asyncio.to_thread(_delete)
watching.notify_change(rel, *rel.parents)
return empty(status=204)
@bp.route("/<name:path>", methods=["MKCOL"])
async def create_folder(request, name):
rel, path = _safe_relpath(name)
if not rel.parts:
raise BadRequest("Refusing to create root folder")
await asyncio.to_thread(path.mkdir, parents=True, exist_ok=False)
watching.notify_change(rel, *rel.parents)
return empty(status=201)
@bp.post("/", name="post_root", strict_slashes=False)
@bp.post("/<name:path>", name="post_path")
async def copy_or_move(request, name=""):
provided_args = set(request.args.keys())
if not provided_args:
raise BadRequest("No query arguments passed")
allowed_args = {"cp", "mv"}
unknown_args = sorted(provided_args - allowed_args)
if unknown_args:
raise BadRequest(f"Unknown query parameter(s): {', '.join(unknown_args)}")
mv_vals = request.args.getlist("mv")
cp_vals = request.args.getlist("cp")
mv_keys: list[str] = []
for value in mv_vals:
mv_keys.extend(k for k in value.split() if k)
cp_keys: list[str] = []
for value in cp_vals:
cp_keys.extend(k for k in value.split() if k)
if not mv_keys and not cp_keys:
raise BadRequest("No keys given")
dst_rel, dst_abs = _safe_relpath(name)
dst_exists = dst_abs.exists()
dst_is_dir = dst_exists and dst_abs.is_dir()
ordered_keys = cp_keys + mv_keys
key_paths = _get_key_paths(set(ordered_keys))
missing = [key for key in ordered_keys if key not in key_paths]
if missing:
raise NotFound("Files not found", context={"missing": missing})
# Validate target shape/type before mutating anything.
for op_name, op_keys in (("cp", cp_keys), ("mv", mv_keys)):
if len(op_keys) > 1 and not dst_is_dir:
raise BadRequest("Destination must be an existing directory for multiple keys")
if not op_keys:
continue
if not dst_is_dir:
if not dst_rel.parts:
raise BadRequest("Destination file path is required")
parent_abs = dst_abs.parent
if not parent_abs.is_dir():
raise BadRequest("Destination parent folder does not exist")
if dst_exists and dst_abs.is_file():
for key in op_keys:
src_abs = _resolve_from_relpath(key_paths[key])
if src_abs.is_dir():
raise BadRequest("Cannot move/copy a directory to an existing file")
changed: set[PurePosixPath] = set()
completed: list[dict[str, str]] = []
class _FileOpFailed(Exception):
def __init__(self, op_name: str, key: str, error: Exception):
self.op_name = op_name
self.key = key
self.error = error
super().__init__(str(error))
def _apply():
for op_name, op_keys in (("cp", cp_keys), ("mv", mv_keys)):
op_multi = len(op_keys) > 1
for key in op_keys:
try:
src_rel = key_paths[key]
src_abs = _resolve_from_relpath(src_rel)
if op_multi:
if not dst_is_dir:
raise BadRequest(
"Destination must be an existing directory for multiple keys"
)
dst_item_rel = (
dst_rel / src_rel.name if dst_rel.parts else PurePosixPath(src_rel.name)
)
elif dst_is_dir:
dst_item_rel = (
dst_rel / src_rel.name if dst_rel.parts else PurePosixPath(src_rel.name)
)
else:
if not dst_rel.parts:
raise BadRequest("Destination file path is required")
parent_abs = dst_abs.parent
if not parent_abs.is_dir():
raise BadRequest("Destination parent folder does not exist")
if src_abs.is_dir() and dst_exists and dst_abs.is_file():
raise BadRequest(
"Cannot move/copy a directory to an existing file"
)
dst_item_rel = dst_rel
dst_item_abs = _resolve_from_relpath(dst_item_rel)
if op_name == "mv":
