Codex
1458 lines
73 KiB
Python
1458 lines
73 KiB
Python
import contextlib
|
|
import io
|
|
import json
|
|
import os
|
|
import tempfile
|
|
import unittest
|
|
from pathlib import Path
|
|
from unittest.mock import Mock
|
|
|
|
os.environ.setdefault("DATABASE_URL", "postgresql+psycopg://user:pass@localhost:5432/ai_ict_test")
|
|
|
|
from scripts import run_until_blocked
|
|
|
|
|
|
class RunUntilBlockedTests(unittest.TestCase):
|
|
def setUp(self) -> None:
|
|
self._status_tmp = tempfile.TemporaryDirectory()
|
|
self._original_runner_status_path = run_until_blocked.RUNNER_STATUS_PATH
|
|
self._original_monitor_history_path = run_until_blocked.ANALYSIS_DECISION_MONITOR_HISTORY_PATH
|
|
self._original_monitor_trend_path = run_until_blocked.ANALYSIS_DECISION_MONITOR_TREND_PATH
|
|
self._original_monitor_quality_gate_path = run_until_blocked.ANALYSIS_DECISION_MONITOR_QUALITY_GATE_PATH
|
|
self._original_monitor_maturity_rollup_path = run_until_blocked.ANALYSIS_DECISION_MONITOR_MATURITY_ROLLUP_PATH
|
|
self._original_monitor_remote_drift_check_path = run_until_blocked.ANALYSIS_DECISION_MONITOR_REMOTE_DRIFT_CHECK_PATH
|
|
self._original_support_quality_review_path = run_until_blocked.ANALYSIS_DECISION_SUPPORT_QUALITY_REVIEW_PATH
|
|
self._original_monitor_p7_selector_path = run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_PATH
|
|
run_until_blocked.RUNNER_STATUS_PATH = Path(self._status_tmp.name) / "automation_status.json"
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_HISTORY_PATH = Path(self._status_tmp.name) / "analysis_decision_monitor_history.jsonl"
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_TREND_PATH = Path(self._status_tmp.name) / "analysis_decision_monitor_trend.json"
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_QUALITY_GATE_PATH = Path(self._status_tmp.name) / "analysis_decision_monitor_quality_gate.json"
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_MATURITY_ROLLUP_PATH = Path(self._status_tmp.name) / "analysis_decision_monitor_maturity_rollup.json"
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_REMOTE_DRIFT_CHECK_PATH = Path(self._status_tmp.name) / "analysis_decision_monitor_remote_drift_check.json"
|
|
run_until_blocked.ANALYSIS_DECISION_SUPPORT_QUALITY_REVIEW_PATH = Path(self._status_tmp.name) / "analysis_decision_support_quality_review.json"
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_PATH = Path(self._status_tmp.name) / "analysis_decision_monitor_p7_slice_selector.json"
|
|
|
|
def tearDown(self) -> None:
|
|
run_until_blocked.RUNNER_STATUS_PATH = self._original_runner_status_path
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_HISTORY_PATH = self._original_monitor_history_path
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_TREND_PATH = self._original_monitor_trend_path
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_QUALITY_GATE_PATH = self._original_monitor_quality_gate_path
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_MATURITY_ROLLUP_PATH = self._original_monitor_maturity_rollup_path
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_REMOTE_DRIFT_CHECK_PATH = self._original_monitor_remote_drift_check_path
|
|
run_until_blocked.ANALYSIS_DECISION_SUPPORT_QUALITY_REVIEW_PATH = self._original_support_quality_review_path
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_PATH = self._original_monitor_p7_selector_path
|
|
self._status_tmp.cleanup()
|
|
|
|
def test_selector_picks_first_unfinished_execution_layer_slice(self) -> None:
|
|
selection = run_until_blocked.select_automation_execution_slice(
