jiang
be6317eb48
Add project-bound OAuth MCP queries and controlled analysis with scoped evidence, limits, timeouts and a nearby-model spatial query using EPSG:4326 inputs. Default nearby searches to 200 meters and five results. Integrate investigation messaging/presentation, health/readiness and typed history; remove duplicate web search and fix online simulation and compressed baseline writes. Include project migrations 023-025; keep local credentials and demo seeding excluded. Validation: 319 focused pytest checks; real PostGIS temporary-table regression; read-only Jinqian bridge radius comparison; cross-language MCP discovery/calls.
448 lines
23 KiB
Python
448 lines
23 KiB
Python
#!/usr/bin/env python3
|
||
"""Synchronize the explicitly synthetic peripheral work orders into task projections."""
|
||
from __future__ import annotations
|
||
|
||
import argparse
|
||
import asyncio
|
||
import json
|
||
import logging
|
||
import os
|
||
import sys
|
||
from pathlib import Path
|
||
from uuid import NAMESPACE_URL, UUID, uuid5
|
||
|
||
import httpx
|
||
|
||
sys.path.insert(0, str(Path(__file__).resolve().parents[1]))
|
||
|
||
from samples.peripheral.seed_identity import SEED_ACTOR_ID
|
||
from app.domain.schemas.workbench_events import WorkbenchEvent
|
||
|
||
try:
|
||
from scripts.run_operational_worker import CredentialProvider, WorkerSettings
|
||
except ModuleNotFoundError:
|
||
from run_operational_worker import CredentialProvider, WorkerSettings
|
||
|
||
|
||
logger = logging.getLogger("peripheral-task-sync")
|
||
STATUS = {"draft": "scheduled", "open": "assigned", "in_progress": "in_progress",
|
||
"awaiting_reply": "awaiting_feedback", "completed": "completed", "cancelled": "cancelled"}
|
||
EVENT_WORK_ORDER_TYPES = frozenset({"water_quality_complaint", "burst_report"})
|
||
MAX_TARGETED_ORDERS = 200
|
||
|
||
|
||
def peripheral_order_id(external_id: str | None) -> str | None:
|
||
"""Return only canonical IDs owned by this synchronizer's namespace."""
|
||
if not external_id or not external_id.startswith("peripheral_sample:"):
|
||
return None
|
||
try:
|
||
return str(UUID(external_id.removeprefix("peripheral_sample:")))
|
||
except ValueError:
|
||
return None
|
||
|
||
|
||
def should_project_work_order(order: dict, linked_order_ids: set[str]) -> bool:
|
||
"""Exclude retired seed checks while preserving user and workflow orders."""
|
||
work_order_id = order["work_order_id"]
|
||
work_order_type = order.get("work_order_type", "general")
|
||
if work_order_type in EVENT_WORK_ORDER_TYPES or work_order_id in linked_order_ids:
|
||
return True
|
||
return not (
|
||
work_order_type == "general"
|
||
and order.get("origin") == "synthetic"
|
||
and order.get("created_by") == str(SEED_ACTOR_ID)
|
||
)
|
||
|
||
|
||
def work_order_snapshot(order: dict) -> dict:
|
||
"""Preserve the flat order and its recorded activity without inventing stages."""