# A no-op rename should still return success.
if src_abs != dst_item_abs:
shutil.move(src_abs, dst_item_abs)
changed.add(src_rel)
changed.add(src_rel.parent)
elif src_abs.is_dir():
shutil.copytree(
src_abs,
dst_item_abs,
dirs_exist_ok=True,
ignore_dangling_symlinks=True,
)
else:
shutil.copy2(src_abs, dst_item_abs)
changed.add(dst_item_rel)
changed.add(dst_item_rel.parent)
completed.append({"op": op_name, "key": key})
except Exception as e:
raise _FileOpFailed(op_name, key, e) from e
try:
await asyncio.to_thread(_apply)
except _FileOpFailed as e:
raise BadRequest(
"File operation failed after partial progress",
context={
"failed_op": e.op_name,
"failed_key": e.key,
"error": str(e.error),
"completed": completed,
},
) from e
notify_paths = [p for p in changed if p.parts]
if notify_paths:
watching.notify_change(*notify_paths)
return json(
{
"status": "ack",
"counts": {"cp": len(cp_keys), "mv": len(mv_keys)},
}
)
@bp.get("/<name:path>")
async def get_file(request, name=""):
return await _send_static_file(request, name, head_only=False)
@bp.head("/<name:path>")
async def head_file(request, name=""):
return await _send_static_file(request, name, head_only=True)
@bp.route("/", methods=["OPTIONS"], name="options_root", strict_slashes=False)
@bp.route("/<name:path>", methods=["OPTIONS"], name="options_path")
async def dav_options(request, name=""):
return HTTPResponse(
status=200,
headers={
"Allow": "OPTIONS, GET, HEAD, PUT, DELETE, MKCOL, COPY, MOVE, PROPFIND, POST",
"DAV": "1",
"MS-Author-Via": "DAV",
},
)
@bp.route("/", methods=["PROPFIND"], name="propfind_root", strict_slashes=False)
@bp.route("/<name:path>", methods=["PROPFIND"], name="propfind_path")
async def dav_propfind(request, name=""):
rel, path = _safe_relpath(name)
if not path.exists():
raise NotFound(f"Not found: {name}")
depth = request.headers.get("depth", "1").strip()
if depth == "infinity":
return HTTPResponse(status=403)
entries = await asyncio.to_thread(_collect_propfind_entries, rel, path, depth)
return HTTPResponse(
body=_build_propfind_xml(entries),
status=207,
content_type='application/xml; charset="utf-8"',
)
@bp.route("/", methods=["COPY"], name="copy_root", strict_slashes=False)
@bp.route("/<name:path>", methods=["COPY"], name="copy_path")
async def dav_copy(request, name=""):
dest_header = request.headers.get("destination")
if not dest_header:
raise BadRequest("Missing Destination header")
overwrite = request.headers.get("overwrite", "T").strip().upper() != "F"
src_rel, src_abs = _safe_relpath(name)
dst_rel, dst_abs = _parse_webdav_destination(dest_header)
request.ctx._log_extra = f"{dst_rel}"
if not src_abs.exists():
raise NotFound(f"Source not found: {name}")
if src_abs == dst_abs:
raise BadRequest("Source and destination are the same")
dst_existed = dst_abs.exists()
if dst_existed and not overwrite:
return HTTPResponse(status=412)
if not dst_abs.parent.is_dir():
return HTTPResponse(status=409)
def _do_copy():
if dst_existed:
shutil.rmtree(dst_abs) if dst_abs.is_dir() else dst_abs.unlink()
if src_abs.is_dir():
shutil.copytree(src_abs, dst_abs, ignore_dangling_symlinks=True)
else:
shutil.copy2(src_abs, dst_abs)
await asyncio.to_thread(_do_copy)
watching.notify_change(dst_rel, *dst_rel.parents)
return HTTPResponse(status=201 if not dst_existed else 204)
@bp.route("/", methods=["MOVE"], name="move_root", strict_slashes=False)
@bp.route("/<name:path>", methods=["MOVE"], name="move_path")
async def dav_move(request, name=""):
dest_header = request.headers.get("destination")
if not dest_header:
raise BadRequest("Missing Destination header")
overwrite = request.headers.get("overwrite", "T").strip().upper() != "F"
src_rel, src_abs = _safe_relpath(name)
dst_rel, dst_abs = _parse_webdav_destination(dest_header)
request.ctx._log_extra = f"{dst_rel}"
if not src_abs.exists():
raise NotFound(f"Source not found: {name}")
if src_abs == dst_abs:
return HTTPResponse(status=204)
dst_existed = dst_abs.exists()
if dst_existed and not overwrite:
return HTTPResponse(status=412)
if not dst_abs.parent.is_dir():
return HTTPResponse(status=409)
def _do_move():
if dst_existed:
shutil.rmtree(dst_abs) if dst_abs.is_dir() else dst_abs.unlink()
shutil.move(src_abs, dst_abs)
await asyncio.to_thread(_do_move)
watching.notify_change(src_rel, *src_rel.parents, dst_rel, *dst_rel.parents)
return HTTPResponse(status=201 if not dst_existed else 204)
def _parse_content_range(header: str, body_len: int) -> tuple[int, int, int]:
m = _CONTENT_RANGE_RE.fullmatch(header.strip())
if m is None:
raise BadRequest("Invalid Content-Range format")
start, end_inclusive, total = (int(v) for v in m.groups())
if total <= 0:
raise BadRequest("Invalid Content-Range total size")
if start > end_inclusive:
raise BadRequest("Invalid Content-Range range")
if end_inclusive >= total:
raise BadRequest("Content-Range exceeds total size")
expected_len = end_inclusive - start + 1
if expected_len != body_len:
raise BadRequest(
f"Content length mismatch for range: expected {expected_len}, got {body_len}"
)
return start, end_inclusive + 1, total
def _to_mib_int(value_bytes: int) -> int:
return round(value_bytes / (1 << 20))
def _safe_relpath(path: str) -> tuple[PurePosixPath, Path]:
"""Resolve a user path under storage root and enforce containment."""
base = config.config.path.resolve()
try:
sanitized = filename.sanitize(unquote(path))
except ValueError as e:
raise BadRequest(f"Invalid path: {e}") from e
resolved = (base / sanitized).resolve()
if not resolved.is_relative_to(base):
raise BadRequest("Invalid path")
rel = PurePosixPath(resolved.relative_to(base).as_posix())
return rel, resolved
def _resolve_from_relpath(rel: PurePosixPath) -> Path:
"""Resolve a relative path under storage root and enforce containment."""
base = config.config.path.resolve()
resolved = (base / rel).resolve()
if not resolved.is_relative_to(base):
raise BadRequest("Invalid path")
return resolved
async def _send_static_file(request, name: str, *, head_only: bool):
_, path = _safe_relpath(name)
st = await asyncio.to_thread(path.stat)
if path.is_dir():
raise NotFound(f"Not a file: {name}")
size = st.st_size
start = 0
end_excl = size
status = 200
range_header = request.headers.get("range")
if range_header is not None:
parsed = _parse_range_header(range_header, size)
if parsed is None:
return empty(
status=416,
headers={
"accept-ranges": "bytes",
"content-range": f"bytes */{size}",
},
)
start, end_excl = parsed
status = 206
length = end_excl - start
mime = mimetypes.guess_type(path.name)[0] or "application/octet-stream"
headers = {
"accept-ranges": "bytes",
"cache-control": "no-cache",
"content-length": str(length),
"content-type": mime,
"last-modified": format_date_time(st.st_mtime),
}
if status == 206:
headers["content-range"] = f"bytes {start}-{end_excl - 1}/{size}"
if head_only:
return empty(status=status, headers=headers)
res = await request.respond(status=status, headers=headers)
fd = await asyncio.to_thread(os.open, path, os.O_RDONLY)
try:
pos = start
while pos < end_excl:
chunk = await asyncio.to_thread(
os.pread,
fd,
min(_FILE_CHUNK_SIZE, end_excl - pos),
pos,
)
if not chunk:
break
pos += len(chunk)
await res.send(chunk)
finally:
await asyncio.to_thread(os.close, fd)
def _parse_range_header(header: str, size: int) -> tuple[int, int] | None:
value = header.strip()
if "," in value:
return None
m = _RANGE_RE.fullmatch(value)
if m is None:
return None
start_s, end_s = m.groups()
if not start_s and not end_s:
return None
if start_s:
start = int(start_s)
if start >= size:
return None
end_inclusive = int(end_s) if end_s else (size - 1)
if end_inclusive < start:
return None
end_inclusive = min(end_inclusive, size - 1)
return start, end_inclusive + 1
suffix_len = int(end_s)
if suffix_len <= 0:
return None
if suffix_len >= size:
return 0, size
start = size - suffix_len
return start, size
def _get_key_paths(wanted: set[str]) -> dict[str, PurePosixPath]:
"""Map file keys to their current relative filesystem paths."""
