fix(webhooks): wire receiver to vector sync queue and fix registered URI
The /webhooks/nextcloud endpoint was a no-op stub that logged the payload and returned 200 OK; webhook deletions never reached Qdrant. Compounding that, _get_webhook_uri() registered the docker-compose internal hostname (http://mcp:8000) with Nextcloud whenever /.dockerenv existed — including ECS Fargate — so cloud deployments were registering a URL NC could not resolve. - New vector/webhook_parser.py extracts a DocumentTask from NodeCreatedEvent / NodeWrittenEvent / BeforeNodeDeletedEvent payloads scoped to */files/Notes/*.md (matching the registered preset filters). - New vector/webhook_receiver.py pushes that task onto the same send-stream the scanner uses (app.state.document_send_stream), with 503 when sync is not running so NC retries delivery. - _get_webhook_uri() now prefers NEXTCLOUD_MCP_SERVER_URL over the /.dockerenv branch, so the explicit public URL set on cloud tasks wins; docker-compose dev still falls back to the internal name when no public URL is configured. Calendar / Tables event parsing is intentionally out of scope here. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Chris Coutinho
co-authored by
Claude Opus 4.7
Claude Opus 4.7
parent
111ae988f9
commit
2e2a098bee
@@ -0,0 +1,86 @@
|
||||
"""Parse Nextcloud webhook payloads into DocumentTask objects.
|
||||
|
||||
Maps Nextcloud webhook events to vector-sync DocumentTasks. The handler at
|
||||
``/webhooks/nextcloud`` calls :func:`extract_document_task` and forwards any
|
||||
non-None result to the same processor send-stream the scanner uses.
|
||||
|
||||
Currently scoped to file (note) events. Calendar / Tables events fall through
|
||||
to ``None`` for now; those parsers can be added in follow-up changes.
|
||||
|
||||
See ADR-010 for the design and ``webhook-testing-findings.md`` for real
|
||||
captured payloads.
|
||||
"""
|
||||
|
||||
import logging
|
||||
import re
|
||||
|
||||
from nextcloud_mcp_server.vector.scanner import DocumentTask
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
_FILE_EVENT_CREATED = "OCP\\Files\\Events\\Node\\NodeCreatedEvent"
|
||||
_FILE_EVENT_WRITTEN = "OCP\\Files\\Events\\Node\\NodeWrittenEvent"
|
||||
_FILE_EVENT_BEFORE_DELETED = "OCP\\Files\\Events\\Node\\BeforeNodeDeletedEvent"
|
||||
|
||||
# Matches paths inside any user's Notes folder ending in .md, e.g.
|
||||
# "/admin/files/Notes/Sub/Note.md" or "/alice/files/Notes/foo.md".
|
||||
_NOTES_PATH_RE = re.compile(r"^/[^/]+/files/Notes/.+\.md$")
|
||||
|
||||
|
||||
def extract_document_task(payload: dict) -> DocumentTask | None:
|
||||
"""Convert a Nextcloud webhook payload into a DocumentTask.
|
||||
|
||||
Returns None for any event we don't (yet) handle, or any event whose
|
||||
target isn't a markdown file under a user's Notes folder. Callers should
|
||||
treat None as "ignored" — not an error.
|
||||
"""
|
||||
try:
|
||||
event = payload["event"]
|
||||
event_class = event["class"]
|
||||
user_id = payload["user"]["uid"]
|
||||
except (KeyError, TypeError):
|
||||
logger.debug("Webhook payload missing user/event/class fields")
|
||||
return None
|
||||
|
||||
if event_class in (
|
||||
_FILE_EVENT_CREATED,
|
||||
_FILE_EVENT_WRITTEN,
|
||||
_FILE_EVENT_BEFORE_DELETED,
|
||||
):
|
||||
return _parse_file_event(event_class, event, user_id, payload.get("time", 0))
|
||||
|
||||
logger.debug("Ignoring webhook for unsupported event: %s", event_class)
|
||||
return None
|
||||
|
||||
|
||||
def _parse_file_event(
|
||||
event_class: str, event: dict, user_id: str, time: int
|
||||
) -> DocumentTask | None:
|
||||
node = event.get("node") or {}
|
||||
path = node.get("path", "")
|
||||
node_id = node.get("id")
|
||||
|
||||
if not _NOTES_PATH_RE.match(path):