|
||
required = (
|
||
"work_order_id", "revision", "title", "summary", "status", "priority",
|
||
"origin", "updated_at",
|
||
)
|
||
snapshot = {key: order[key] for key in required}
|
||
snapshot.update({key: order.get(key) for key in (
|
||
"assignee", "scheduled_at", "location", "scenario_code",
|
||
)})
|
||
snapshot["feedback_requirements"] = order.get("feedback_requirements") or []
|
||
snapshot["activity"] = order.get("activity") or []
|
||
snapshot["work_order_type"] = order.get("work_order_type", "general")
|
||
return snapshot
|
||
|
||
|
||
async def work_order_pages(client, url, headers, work_order_ids, deleted_ids):
|
||
if work_order_ids is not None:
|
||
for order_id in sorted(work_order_ids):
|
||
response = await client.get(f"{url}/{order_id}", headers=headers)
|
||
if response.status_code == 404:
|
||
if response.json().get("code") != "WORK_ORDER_NOT_FOUND":
|
||
raise ValueError("peripheral work order lookup did not confirm deletion")
|
||
deleted_ids.add(order_id)
|
||
continue
|
||
response.raise_for_status()
|
||
order = response.json()
|
||
if order["work_order_id"] != order_id:
|
||
raise ValueError("peripheral work order identity mismatch")
|
||
yield {"items": [order]}
|
||
return
|
||
cursor = None
|
||
while True:
|
||
params = {"limit": 100}
|
||
if cursor:
|
||
params["cursor"] = cursor
|
||
response = await client.get(url, headers=headers, params=params)
|
||
response.raise_for_status()
|
||
page = response.json()
|
||
yield page
|
||
cursor = page.get("next_cursor")
|
||
if not cursor:
|
||
return
|
||
|
||
|
||
async def sync_once(client: httpx.AsyncClient, settings: WorkerSettings, peripheral_url: str,
|
||
credentials: CredentialProvider, *, work_order_ids: set[str] | None = None) -> int:
|
||
if work_order_ids is not None:
|
||
work_order_ids = {str(UUID(order_id)) for order_id in work_order_ids}
|
||
headers = await credentials.headers(client)
|
||
tasks_url = f"{settings.server_url.rstrip('/')}/api/v1/operational-tasks"
|
||
existing: dict[str, dict] = {}
|
||
event_tasks: dict[str | None, dict] = {}
|
||
investigations: dict[str, dict] = {}
|
||
for order_id in sorted(work_order_ids) if work_order_ids is not None else [None]:
|
||
offset = 0
|
||
while True:
|
||
params = {"limit": 100, "offset": offset}
|
||
if order_id is not None:
|
||
params["peripheral_work_order_id"] = order_id
|
||
response = await client.get(tasks_url, headers=headers, params=params)
|
||
response.raise_for_status()
|
||
page = response.json()
|
||
for task in page["items"]:
|
||
if task["task_type"] == "external_work_order":
|
||
existing[task["source"].get("external_id")] = task
|
||
elif task["task_type"] == "event_investigation":
|
||
event_tasks[task.get("source", {}).get("external_id")] = task
|
||
if task["status"] in {"queued", "waiting_human", "running"} or (
|
||
work_order_ids is not None and task["status"] == "completed"
|
||
):
|
||
investigations[task["task_id"]] = task
|
||
if page.get("next_offset") is None:
|
||
break
|
||
offset = page["next_offset"]
|
||
|
||
linked: dict[str, list[str]] = {}
|
||
for task in investigations.values():
|
||
response = await client.get(f"{tasks_url}/{task['task_id']}", headers=headers)
|
||
response.raise_for_status()
|
||
for action in response.json().get("actions", []):
|
||
receipt = action.get("receipt")
|
||
if action["kind"] == "create_work_order" and receipt and receipt["adapter"] == "peripheral_sample":
|
||
linked.setdefault(receipt["receipt_id"], []).append(task["task_id"])
|
||
linked_order_ids = set(linked)
|
||
|
||
synchronized = 0
|
||
deleted_ids: set[str] = set()
|
||
seen_order_ids: set[str] = set()
|
||
async for page in work_order_pages(
|
||
client, f"{peripheral_url.rstrip('/')}/api/sample/v1/work-orders", headers,
|
||
work_order_ids, deleted_ids,
|
||
):
|
||
for order in page["items"]:
|
||
if order["project_id"] != settings.project_id or order["origin"] != "synthetic":
|
||
raise ValueError("peripheral work order scope or origin mismatch")
|
||
seen_order_ids.add(order["work_order_id"])
|
||
if not should_project_work_order(order, linked_order_ids):
|
||
continue
|
||
external_id = f"peripheral_sample:{order['work_order_id']}"