loc = PurePosixPath()
ret: dict[str, PurePosixPath] = {}
with watching.state.lock:
root = watching.state.root
for f in root:
loc = PurePosixPath(*loc.parts[: f.level - 1]) / f.name
if f.key in wanted and f.key not in ret:
ret[f.key] = loc
if len(ret) == len(wanted):
break
return ret
# ---------------------------------------------------------------------------
# WebDAV helpers
# ---------------------------------------------------------------------------
def _parse_webdav_destination(dest_header: str) -> tuple[PurePosixPath, Path]:
"""Parse a WebDAV Destination header and resolve it to a storage path."""
parsed = urlparse(dest_header)
raw_path = parsed.path # still percent-encoded
prefix = "/files"
if raw_path in (prefix, prefix + "/"):
rel_str = ""
elif raw_path.startswith(prefix + "/"):
rel_str = raw_path[len(prefix) + 1:]
else:
raise BadRequest("Destination must be within /files")
return _safe_relpath(rel_str)
def _rel_to_href(rel: PurePosixPath, is_dir: bool) -> str:
"""Build a DAV href from a storage-relative path."""
parts = rel.parts
if not parts:
return "/files/"
encoded = "/".join(url_quote(p, safe="") for p in parts)
href = f"/files/{encoded}"
return href + "/" if is_dir else href
def _dav_xml(element: ET.Element) -> bytes:
"""Serialise an ElementTree element to UTF-8 bytes with XML declaration."""
return (
b'<?xml version="1.0" encoding="UTF-8"?>'
+ ET.tostring(element, encoding="unicode").encode("utf-8")
)
def _collect_propfind_entries(rel: PurePosixPath, path: Path, depth: str) -> list[dict]:
entries = [_propfind_entry(rel, path)]
if depth == "1" and path.is_dir():
for child in sorted(path.iterdir(), key=lambda p: (p.is_file(), p.name)):
child_rel = rel / child.name if rel.parts else PurePosixPath(child.name)
try:
entries.append(_propfind_entry(child_rel, child))
except OSError:
pass
return entries
def _propfind_entry(rel: PurePosixPath, path: Path) -> dict:
st = path.stat()
is_dir = path.is_dir()
return {
"href": _rel_to_href(rel, is_dir),
"name": rel.parts[-1] if rel.parts else "",
"is_dir": is_dir,
"size": st.st_size,
"etag": f'"{st.st_mtime:.0f}-{st.st_size}"',
"content_type": mimetypes.guess_type(path.name)[0] or "application/octet-stream",
"last_modified": format_date_time(st.st_mtime),
"created": datetime.fromtimestamp(st.st_ctime, tz=timezone.utc).strftime(
"%Y-%m-%dT%H:%M:%SZ"
),
}
def _build_propfind_xml(entries: list[dict]) -> bytes:
multistatus = ET.Element(_dav_tag("multistatus"))
for e in entries:
response = ET.SubElement(multistatus, _dav_tag("response"))
ET.SubElement(response, _dav_tag("href")).text = e["href"]
propstat = ET.SubElement(response, _dav_tag("propstat"))
prop = ET.SubElement(propstat, _dav_tag("prop"))
rt = ET.SubElement(prop, _dav_tag("resourcetype"))
if e["is_dir"]:
ET.SubElement(rt, _dav_tag("collection"))
ET.SubElement(prop, _dav_tag("displayname")).text = e["name"]
ET.SubElement(prop, _dav_tag("getlastmodified")).text = e["last_modified"]
ET.SubElement(prop, _dav_tag("creationdate")).text = e["created"]
if not e["is_dir"]:
ET.SubElement(prop, _dav_tag("getcontentlength")).text = str(e["size"])
ET.SubElement(prop, _dav_tag("getcontenttype")).text = e["content_type"]
