"""aiohttp server for vacuum-walld.
Listens on a Unix socket, serves the daemon API to the web UI.
Handles routing, batching, and request/response lifecycle.
"""
import asyncio
import hashlib
import json
import logging
import os
import signal
from collections.abc import Callable
from pathlib import Path
from typing import Any
from aiohttp import web
from daemon.iface import PathLike
from lib.state import _DEFAULT_POLL_INTERVALS
from lib.state import state as state_store
logger = logging.getLogger(__name__)
PROJECT_DIR = Path(__file__).resolve().parent.parent
SOCKET_PATH = PROJECT_DIR / "data" / "daemon.sock"
_WS_PORT = int(os.environ.get("VACUUM_WALLD_WS_PORT", "9091"))
# Polling intervals per subsystem (seconds). VACUUM_WALL_POLL_INTERVALS overrides.
_RAW_POLL = os.environ.get("VACUUM_WALL_POLL_INTERVALS", "")
if _RAW_POLL:
_POLL_OVERRIDE: dict[str, int] = {}
for pair in _RAW_POLL.split(","):
if ":" in pair:
name, _, val = pair.partition(":")
try:
_POLL_OVERRIDE[name.strip()] = int(val.strip())
except ValueError:
logger.warning(
"Invalid poll interval value %r for %r, skipping", val, name
)
_POLL_INTERVALS = {**_DEFAULT_POLL_INTERVALS, **_POLL_OVERRIDE}
else:
_POLL_INTERVALS = dict(_DEFAULT_POLL_INTERVALS)
class Handler:
"""Wrapper for a daemon handler function.
Attributes:
method: HTTP method (e.g. "GET", "POST").
path: URL path pattern.
"""
def __init__(self, method: str, path: str) -> None:
"""Initialize the handler metadata.
Args:
method: HTTP method.
path: URL path pattern.
"""
self.method = method.upper()
self.path = path
class Registry:
"""Route registry for daemon handlers.
Attributes:
_routes: Mapping of (method, path) tuples to handler callables.
"""
def __init__(self) -> None:
"""Initialize an empty route registry."""
self._routes: dict[tuple[str, str], Callable] = {}
def register(self, method: PathLike, path: str | None = None):
"""Decorator that registers a handler for the given method and path.
Accepts either two separate arguments (``method``, ``path``) or a
single :class:`daemon.iface.Endpoint` tuple.
Args:
method: HTTP method string, or an :class:`Endpoint` tuple.
path: URL path (omit when passing an :class:`Endpoint`).
Returns:
Decorator function wrapping the handler.
"""
if isinstance(method, tuple):
ep_method, ep_path = method
path = ep_path
method = ep_method
def decorator(fn: Callable) -> Callable:
self._routes[(method.upper(), path)] = fn # type: ignore[arg-type]
fn._handler = Handler(method, path) # type: ignore[attr-defined,reportArgumentType]
return fn
return decorator
def get(self, method: str, path: str) -> Callable | None:
"""Look up a registered handler by method and path.
Args:
method: HTTP method to match.
path: URL path to match.
Returns:
The handler callable, or None if not registered.
"""
return self._routes.get((method.upper(), path))
def match(self, method: str, path: str):
"""Match *path* against registered patterns, returning handler + params.
Patterns may contain ```` segments (e.g. ``/foo/``).
Matching segments are captured into a dict and merged into *body*.
Returns:
Tuple of (handler_fn, params_dict) or (None, None) if no match.
"""
for (reg_method, reg_path), fn in self._routes.items():
if reg_method != method.upper():
continue
pat_params = _match_path(reg_path, path)
if pat_params is not None:
return fn, pat_params
return None, None
registry = Registry()
def refresh_state(subsystems: list[str] | None = None) -> None:
"""Refresh the pre-computed state for the given subsystems (or all).
Args:
subsystems: List of subsystem names to refresh. If None, all subsystems are refreshed.