|
||
# Complaint intake is an event even if no browser was open when the
|
||
# external system created it. Use the same source key as the UI.
|
||
event_kind = order.get("work_order_type")
|
||
if (
|
||
event_kind in EVENT_WORK_ORDER_TYPES
|
||
and order["work_order_id"] not in event_tasks
|
||
and order["work_order_id"] not in linked_order_ids
|
||
):
|
||
event_request_id = str(uuid5(NAMESPACE_URL, f"{settings.project_id}:{external_id}:investigation"))
|
||
event = await client.post(tasks_url, headers=headers, json={
|
||
"request_id": event_request_id, "task_type": "event_investigation", "workflow": event_kind,
|
||
"title": "水质投诉调查" if event_kind == "water_quality_complaint" else "爆管事件调查", "summary": f"合成外围事件：{order['summary']}",
|
||
"priority": "urgent" if order["priority"] in {"high", "urgent"} else "normal",
|
||
"source": {"kind": "external_system", "external_id": order["work_order_id"],
|
||
"external_revision": order["revision"], "event_kind": event_kind},
|
||
"trigger": {"kind": "event", "occurred_at": order.get("scheduled_at") or order["created_at"],
|
||
"detail": f"外围样板事件上报 · {order.get('location') or '位置待确认'}"},
|
||
})
|
||
event.raise_for_status()
|
||
event_tasks[order["work_order_id"]] = event.json()
|
||
task = existing.get(external_id)
|
||
if task is None:
|
||
request_id = str(uuid5(NAMESPACE_URL, f"{settings.project_id}:{external_id}:create"))
|
||
result = await client.post(tasks_url, headers=headers, json={
|
||
"request_id": request_id, "task_type": "external_work_order", "workflow": "external-work-order",
|
||
"title": order["title"], "summary": f"合成外围工单：{order['summary']}",
|
||
"priority": "urgent" if order["priority"] in {"high", "urgent"} else "normal",
|
||
"source": {"kind": "external_system", "external_id": external_id,
|
||
"event_kind": order.get("work_order_type", "general")},
|
||
"trigger": {"kind": "external_sync", "occurred_at": order["created_at"],
|
||
"detail": "外围样板工单同步，进度以外部系统反馈为准"},
|
||
"scheduled_for": order["scheduled_at"],
|
||
})
|
||
result.raise_for_status()
|
||
task = result.json()
|
||
existing[external_id] = task
|
||
changed = (task["source"].get("external_revision") or 0) < order["revision"]
|
||
if work_order_ids is None and changed and order["revision"] > 1:
|
||
# Only changed sources need completed-parent receipt discovery.
|
||
# Never fetch every historical completed investigation each tick.
|
||
offset = 0
|
||
while True:
|
||
related = await client.get(tasks_url, headers=headers, params={
|
||
"limit": 100, "offset": offset, "peripheral_work_order_id": order["work_order_id"],
|
||
})
|
||
related.raise_for_status()
|
||
related_page = related.json()
|
||
for parent in related_page["items"]:
|
||
if parent["task_type"] != "event_investigation" or parent.get("status") != "completed":
|
||
continue
|
||
detail = await client.get(f"{tasks_url}/{parent['task_id']}", headers=headers)
|
||
detail.raise_for_status()
|
||
for action in detail.json().get("actions", []):
|
||
receipt = action.get("receipt")
|
||
if (action["kind"] == "create_work_order" and receipt
|
||
and receipt["adapter"] == "peripheral_sample"
|
||
and receipt["receipt_id"] == order["work_order_id"]):
|
||
linked.setdefault(order["work_order_id"], []).append(parent["task_id"])
|
||
if related_page.get("next_offset") is None:
|
||
break
|
||
offset = related_page["next_offset"]
|
||
event_id = f"{external_id}:{order['revision']}"
|
||
notes = [activity["detail"] for activity in order.get("activity", [])
|
||
if activity.get("kind") == "feedback" and activity.get("detail")]
|
||
evidence_summary = order["summary"]
|
||
if notes:
|
||
evidence_summary += "\n现场反馈：\n" + "\n".join(notes[-10:])
|
||
feedback = {
|
||
"request_id": str(uuid5(NAMESPACE_URL, event_id)), "source_event_id": external_id,
|
||
"source_revision": order["revision"], "status": STATUS[order["status"]],
|
||
"occurred_at": order["updated_at"],
|
||
"detail": f"合成外围工单 {order['work_order_id']}：{order['status']}",
|
||
"evidence": {"evidence_id": event_id, "kind": "synthetic", "summary": evidence_summary,
|
||
"source_label": "外围样板合成工单", "source_ref": external_id,
|
||
"observed_at": order["updated_at"]},
|
||
}
|
||
# A child order's creation is already represented by its dispatch
|
||
# receipt. Only a subsequent source revision is new field input.
|
||
targets = set(linked.get(order["work_order_id"], [])) if order["revision"] > 1 else set()
|
||
if work_order_ids is None and not changed:
|
||
targets.clear()
|
||
event_task = event_tasks.get(order["work_order_id"])
|
||
if (event_task and event_task.get("status") != "cancelled"
|
||
and order["revision"] > (event_task["source"].get("external_revision") or 0)):
|
||
targets.add(event_task["task_id"])
|
||
# The external projection is the reconciliation cursor. Advance it
|
||
# only after all investigation feedback was accepted; retry remains
|
||
# safe if a response is lost or a later target fails.
|
||
ordered_targets = sorted(targets)
|
||
if changed:
|
||
ordered_targets.append(task["task_id"])
|
||
for target in ordered_targets:
|
||
payload = dict(feedback)
|
||
if target == task["task_id"]:
|
||
payload["work_order_snapshot"] = work_order_snapshot(order)
|
||
result = await client.post(f"{tasks_url}/{target}/external-feedback", headers=headers, json=payload)
|
||
result.raise_for_status()
|
||
synchronized += 1
|
||
|
||
for external_id, task in existing.items():
|
||
order_id = peripheral_order_id(external_id)
|
||
if order_id is None:
|
||
continue
|
||
if order_id in seen_order_ids:
|
||
continue
|
||
if work_order_ids is not None and order_id not in work_order_ids:
|
||
continue
|
||
# The list may change during pagination. Only a scoped 404 proves that
|
||
# the source order is now absent; transient errors must stop this tick.
|
||
if order_id not in deleted_ids:
|
||
response = await client.get(
|
||
f"{peripheral_url.rstrip('/')}/api/sample/v1/work-orders/{order_id}",
|
||
headers=headers,
|
||
)
|
||
if response.status_code == 200:
|
||
continue
|
||
if response.status_code != 404:
|
||
response.raise_for_status()
|
||
if response.json().get("code") != "WORK_ORDER_NOT_FOUND":
|
||
raise ValueError("peripheral work order lookup did not confirm deletion")
|
||
result = await client.post(
|
||
f"{tasks_url}/{task['task_id']}/withdraw-external-source",
|
||
headers=headers,
|
||
json={
|
||
"request_id": str(uuid5(NAMESPACE_URL, f"{settings.project_id}:{external_id}:withdraw")),
|
||
"source_event_id": external_id,
|
||
},
|
||
)
|
||
result.raise_for_status()
|
||
synchronized += 1
|
||
return synchronized
|
||
|
||
|
||
class PeripheralWorkOrderSynchronizer:
|
||
"""Wake on committed source events; periodically reconcile missed changes."""