ET.SubElement(prop, _dav_tag("getetag")).text = e["etag"]
ET.SubElement(propstat, _dav_tag("status")).text = "HTTP/1.1 200 OK"
return _dav_xml(multistatus)
+2 -6
View File
@@ -221,9 +221,7 @@ class _PreviewWorkerPool:
logger.warning(
"Preview worker protocol failure for %s: %s", filepath.name, e
)
raise PreviewError(
f"worker protocol failure for {filepath.name}: {e}"
)
raise PreviewError(f"worker protocol failure for {filepath.name}: {e}")
finally:
if replace:
await self._replace_worker(worker)
@@ -470,9 +468,7 @@ def process_pdf(path, *, maxsize, maxzoom, quality, page_number=0):
img = pyvips.Image.new_from_memory(
pix.samples_mv, pix.width, pix.height, pix.n, "uchar"
)
ret = img.write_to_buffer(
".avif", Q=quality, effort=AVIF_FAST_EFFORT, strip=True
)
ret = img.write_to_buffer(".avif", Q=quality, effort=AVIF_FAST_EFFORT, strip=True)
backend = "pdf+pyvips"
t_save_end = perf_counter()
-115
View File
@@ -1,125 +1,10 @@
from __future__ import annotations
import shutil
from pathlib import PurePosixPath
from typing import Any
import msgspec
from sanic import BadRequest
from cista import config
from cista.util import filename
## Control commands
class ControlBase(msgspec.Struct, tag_field="op", tag=str.lower):
def __call__(self):
raise NotImplementedError
def affected_paths(self) -> list[str]:
"""Return list of paths affected by this operation for change notification."""
return []
class MkDir(ControlBase):
path: str
def __call__(self):
path = config.config.path / filename.sanitize(self.path)
path.mkdir(parents=True, exist_ok=False)
def affected_paths(self) -> list[str]:
return [filename.sanitize(self.path)]
class Rename(ControlBase):
path: str
to: str
def __call__(self):
to = filename.sanitize(self.to)
if "/" in to:
raise BadRequest("Rename 'to' name should only contain filename, not path")
path = config.config.path / filename.sanitize(self.path)
path.rename(path.with_name(to))
def affected_paths(self) -> list[str]:
sanitized = filename.sanitize(self.path)
new_path = str(PurePosixPath(sanitized).with_name(filename.sanitize(self.to)))
return [sanitized, new_path]
class Rm(ControlBase):
sel: list[str]
def __call__(self):
root = config.config.path
sel = [root / filename.sanitize(p) for p in self.sel]
for p in sel:
if p.is_dir():
shutil.rmtree(p)
else:
p.unlink()
def affected_paths(self) -> list[str]:
return [filename.sanitize(p) for p in self.sel]
class Mv(ControlBase):
sel: list[str]
dst: str
def __call__(self):
root = config.config.path
sel = [root / filename.sanitize(p) for p in self.sel]
dst = root / filename.sanitize(self.dst)
if not dst.is_dir():
raise BadRequest("The destination must be a directory")
for p in sel:
shutil.move(p, dst)
def affected_paths(self) -> list[str]:
dst = filename.sanitize(self.dst)
paths = [filename.sanitize(p) for p in self.sel]
# Include new locations in dst
paths.extend(f"{dst}/{PurePosixPath(p).name}" for p in self.sel)
return paths
class Cp(ControlBase):
sel: list[str]
dst: str
def __call__(self):
root = config.config.path
sel = [root / filename.sanitize(p) for p in self.sel]
dst = root / filename.sanitize(self.dst)
if not dst.is_dir():
raise BadRequest("The destination must be a directory")
for p in sel:
if p.is_dir():