|
|
anchor_state={
|
|
"missing_required_anchors": [],
|
|
"completed_execution_slices": ["automation_execution_layer_p0_governance_gate"],
|
|
"in_progress_execution_slices": ["automation_execution_layer_p1_anchor_slice_selector"],
|
|
},
|
|
tick_payload={
|
|
"status": "no_op_healthy",
|
|
"recommended_actions": ["continue_periodic_optimization_cycle"],
|
|
},
|
|
)
|
|
|
|
self.assertEqual(selection["status"], "ready")
|
|
self.assertEqual(selection["slice"], "automation_execution_layer_p1_anchor_slice_selector")
|
|
self.assertTrue(selection["requires_code"])
|
|
self.assertTrue(selection["already_in_progress"])
|
|
|
|
def test_required_automation_anchor_paths_match_new_kickoff_prompt(self) -> None:
|
|
anchors = run_until_blocked.get_automation_required_anchor_paths()
|
|
|
|
self.assertEqual(
|
|
[path.name for path in anchors],
|
|
[
|
|
"AI_ICT_实盘交易分析决策助手下一阶段开发方案_v1.md",
|
|
"SESSION.md",
|
|
"TASKS.md",
|
|
],
|
|
)
|
|
|
|
def test_render_session_uses_new_kickoff_anchors(self) -> None:
|
|
text = run_until_blocked.render_session(
|
|
completed_stages=[],
|
|
current_stage="P0-2",
|
|
last_result={},
|
|
blocked_reason=None,
|
|
)
|
|
|
|
self.assertIn("AI_ICT_实盘交易分析决策助手下一阶段开发方案_v1.md", text)
|
|
self.assertIn("SESSION.md", text)
|
|
self.assertIn("TASKS.md", text)
|
|
self.assertNotIn("AI_ICT_实盘辅助决策系统开发方案_v2.md", text)
|
|
|
|
def test_selector_blocks_when_required_anchors_are_missing(self) -> None:
|
|
selection = run_until_blocked.select_automation_execution_slice(
|
|
anchor_state={
|
|
"missing_required_anchors": ["docs/missing.md"],
|
|
"completed_execution_slices": [],
|
|
"in_progress_execution_slices": [],
|
|
},
|
|
tick_payload={"status": "no_op_healthy"},
|
|
)
|
|
|
|
self.assertEqual(selection["status"], "blocked")
|
|
self.assertEqual(selection["slice"], "repair_missing_automation_anchors")
|
|
self.assertEqual(selection["blocked_reason"], "missing_required_anchors:docs/missing.md")
|
|
self.assertFalse(selection["requires_code"])
|
|
|
|
def test_validation_plan_for_execution_layer_slice_uses_focused_checks(self) -> None:
|
|
plan = run_until_blocked.build_validation_plan(
|
|
selected_slice={
|
|
"slice": "automation_execution_layer_p2_validation_quality_router",
|
|
"requires_code": True,
|
|
},
|
|
changed_files=["scripts/run_until_blocked.py", "tests/unit/test_run_until_blocked.py"],
|
|
)
|
|
|
|
self.assertIn("D:\\ai-exchange\\.venv\\Scripts\\python.exe -m unittest tests.unit.test_run_until_blocked", plan["focused_commands"])
|
|
self.assertIn("D:\\ai-exchange\\.venv\\Scripts\\python.exe scripts/run_script_static_checks.py", plan["quality_gates"])
|
|
self.assertIn("npm run lint", plan["quality_gates"])
|
|
self.assertFalse(plan["deploy_required"])
|
|
|
|
def test_validation_plan_requires_deploy_for_declared_runtime_api_changes(self) -> None:
|
|
plan = run_until_blocked.build_validation_plan(
|
|
selected_slice={
|
|
"slice": "automation_execution_layer_p3_deploy_smoke_router",
|
|
"requires_code": True,
|
|
},
|
|
changed_files=["src/api/research_council.py"],
|
|
)
|
|
|
|
self.assertTrue(plan["deploy_required"])
|
|
self.assertEqual(plan["deploy_reason"], "declared_runtime_api_db_or_workbench_change")
|
|
self.assertIn("deploy_smoke_required_by_declared_runtime_api_db_or_workbench_change", plan["notes"])
|
|
self.assertTrue(plan["deploy_smoke_plan"]["deploy_required"])
|
|
self.assertIn("/health", plan["deploy_smoke_plan"]["smoke_targets"])
|
|
self.assertIn("POST /trading/research-council/governance-tick", plan["deploy_smoke_plan"]["smoke_targets"])