|
||||
# Not a note file — could be a parent folder, an unrelated file, etc.
|
||||
return None
|
||||
|
||||
if node_id is None:
|
||||
# BeforeNodeDeletedEvent should still carry node.id; if it doesn't
|
||||
# we can't address the Qdrant points to delete. Skip rather than
|
||||
# guess — the polling scanner will catch up via its grace period.
|
||||
logger.warning(
|
||||
"Webhook %s for note %s missing node.id; falling back to scanner",
|
||||
event_class,
|
||||
path,
|
||||
)
|
||||
return None
|
||||
|
||||
operation = "delete" if event_class == _FILE_EVENT_BEFORE_DELETED else "index"
|
||||
|
||||
return DocumentTask(
|
||||
user_id=user_id,
|
||||
doc_id=str(node_id),
|
||||
doc_type="note",
|
||||
operation=operation,
|
||||
modified_at=int(time),
|
||||
)
|
||||
@@ -0,0 +1,82 @@
|
||||
"""HTTP receiver for Nextcloud webhooks.
|
||||
|
||||
Routes inbound webhooks to the same processor send-stream the scanner uses.
|
||||
The receiver is registered as a Starlette route at ``/webhooks/nextcloud``
|
||||
in :mod:`nextcloud_mcp_server.app`.
|
||||
"""
|
||||
|
||||
import logging
|
||||
|
||||
from starlette.requests import Request
|
||||
from starlette.responses import JSONResponse
|
||||
|
||||
from nextcloud_mcp_server.vector.webhook_parser import extract_document_task
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
async def handle_nextcloud_webhook(request: Request) -> JSONResponse:
|
||||
"""Receive a Nextcloud webhook and queue a DocumentTask for vector sync.
|
||||
|
||||
Returns quickly so NC's webhook worker is not blocked. The send-stream is
|
||||
read from ``request.app.state.document_send_stream``; when vector sync
|
||||
isn't running we return 503 so NC retries delivery.
|
||||
"""
|
||||
try:
|
||||
payload = await request.json()
|
||||
except Exception as e:
|
||||
logger.warning(f"Webhook payload was not valid JSON: {e}")
|
||||
return JSONResponse(
|
||||
{"status": "error", "message": "invalid JSON"},
|
||||
status_code=400,
|
||||
)
|
||||
|
||||
task = extract_document_task(payload)
|
||||
if task is None:
|
||||
event_class = (payload.get("event") or {}).get("class", "<missing>")
|
||||
logger.debug("Webhook ignored (unsupported event): %s", event_class)
|
||||
return JSONResponse(
|
||||
{"status": "ignored", "reason": "unsupported event"},
|
||||
status_code=200,
|
||||
)
|
||||
|
||||
send_stream = getattr(request.app.state, "document_send_stream", None)
|
||||
if send_stream is None:
|
||||
logger.warning(
|
||||
"Webhook received but vector sync is not running; rejecting so NC retries"
|
||||
)
|
||||
return JSONResponse(
|
||||
{"status": "unavailable", "reason": "vector sync not running"},
|
||||
status_code=503,
|
||||
)
|
||||
|
||||
try:
|
||||
await send_stream.send(task)
|
||||
except Exception as e:
|
||||
logger.error(
|
||||
"Failed to queue webhook task for %s_%s: %s",
|
||||
task.doc_type,
|
||||
task.doc_id,
|
||||
e,
|
||||
)
|
||||
return JSONResponse(
|
||||
{"status": "error", "message": "queue unavailable"},
|
||||
status_code=500,
|
||||
)
|
||||
|
||||
logger.info(
|
||||
"Webhook queued %s_%s (%s) for user %s",
|
||||
task.doc_type,
|
||||
task.doc_id,
|
||||
task.operation,
|
||||
task.user_id,
|
||||
)
|
||||
return JSONResponse(
|
||||
{
|
||||
"status": "queued",
|
||||
"doc_type": task.doc_type,
|
||||
"doc_id": task.doc_id,
|
||||
"operation": task.operation,
|
||||
},
|
||||
status_code=200,
|
||||
)
|
||||
Reference in New Issue
Block a user