# Note: copies as dst rather than in dst unless name is appended.
shutil.copytree(
p,
dst / p.name,
dirs_exist_ok=True,
ignore_dangling_symlinks=True,
)
else:
shutil.copy2(p, dst)
def affected_paths(self) -> list[str]:
dst = filename.sanitize(self.dst)
# Only destinations are new (sources unchanged)
return [f"{dst}/{PurePosixPath(filename.sanitize(p)).name}" for p in self.sel]
ControlTypes = MkDir | Rename | Rm | Mv | Cp
class StatusMsg(msgspec.Struct):
status: str
req: Any
class ErrorMsg(msgspec.Struct):
+23 -15
View File
@@ -8,18 +8,18 @@ from ipaddress import IPv6Address
logger = logging.getLogger("cista.access")
_RESET = "\033[0m"
_STATUS_INFO = "\033[32m" # 1xx (green)
_STATUS_OK = "\033[1;92m" # 2xx (bright green)
_STATUS_INFO = "\033[32m" # 1xx (green)
_STATUS_OK = "\033[1;92m" # 2xx (bright green)
_STATUS_REDIRECT = "\033[32m" # 3xx (green)
_STATUS_CLIENT_ERR = "\033[0;31m" # 4xx (red)
_STATUS_SERVER_ERR = "\033[1;91m" # 5xx (bold bright red)
_METHOD_READ = "\033[0;34m" # GET, HEAD, OPTIONS (blue)
_METHOD_WRITE = "\033[1;94m" # POST, PUT, DELETE, PATCH (bold bright blue)
_HOST = "\033[38;5;242m" # hostname (dark grey)
_PATH = "\033[38;5;250m" # path (light grey)
_TIMING = "\033[38;5;242m" # timing (dark grey)
_WS_OPEN = "\033[1;93m" # WebSocket connect (bold bright yellow)
_WS_CLOSE = "\033[33m" # WebSocket disconnect (yellow)
_STATUS_CLIENT_ERR = "\033[0;31m" # 4xx (red)
_STATUS_SERVER_ERR = "\033[1;91m" # 5xx (bold bright red)
_METHOD_READ = "\033[0;34m" # GET, HEAD, OPTIONS (blue)
_METHOD_WRITE = "\033[1;94m" # POST, PUT, DELETE, PATCH (bold bright blue)
_HOST = "\033[38;5;242m" # hostname (dark grey)
_PATH = "\033[38;5;250m" # path (light grey)
_TIMING = "\033[38;5;242m" # timing (dark grey)
_WS_OPEN = "\033[1;93m" # WebSocket connect (bold bright yellow)
_WS_CLOSE = "\033[33m" # WebSocket disconnect (yellow)
_WS_STATUS = "\033[38;5;250m" # WebSocket close status (normal white)
@@ -113,7 +113,12 @@ def _format_method_label(label: str, *, color: str | None = None) -> str:
def format_access_log(
client: str, status: int, method: str, host: str, path: str, duration_ms: float,
client: str,
status: int,
method: str,
host: str,
path: str,
duration_ms: float,
extra: str | None = None,
) -> str:
ip = _format_left(format_client_ip(client))
@@ -123,7 +128,9 @@ def format_access_log(
path_str = f"{_PATH}{path}{_RESET}"
timing_str = f"{_TIMING}{format_duration_ms(duration_ms)}{_RESET}"
extra_str = f" {_TIMING}{extra}{_RESET}" if extra else ""
return f"{ip} {status_str} {method_str} {host_str}{path_str}{extra_str} {timing_str}"
return (
f"{ip} {status_str} {method_str} {host_str}{path_str}{extra_str} {timing_str}"
)
_ws_counter = 1
@@ -194,7 +201,7 @@ WS_CLOSE_CODES = {
}
def log_ws_close(ws_id: int, close_code: int | None, duration: float) -> None:
def log_ws_close(ws_id: int, close_code: int | None, duration: float, extra: str | None = None) -> None:
"""Log WebSocket connection close with duration and status."""