"""
state_store.populate(subsystems)
targets = subsystems or state_store.SUBSYSTEMS
for name in targets:
state_store.bump(name)
try:
asyncio.get_running_loop()
except RuntimeError:
pass
else:
task = asyncio.create_task(broadcast_versions())
task.add_done_callback(_ws_tasks.discard)
_ws_tasks.add(task)
class NotFoundError(Exception):
"""Raised when a requested resource is not found."""
pass
def ok(data: Any = None) -> web.Response:
"""Create a success JSON response.
Args:
data: Response payload. Defaults to None.
Returns:
A JSON Response with `ok` set to True.
"""
return web.json_response({"ok": True, "data": data})
def error(msg: str, code: int = 400) -> web.Response:
"""Create an error JSON response.
Args:
msg: Error message.
code: HTTP status code. Defaults to 400.
Returns:
A JSON Response with `ok` set to False.
"""
return web.json_response({"ok": False, "error": msg}, status=code)
def _match_path(pattern: str, path: str) -> dict[str, str] | None:
"""Match *path* against a URL pattern containing ```` segments.
Args:
pattern: URL pattern like ``/network/interfaces/``.
path: Actual request path like ``/network/interfaces/eth1``.
Returns:
Dict mapping param names to their matched values, or ``None`` if no match.
"""
p_parts = pattern.strip("/").split("/")
r_parts = path.strip("/").split("/")
if len(p_parts) != len(r_parts):
return None
params: dict[str, str] = {}
for p_seg, r_seg in zip(p_parts, r_parts, strict=True):
if p_seg.startswith("<") and p_seg.endswith(">"):
params[p_seg[1:-1]] = r_seg
elif p_seg != r_seg:
return None
return params
async def _handle_request(request: web.Request) -> web.Response:
"""Dispatch a request to the appropriate handler.
Args:
request: The incoming HTTP request.
Returns:
The handler's response.
"""
handler_fn, pat_params = registry.match(request.method, request.path)
if handler_fn is None:
return error(f"Method {request.method} not allowed for {request.path}", 404)
# Build body — merge order (highest wins): path params > JSON body > query params.
# Path params come from the URL path (e.g. /interfaces/eth0) and should not
# be overridable by body or query parameters.
body: dict[str, Any] | None = pat_params if pat_params else None
if request.content_type == "application/json":
try:
json_body = await request.json()
body = {**json_body, **body} if body is not None else json_body
except json.JSONDecodeError:
return error("Invalid JSON body", 400)
query_dict = dict(request.query)
if query_dict:
query_body = {k: v[0] if len(v) == 1 else v for k, v in query_dict.items()}
body = {**query_body, **body} if body is not None else query_body
try:
if body is not None:
result = handler_fn(request, body)
if asyncio.iscoroutine(result):
result = await result
else:
result = handler_fn(request, None)
if asyncio.iscoroutine(result):
result = await result
except NotFoundError as exc:
return error(str(exc), 404)
except ValueError as exc:
return error(str(exc), 400)
except RuntimeError as exc:
logger.error("Handler error: %s", exc)
return error(str(exc), 500)
except Exception as exc:
logger.exception(
"Unexpected handler error in %s %s", request.method, request.path
)
return error(f"Internal error: {exc}", 500)
# Convert result to response if not already
if isinstance(result, web.Response):
return result
if isinstance(result, dict) and result.get("ok") is False:
return error(result["error"], result.get("code", 400))
return ok(result)
async def _handle_batch(request: web.Request) -> web.Response:
"""Handle batch requests: execute operations in order, return keyed results.
Args:
request: The incoming HTTP request containing operations.
Returns:
A JSON response keyed by operation IDs.