|
|
|
|
def test_deploy_smoke_plan_stays_local_for_script_docs_state_changes(self) -> None:
|
|
plan = run_until_blocked.build_validation_plan(
|
|
selected_slice={
|
|
"slice": "automation_execution_layer_p3_deploy_smoke_router",
|
|
"requires_code": True,
|
|
},
|
|
changed_files=[
|
|
"scripts/run_until_blocked.py",
|
|
"tests/unit/test_run_until_blocked.py",
|
|
"docs/AI_ICT_自动化流程使用教程_v1.md",
|
|
"SESSION.md",
|
|
"TASKS.md",
|
|
],
|
|
)
|
|
|
|
self.assertFalse(plan["deploy_required"])
|
|
self.assertFalse(plan["deploy_smoke_plan"]["deploy_required"])
|
|
self.assertEqual(plan["deploy_smoke_plan"]["smoke_targets"], [])
|
|
|
|
def test_state_writeback_plan_tracks_status_and_next_slice(self) -> None:
|
|
summary = run_until_blocked.build_automation_execution_summary(
|
|
started_ts="2026-05-11T00:00:00+00:00",
|
|
tick_payload={
|
|
"status": "no_op_healthy",
|
|
"effective_health_state": "healthy",
|
|
"safe_to_auto_continue": True,
|
|
"recommended_actions": ["continue_periodic_optimization_cycle"],
|
|
"boundary_notes": [],
|
|
},
|
|
anchor_state={
|
|
"missing_required_anchors": [],
|
|
"completed_execution_slices": [
|
|
"automation_execution_layer_p0_governance_gate",
|
|
"automation_execution_layer_p1_anchor_slice_selector",
|
|
"automation_execution_layer_p2_validation_quality_router",
|
|
"automation_execution_layer_p3_deploy_smoke_router",
|
|
],
|
|
"in_progress_execution_slices": ["automation_execution_layer_p4_state_writeback_renderer"],
|
|
},
|
|
changed_files=["scripts/run_until_blocked.py"],
|
|
)
|
|
|
|
plan = summary["state_writeback_plan"]
|
|
self.assertEqual(summary["next_slice"], "automation_execution_layer_p4_state_writeback_renderer")
|
|
self.assertEqual(plan["state"], "automation_execution_layer_p3_deploy_smoke_router_closed")
|
|
self.assertIn("SESSION.md", plan["files_to_write"])
|
|
self.assertIn("TASKS.md", plan["files_to_write"])
|
|
self.assertTrue(plan["no_new_phase_created"])
|
|
self.assertIn("execution_allowed", plan["guards_unchanged"])
|
|
|
|
def test_automation_execution_layer_noops_on_healthy_governance_tick(self) -> None:
|
|
class _HealthyGovernanceTickService:
|
|
def run_tick(self, **kwargs):
|
|
return {
|
|
"status": "no_op_healthy",
|
|
"passed": True,
|
|
"instrument_id": kwargs["instrument_id"],
|
|
"effective_health_state": "healthy",
|
|
"safe_to_auto_continue": True,
|
|
"requires_cycle_run": False,
|
|
"requires_bounded_slice": False,
|
|
"recommended_actions": ["continue_periodic_optimization_cycle"],
|
|
"boundary_notes": [
|
|
"governance_tick_does_not_change_execution_gate",
|
|
"governance_tick_does_not_change_live_readiness",
|
|
"governance_tick_does_not_change_trust_bucket",
|
|
],
|
|
}
|
|
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text(
|
|
"# ai-exchange Session\n\n- state: automation_flow_p0_p6_review_closed\n",
|
|
encoding="utf-8",
|
|
)
|
|
tasks_path.write_text(
|
|
"# ai-exchange Tasks\n\n- [ ] Start automation execution-layer development\n",
|
|
encoding="utf-8",
|
|
)
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
original_service = run_until_blocked.CouncilGovernanceTickService
|
|
original_rehearsal = run_until_blocked.run_trading_delivery_rehearsal
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
run_until_blocked.CouncilGovernanceTickService = _HealthyGovernanceTickService
|
|
run_until_blocked.run_trading_delivery_rehearsal = lambda **kwargs: (_ for _ in ()).throw(
|
|
AssertionError("legacy stage runner should not be called")
|
|
)
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_OUTPUT_JSON": "0"})