id_str = _format_ws_id(ws_id)
timing = format_duration_ms(duration * 1000)
@@ -209,8 +216,9 @@ def log_ws_close(ws_id: int, close_code: int | None, duration: float) -> None:
method_str = _format_method_label("closed", color=_TIMING)
status_str = f"{_WS_STATUS}{code} {status}{_RESET}"
timing_str = f"{_TIMING}{timing}{_RESET}"
extra_str = f" {_TIMING}{extra}{_RESET}" if extra else ""
logger.info("%s %s %s %s %s", " " * 19, id_str, method_str, status_str, timing_str)
logger.info("%s %s %s %s %s%s", " " * 19, id_str, method_str, status_str, timing_str, extra_str)
def configure_access_logging() -> None:
+4 -4
View File
@@ -19,21 +19,21 @@ def get(request):
return False if "s" in request.cookies else None
def create(res, username, **kwargs):
def create(res, username, *, secure: bool = True, **kwargs):
data = {
"exp": int(time()) + max_age,
"username": username,
**kwargs,
}
s = jwt.encode(data, session_secret())
res.cookies.add_cookie("s", s, httponly=True, max_age=max_age)
res.cookies.add_cookie("s", s, httponly=True, max_age=max_age, secure=secure)
def update(res, s, **kwargs):
def update(res, s, *, secure: bool = True, **kwargs):
s.update(kwargs)
s = jwt.encode(s, session_secret())
max_age = max(1, s["exp"] - int(time())) # type: ignore
res.cookies.add_cookie("s", s, httponly=True, max_age=max_age)
res.cookies.add_cookie("s", s, httponly=True, max_age=max_age, secure=secure)
def delete(res):
+55
View File
@@ -152,6 +152,61 @@ async def validate_sso_request(request, *, perm: str = "cista:login") -> dict |
)
async def check_permissions(user_id: str, perm: str) -> dict:
"""Check if a Paskia user has the given permission.
Args:
user_id: The Paskia user UUID
perm: Permission to check (e.g. cista:login or cista:admin)
Returns:
User info dict if permission is granted
Raises:
Forbidden: If permission is denied or check fails
SanicException: If the auth service is unreachable
"""
if not paskia_enabled():
raise ValueError("Paskia not enabled")
client = await get_client()
url = f"{PASKIA_BACKEND_URL}/auth/api/check-permissions"
try:
response = await client.post(
url,
json={"user_id": user_id, "perm": perm},
headers={"accept": "application/json"},
)
if response.status_code == 200:
return response.json()
try:
error_data = response.json()
except Exception:
error_data = {"detail": response.text or "Permission check failed"}
if response.status_code == 403:
raise Forbidden(
error_data.get("detail", "Access denied"),
quiet=True,
)
else:
raise Forbidden(
error_data.get("detail", "Permission check failed"),
quiet=True,
)
except httpx.RequestError as e:
logger.error(f"Permission check {url} network error: {e}")
raise SanicException(
"Authentication service unavailable",
status_code=502,
quiet=True,
)
async def proxy_auth_request(request):
"""Proxy a request to the auth backend.
+6 -1
View File
@@ -24,10 +24,12 @@ def jres(data, **kwargs):
async def handle_sanic_exception(request, e):
context, code = {}, 500
headers = None
message = str(e)
if isinstance(e, SanicException):
context = e.context or {}
code = e.status_code
headers = getattr(e, "headers", None)
if not message or not request.app.debug and code == 500:
message = "Internal Server Error"
message = f"⚠️ {message}" if code < 500 else f"🛑 {message}"
@@ -41,6 +43,7 @@ async def handle_sanic_exception(request, e):
return jres(
response_data,
status=code,
headers=headers,
)
# Redirections flash the error message via cookies
if "redirect" in context:
@@ -60,6 +63,7 @@ def websocket_wrapper(handler):
extra = username if username else None
start = time.perf_counter()
ws_id = log_ws_open(request, extra=extra)
close_extra = None
try:
await auth.verify(request)
await handler(request, ws, *args, **kwargs)
@@ -72,6 +76,7 @@ def websocket_wrapper(handler):
await asend(ws, ErrorMsg({"code": code, "message": message, **context}))
if not getattr(e, "quiet", False) or code == 500:
logger.exception(f"{code} {e!r}")
close_extra = f"{code} {message}"
raise
finally:
duration = time.perf_counter() - start
@@ -86,6 +91,6 @@ def websocket_wrapper(handler):
close_code = p.close_code
except AttributeError:
pass
log_ws_close(ws_id, close_code, duration)
log_ws_close(ws_id, close_code, duration, extra=close_extra)
return wrapper