"""
try:
body = await request.json()
except json.JSONDecodeError:
return error("Invalid JSON body", 400)
ops = body.get("ops", [])
if not isinstance(ops, list):
return error("'ops' must be a list", 400)
results: dict[str, Any] = {}
for op in ops:
op_id = op.get("id")
method = op.get("method", "GET").upper()
path = op.get("path", "")
if not op_id or not path:
results[op_id] = {"ok": False, "error": "'id' and 'path' are required"}
continue
handler_fn, pat_params = registry.match(method, path)
if handler_fn is None:
results[op_id] = {
"ok": False,
"error": f"Endpoint not found: {method} {path}",
}
continue
op_body = op.get("body")
if pat_params:
op_body = {**(op_body or {}), **pat_params}
try:
result = handler_fn(None, op_body)
if asyncio.iscoroutine(result):
result = await result
except Exception as exc:
logger.error("Batch handler error for %s: %s", op_id, exc)
results[op_id] = {"ok": False, "error": str(exc)}
continue
# Strip ok/data wrapper for batch results
if isinstance(result, dict) and result.get("ok") is not None:
results[op_id] = result
else:
results[op_id] = {"ok": True, "data": result}
return ok(results)
def create_app() -> web.Application:
"""Create and configure the aiohttp application.
Returns:
A configured web.Application instance.
"""
app = web.Application()
app.router.add_route("GET", "/health", _health)
app.router.add_route("GET", "/status/all", get_status_all)
app.router.add_route("POST", "/status/refresh", refresh_status)
app.router.add_route("POST", "/batch", _handle_batch)
app.router.add_route("GET", "/ws", _handle_ws)
app.router.add_route("*", "/{tail:.*}", _catch_all)
return app
# WebSocket subscribers
_ws_subscribers: set[web.WebSocketResponse] = set()
_ws_tasks: set[asyncio.Task[None]] = set()
_poll_tasks: set[asyncio.Task[None]] = set()
async def _handle_ws(request: web.Request) -> web.Response:
"""WebSocket endpoint for real-time state change notifications.
On connect: sends current versions. On state change: broadcasts
updated subsystem versions. Clients disconnect to unsubscribe.
"""
ws = web.WebSocketResponse()
await ws.prepare(request)
_ws_subscribers.add(ws)
await ws.send_json({"type": "init", "versions": state_store.get_versions()})
try:
async for msg in ws:
if msg.type == web.WSMsgType.ERROR:
break
if msg.type == web.WSMsgType.CLOSE:
break
finally:
_ws_subscribers.discard(ws)
return ws
async def broadcast_versions() -> None:
"""Broadcast updated subsystem versions to all WebSocket clients."""
updated = state_store.get_updated_versions()
if not updated or not _ws_subscribers:
return
data = json.dumps({"type": "versions", "updated": updated})
dead: set[web.WebSocketResponse] = set()
for ws in _ws_subscribers:
try:
await ws.send_str(data)
except Exception:
dead.add(ws)
_ws_subscribers.difference_update(dead)
if dead:
logger.warning("Removed %d dead WS subscribers", len(dead))
async def broadcast_tick(subsystems: list[str]) -> None:
"""Broadcast a lightweight tick to WS clients without version payload."""
data = json.dumps({"type": "tick", "subsystems": subsystems})
dead: set[web.WebSocketResponse] = set()
for ws in _ws_subscribers:
try:
await ws.send_str(data)
except Exception:
dead.add(ws)
_ws_subscribers.difference_update(dead)
if dead:
logger.warning("Removed %d dead WS subscribers", len(dead))
async def _poll_loop(subsystem: str, interval: int) -> None:
"""Periodically poll a subsystem for state changes and broadcast as needed."""
offset = int(hashlib.md5(subsystem.encode()).hexdigest(), 16) % interval
await asyncio.sleep(offset)
while True:
try:
structural, volatile = state_store.poll(subsystem)
if structural:
state_store.bump(subsystem)
await broadcast_versions()
elif volatile:
await broadcast_tick([subsystem])
except asyncio.CancelledError:
raise
except Exception:
logger.error("Poll loop error for %s", subsystem, exc_info=True)
await asyncio.sleep(interval)
def start_polling() -> None:
"""Start one poll loop task per subsystem."""
for subsystem, interval in _POLL_INTERVALS.items():
task = asyncio.create_task(_poll_loop(subsystem, interval))
task.add_done_callback(_poll_tasks.discard)
_poll_tasks.add(task)
def _stop_polling() -> None:
"""Cancel all polling tasks."""
for task in _poll_tasks:
task.cancel()
_poll_tasks.clear()
async def _health(_request: web.Request) -> web.Response:
"""Return the health check response.