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
status_payload = json.loads(status_path.read_text(encoding="utf-8"))
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
run_until_blocked.CouncilGovernanceTickService = original_service
|
|
run_until_blocked.run_trading_delivery_rehearsal = original_rehearsal
|
|
|
|
self.assertEqual(summary["status"], "completed")
|
|
self.assertEqual(summary["mode"], "automation_execution_layer")
|
|
self.assertEqual(summary["tick_status"], "no_op_healthy")
|
|
self.assertEqual(summary["next_slice"], "automation_execution_layer_p1_anchor_slice_selector")
|
|
self.assertEqual(summary["selected_slice"]["reason"], "first_unfinished_execution_layer_slice")
|
|
self.assertEqual(summary["anchor_state"]["state"], "automation_flow_p0_p6_review_closed")
|
|
self.assertIn("npm run lint", summary["validation_plan"]["quality_gates"])
|
|
self.assertTrue(summary["no_new_phase_created"])
|
|
self.assertIn("automation_execution_layer_p0_governance_gate_closed", session_text)
|
|
self.assertIn("Automation Execution-Layer P0", tasks_text)
|
|
self.assertNotIn("### P7", session_text)
|
|
self.assertNotIn("- [ ] `P7", tasks_text)
|
|
self.assertEqual(status_payload["state"], "completed")
|
|
self.assertEqual(status_payload["mode"], "automation_execution_layer")
|
|
self.assertEqual(status_payload["tick_status"], "no_op_healthy")
|
|
|
|
def test_analysis_decision_monitor_phase_runs_monitor_instead_of_legacy_stages(self) -> None:
|
|
monitor_payload = {
|
|
"passed": True,
|
|
"generated_ts": "2026-05-11T04:45:00+00:00",
|
|
"phase": "analysis_decision_only",
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": {
|
|
"broker_status": "broker_state_ok",
|
|
"broker_fills": 0,
|
|
"runtime_health": "ok",
|
|
"market_stream": "market_stream_connected",
|
|
"market_freshness": "market_data_fresh",
|
|
"market_gap": "market_data_gap_clear",
|
|
"real_trade_sample_gate_required": False,
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
"decision_action": "restricted_decision_support",
|
|
"decision_operator_mode": "observe_only",
|
|
"decision_trust_classification": "blacklist",
|
|
"p6_governance_tick_status": "no_op_healthy",
|
|
},
|
|
}
|
|
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text(
|
|
"# ai-exchange Session\n\n"
|
|
"- state: analysis_decision_monitor_runner_deployed\n"
|
|
"- current_phase: Restricted Decision-Support Soft Rollout\n"
|
|
"- repair_next_action: run_analysis_decision_monitor_periodically_without_fill_gate\n",
|
|
encoding="utf-8",
|
|
)
|
|
tasks_path.write_text(
|
|
"# ai-exchange Tasks\n\n"
|
|
"## In Progress\n\n"
|
|
"- [ ] `run_analysis_decision_monitor_periodically_without_fill_gate`\n",
|
|
encoding="utf-8",
|
|
)
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
original_monitor = run_until_blocked.run_analysis_decision_monitor.run_monitor
|
|
original_rehearsal = run_until_blocked.run_trading_delivery_rehearsal
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
run_until_blocked.run_analysis_decision_monitor.run_monitor = lambda **kwargs: monitor_payload
|
|
run_until_blocked.run_trading_delivery_rehearsal = lambda **kwargs: (_ for _ in ()).throw(
|
|
AssertionError("legacy stage runner should not be called")
|
|
)
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_OUTPUT_JSON": "0"})
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
status_payload = json.loads(status_path.read_text(encoding="utf-8"))
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