|
||
|
||
def __init__(self, client: httpx.AsyncClient, settings: WorkerSettings,
|
||
peripheral_url: str, credentials: CredentialProvider) -> None:
|
||
self.client = client
|
||
self.settings = settings
|
||
self.peripheral_url = peripheral_url
|
||
self.credentials = credentials
|
||
self.wake = asyncio.Event()
|
||
self.cursor: int | None = None
|
||
self.pending_order_ids: set[str] = set()
|
||
self.full_reconcile = False
|
||
|
||
def _queue_order(self, order_id: str) -> None:
|
||
if not self.full_reconcile:
|
||
self.pending_order_ids.add(order_id)
|
||
if len(self.pending_order_ids) > MAX_TARGETED_ORDERS:
|
||
self.pending_order_ids.clear()
|
||
self.full_reconcile = True
|
||
self.wake.set()
|
||
|
||
async def _reset_cursor(self, headers: dict[str, str]) -> None:
|
||
response = await self.client.get(
|
||
f"{self.settings.server_url.rstrip('/')}/api/v1/workbench/sync-snapshot",
|
||
headers=headers,
|
||
)
|
||
response.raise_for_status()
|
||
cursor = response.json()["event_cursor"]
|
||
if not isinstance(cursor, int) or isinstance(cursor, bool) or cursor < 0:
|
||
raise ValueError("invalid workbench event cursor")
|
||
self.cursor = cursor
|
||
self.full_reconcile = True
|
||
# Capture the cursor before scanning sources, so updates during the
|
||
# initial/reset scan are replayed by the following SSE connection.
|
||
self.wake.set()
|
||
|
||
async def listen(self) -> None:
|
||
while True:
|
||
try:
|
||
headers = await self.credentials.headers(self.client)
|
||
if self.cursor is None:
|
||
await self._reset_cursor(headers)
|
||
async with self.client.stream(
|
||
"GET", f"{self.settings.server_url.rstrip('/')}/api/v1/workbench/events",
|
||
headers={**headers, "Accept": "text/event-stream"},
|
||
params={"after": self.cursor}, timeout=None,
|
||
) as response:
|
||
response.raise_for_status()
|
||
event_name = ""
|
||
async for line in response.aiter_lines():
|
||
if line.startswith("event:"):
|
||
event_name = line[6:].strip()
|
||
elif line.startswith("data:") and event_name == "reset_required":
|
||
await self._reset_cursor(headers)
|
||
break
|
||
elif line.startswith("data:") and event_name == "workbench_event":
|
||
event = WorkbenchEvent.model_validate(json.loads(line[5:].strip()))
|
||
if str(event.project_id) != self.settings.project_id:
|
||
raise ValueError("workbench event project mismatch")
|
||
if event.event_seq <= self.cursor:
|
||
continue
|
||
if (event.topic == "workbench.work_order.changed"
|
||
and event.aggregate_type == "peripheral_work_order"
|
||
and event.metadata.get("source") == "peripheral_sample"):
|
||
self._queue_order(str(UUID(event.aggregate_id)))
|
||
elif event.topic == "operational.action.changed":