Returns:
JSON response with process ID and socket path.
"""
return ok({"pid": os.getpid(), "socket": str(SOCKET_PATH)})
async def get_status_all(_request: web.Request) -> web.Response:
"""Return the entire state snapshot in one call.
Returns:
JSON response containing state data for all subsystems.
"""
return ok({name: state_store.get(name) for name in state_store.SUBSYSTEMS})
async def refresh_status(_request: web.Request) -> web.Response:
"""Re-collect all state from system.
Args:
_request: The incoming request. JSON body may contain a "subsystems" list.
Returns:
JSON response with the updated state snapshot.
"""
try:
body = await _request.json()
except (json.JSONDecodeError, ValueError):
body = None
subsystems = None
if body and "subsystems" in body:
subsystems = body["subsystems"]
state_store.populate(subsystems)
return ok({name: state_store.get(name) for name in state_store.SUBSYSTEMS})
async def _catch_all(request: web.Request) -> web.Response:
"""Catch-all for registered routes.
Args:
request: The incoming HTTP request.
Returns:
The dispatched handler's response.
"""
return await _handle_request(request)
def _register_routes() -> None:
"""Import all handler modules to register routes.
Side effect: each imported module's `@registry.register` calls populate the
route registry with their handler endpoints.
"""
from daemon.handlers import (
acme, # noqa: F401
dnsmasq, # noqa: F401
firewall, # noqa: F401
logs, # noqa: F401
network, # noqa: F401
nginx, # noqa: F401
wireguard, # noqa: F401
)
def main() -> None:
"""Entry point for vacuum-walld.
Sets up logging, registers routes, creates the aiohttp application,
and starts listening on the Unix socket.
"""
from lib.logging import setup_logging
setup_logging()
_register_routes()
app = create_app()
socket_path = os.environ.get("VACUUM_WALLD_SOCKET", str(SOCKET_PATH))
socket_dir = Path(socket_path).parent
socket_dir.mkdir(parents=True, exist_ok=True)
if Path(socket_path).exists():
os.unlink(socket_path)
loop = asyncio.new_event_loop()
def _on_shutdown(_sig: int) -> None:
logger.info("Shutting down daemon...")
_stop_polling()
loop.stop()
for sig in (signal.SIGTERM, signal.SIGINT):
loop.add_signal_handler(sig, _on_shutdown, sig)
runner = web.AppRunner(app)
loop.run_until_complete(runner.setup())
site = web.UnixSite(runner, socket_path)
loop.run_until_complete(site.start())
tcp_site = web.TCPSite(runner, "127.0.0.1", _WS_PORT)
loop.run_until_complete(tcp_site.start())
os.chmod(socket_path, 0o660)
# Populate state from system (blocking — OK at startup)
logger.info("Populating system state...")
state_store.populate()
for subsystem in state_store.SUBSYSTEMS:
if state_store.get(subsystem) is not None:
state_store.bump(subsystem)
loop.run_until_complete(start_polling())
logger.info("vacuum-walld listening on %s", socket_path)
logger.info("WebSocket on 127.0.0.1:%d", _WS_PORT)
try:
loop.run_forever()
finally:
loop.run_until_complete(runner.cleanup())
if Path(socket_path).exists():
os.unlink(socket_path)
logger.info("vacuum-walld stopped")
if __name__ == "__main__":
main()