run_until_blocked.run_analysis_decision_monitor.run_monitor = original_monitor
|
|
run_until_blocked.run_trading_delivery_rehearsal = original_rehearsal
|
|
|
|
self.assertEqual(summary["status"], "completed")
|
|
self.assertEqual(summary["mode"], "analysis_decision_monitor")
|
|
self.assertTrue(summary["monitor_passed"])
|
|
self.assertEqual(summary["next_slice"], "continue_periodic_analysis_decision_monitoring")
|
|
self.assertIsNone(summary["blocked_reason"])
|
|
self.assertIn("analysis_decision_monitor_runner_automation_wired", session_text)
|
|
self.assertIn("latest_analysis_decision_monitor_status: passed", session_text)
|
|
self.assertIn("latest_real_trade_sample_gate_status: deferred_analysis_decision_only", session_text)
|
|
self.assertIn("continue_periodic_analysis_decision_monitoring", tasks_text)
|
|
self.assertNotIn("wait_for_and_verify_real_fill_sample", session_text)
|
|
self.assertEqual(status_payload["state"], "completed")
|
|
self.assertEqual(status_payload["mode"], "analysis_decision_monitor")
|
|
self.assertEqual(status_payload["monitor_blockers"], [])
|
|
self.assertEqual(summary["monitor_history"]["record_count"], 1)
|
|
self.assertTrue(run_until_blocked.ANALYSIS_DECISION_MONITOR_HISTORY_PATH.exists())
|
|
history_record = json.loads(run_until_blocked.ANALYSIS_DECISION_MONITOR_HISTORY_PATH.read_text(encoding="utf-8").strip())
|
|
self.assertTrue(history_record["monitor_passed"])
|
|
self.assertEqual(history_record["observed"]["broker_fills"], 0)
|
|
self.assertIn("latest_analysis_decision_monitor_history_records: 1", session_text)
|
|
self.assertIn("latest_analysis_decision_monitor_history_records: 1", tasks_text)
|
|
self.assertTrue(run_until_blocked.ANALYSIS_DECISION_MONITOR_TREND_PATH.exists())
|
|
trend_payload = json.loads(run_until_blocked.ANALYSIS_DECISION_MONITOR_TREND_PATH.read_text(encoding="utf-8"))
|
|
self.assertEqual(trend_payload["record_count"], 1)
|
|
self.assertEqual(trend_payload["consecutive_pass_count"], 1)
|
|
self.assertFalse(trend_payload["stable_monitoring_window"])
|
|
self.assertIn("latest_analysis_decision_monitor_consecutive_passes: 1", session_text)
|
|
self.assertIn("latest_analysis_decision_monitor_stable_window: false", tasks_text)
|
|
self.assertTrue(run_until_blocked.ANALYSIS_DECISION_MONITOR_QUALITY_GATE_PATH.exists())
|
|
quality_gate_payload = json.loads(run_until_blocked.ANALYSIS_DECISION_MONITOR_QUALITY_GATE_PATH.read_text(encoding="utf-8"))
|
|
self.assertEqual(quality_gate_payload["status"], "attention_required")
|
|
self.assertIn("stable_monitoring_window_not_reached", quality_gate_payload["failure_reasons"])
|
|
self.assertIn("latest_analysis_decision_monitor_quality_gate_status: attention_required", session_text)
|
|
|
|
def test_analysis_decision_monitor_routes_to_p7_after_stable_quality_gate(self) -> None:
|
|
observed = {
|
|
"broker_status": "broker_state_ok",
|
|
"broker_fills": 0,
|
|
"runtime_health": "ok",
|
|
"market_stream": "market_stream_connected",
|
|
"market_freshness": "market_data_fresh",
|
|
"market_gap": "market_data_gap_clear",
|
|
"real_trade_sample_gate_required": False,
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
"decision_action": "restricted_decision_support",
|
|
"decision_operator_mode": "observe_only",
|
|
"decision_trust_classification": "blacklist",
|
|
"p6_governance_tick_status": "no_op_healthy",
|
|
}
|
|
existing_records = [
|
|
{
|
|
"recorded_ts": "2026-05-11T04:35:00+00:00",
|
|
"monitor_generated_ts": "2026-05-11T04:35:00+00:00",
|
|
"status": "completed",
|
|
"monitor_passed": True,
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": observed,