|
||
# The source create can commit before its action
|
||
# receipt is saved. Revisit just those receipt IDs.
|
||
task_id = str(UUID(event.metadata["task_id"]))
|
||
task = await self.client.get(
|
||
f"{self.settings.server_url.rstrip('/')}/api/v1/operational-tasks/{task_id}",
|
||
headers=headers,
|
||
)
|
||
task.raise_for_status()
|
||
for action in task.json().get("actions", []):
|
||
receipt = action.get("receipt")
|
||
if (action["kind"] == "create_work_order" and receipt
|
||
and receipt["adapter"] == "peripheral_sample"):
|
||
self._queue_order(str(UUID(receipt["receipt_id"])))
|
||
self.cursor = event.event_seq
|
||
elif not line:
|
||
event_name = ""
|
||
await asyncio.sleep(1)
|
||
except (httpx.HTTPError, ValueError, KeyError):
|
||
logger.warning("Work-order event stream disconnected; reconciliation remains active")
|
||
await asyncio.sleep(5)
|
||
|
||
async def serve(self, reconcile_seconds: int = 300) -> None:
|
||
if reconcile_seconds < 1:
|
||
raise ValueError("reconciliation interval must be at least one second")
|
||
listener = asyncio.create_task(self.listen())
|
||
retry_seconds = 5
|
||
next_reconcile = asyncio.get_running_loop().time() + reconcile_seconds
|
||
try:
|
||
while True:
|
||
try:
|
||
remaining = max(0, next_reconcile - asyncio.get_running_loop().time())
|
||
await asyncio.wait_for(self.wake.wait(), timeout=remaining)
|
||
except TimeoutError:
|
||
self.full_reconcile = True
|
||
# Clear before work, so arrivals during a scan request another
|
||
# pass. A failed scan gets a retry even after its cursor advances.
|
||
self.wake.clear()
|
||
full_reconcile = self.full_reconcile
|
||
self.full_reconcile = False
|
||
order_ids, self.pending_order_ids = self.pending_order_ids, set()
|
||
try:
|
||
count = await sync_once(
|
||
self.client, self.settings, self.peripheral_url, self.credentials,
|
||
**({"work_order_ids": order_ids} if not full_reconcile else {}),
|
||
)
|
||
logger.info("Peripheral progress synchronization completed: %d revisions", count)
|
||
retry_seconds = 5
|
||
if full_reconcile:
|
||
next_reconcile = asyncio.get_running_loop().time() + reconcile_seconds
|
||
except (httpx.HTTPError, ValueError, KeyError):
|
||
logger.error("Peripheral synchronization failed; persisted revisions will be retried")
|
||
self.full_reconcile |= full_reconcile
|
||
for order_id in order_ids:
|
||
self._queue_order(order_id)
|
||
await asyncio.sleep(retry_seconds)
|
||
self.wake.set()
|
||
retry_seconds = min(retry_seconds * 2, 60)
|
||
finally:
|
||
listener.cancel()
|
||
await asyncio.gather(listener, return_exceptions=True)
|
||
|
||
|
||
async def run(once: bool, reconcile_seconds: int = 300) -> None:
|
||
settings = WorkerSettings.from_environment()
|
||
peripheral_url = os.environ["OPERATIONAL_ACTION_PERIPHERAL_URL"]
|
||
credentials = CredentialProvider(settings)
|
||
async with httpx.AsyncClient(timeout=30, follow_redirects=False, trust_env=False) as client:
|
||
if once:
|
||
try:
|
||
count = await sync_once(client, settings, peripheral_url, credentials)
|
||
logger.info("Peripheral progress synchronization completed: %d revisions", count)
|
||
except (httpx.HTTPError, ValueError, KeyError):
|
||
logger.error("Peripheral synchronization failed; persisted revisions remain available")
|
||
raise SystemExit(1) from None
|
||
return
|
||
await PeripheralWorkOrderSynchronizer(client, settings, peripheral_url, credentials).serve(reconcile_seconds)
|
||
|
||
|
||
if __name__ == "__main__":
|
||
parser = argparse.ArgumentParser(description=__doc__)
|
||
parser.add_argument("--once", action="store_true")
|
||
parser.add_argument("--reconcile-seconds", type=int, default=300,
|
||
help="missed-event reconciliation interval (default: 300 seconds)")
|
||
args = parser.parse_args()
|
||
if args.reconcile_seconds < 1:
|
||
parser.error("--reconcile-seconds must be at least 1")
|
||
logging.basicConfig(level=logging.INFO)
|
||
try:
|
||
asyncio.run(run(args.once, args.reconcile_seconds))
|
||
except KeyboardInterrupt:
|
||
pass
|