JiuwenSwarm is an intelligent AI Agent built on openJiuwen. It extends the powerful capabilities of large language models directly to your fingertips through various communication apps you use daily.
Python
"""Tests for bounded inbound payload logging."""
from __future__ import annotations
import json
from unittest.mock import patch
from jiuwenswarm.common.schema.message import ReqMethod
from jiuwenswarm.server.wire_parse import parse_inbound
def test_catalog_sync_logs_summary_without_changing_parsed_request() -> None:
large_marker = "catalog-secret-value-" * 10_000
payload = {
"request_id": "sync-1",
"channel_id": "officeclaw",
"req_method": ReqMethod.SYNC_AGENTS_CONFIGS.value,
"params": {
"revision": "revision-1",
"service_id": "default",
"agents": [
{"agent_id": "office", "config": {"marker": large_marker}},
{"agent_id": "assistant", "config": {}},
],
},
}
raw = json.dumps(payload)
with patch("jiuwenswarm.server.wire_parse.logger.info") as log_info:
result = parse_inbound(raw)
assert result.ok
assert result.request is not None
assert result.request.request_id == payload["request_id"]
assert result.request.req_method is ReqMethod.SYNC_AGENTS_CONFIGS
assert result.request.params == payload["params"]
first_log = log_info.call_args_list[0]
rendered_call = repr(first_log)
assert "Inbound raw payload: <summary>" in first_log.args[0]
assert "agent_count=%s" in first_log.args[0]
assert 2 in first_log.args
assert large_marker not in rendered_call
def test_chat_payload_keeps_existing_masked_diagnostic_log() -> None:
payload = {
"request_id": "chat-1",
"channel_id": "officeclaw",
"req_method": ReqMethod.CHAT_SEND.value,
"params": {"query": "private chat text"},
}
with patch("jiuwenswarm.server.wire_parse.logger.info") as log_info:
result = parse_inbound(json.dumps(payload))
assert result.ok
first_log = log_info.call_args_list[0]
assert first_log.args[0] == "[AgentWebSocketServer] Inbound raw payload: %s"
assert first_log.args[1]["params"]["query"] == "******"