|
|
"next_slice": "continue_periodic_analysis_decision_monitoring",
|
|
"guards_unchanged": ["execution_allowed"],
|
|
},
|
|
{
|
|
"recorded_ts": "2026-05-11T04:40:00+00:00",
|
|
"monitor_generated_ts": "2026-05-11T04:40:00+00:00",
|
|
"status": "completed",
|
|
"monitor_passed": True,
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": observed,
|
|
"next_slice": "continue_periodic_analysis_decision_monitoring",
|
|
"guards_unchanged": ["execution_allowed"],
|
|
},
|
|
]
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_HISTORY_PATH.write_text(
|
|
"\n".join(json.dumps(record, sort_keys=True) for record in existing_records) + "\n",
|
|
encoding="utf-8",
|
|
)
|
|
monitor_payload = {
|
|
"passed": True,
|
|
"generated_ts": "2026-05-11T04:45:00+00:00",
|
|
"phase": "analysis_decision_only",
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": observed,
|
|
}
|
|
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text(
|
|
"# ai-exchange Session\n\n"
|
|
"- state: analysis_decision_monitor_runner_deployed\n"
|
|
"- current_phase: Restricted Decision-Support Soft Rollout\n"
|
|
"- repair_next_action: continue_periodic_analysis_decision_monitoring\n",
|
|
encoding="utf-8",
|
|
)
|
|
tasks_path.write_text(
|
|
"# ai-exchange Tasks\n\n"
|
|
"## In Progress\n\n"
|
|
"- [ ] `continue_periodic_analysis_decision_monitoring`\n",
|
|
encoding="utf-8",
|
|
)
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
original_monitor = run_until_blocked.run_analysis_decision_monitor.run_monitor
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
run_until_blocked.run_analysis_decision_monitor.run_monitor = lambda **kwargs: monitor_payload
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_OUTPUT_JSON": "0"})
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
run_until_blocked.run_analysis_decision_monitor.run_monitor = original_monitor
|
|
|
|
self.assertEqual(summary["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE)
|
|
self.assertEqual(summary["monitor_history"]["record_count"], 3)
|
|
self.assertEqual(summary["monitor_history"]["quality_gate"]["status"], "passed")
|
|
self.assertEqual(
|
|
summary["monitor_history"]["quality_gate"]["next_slice"],
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE,
|
|
)
|
|
self.assertIn(run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE, session_text)
|
|
self.assertIn("automation execution-layer P7", session_text)
|
|
self.assertIn("Stable monitor quality may enter automation execution-layer P7", tasks_text)
|
|
|
|
def test_p7_monitor_maturity_rollup_closes_and_routes_to_remote_drift_check(self) -> None:
|
|
observed = {
|
|
"broker_status": "broker_state_ok",
|
|
"broker_fills": 0,
|
|
"runtime_health": "ok",
|
|
"market_gap": "market_data_gap_clear",
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
"decision_action": "restricted_decision_support",
|
|
"decision_operator_mode": "observe_only",
|
|
"decision_trust_classification": "blacklist",
|
|
}
|
|
trend = {
|
|
"record_count": 9,
|
|
"consecutive_pass_count": 9,
|
|
"stable_sample_threshold": 3,
|
|
"stable_monitoring_window": True,
|
|
"latest_monitor_passed": True,
|
|
"latest_monitor_generated_ts": "2026-05-11T08:47:56+00:00",
|
|
"latest_blockers": [],
|
|
"latest_warnings": [],
|
|
"latest_observed": observed,
|
|
"next_slice": run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE,
|
|
"guards_unchanged": [
|
|
"execution_allowed",
|
|
"real_live_ready",
|
|
"risk_governor",
|
|
"trust_bucket",
|
|
"operator_confirmation_protocol",
|
|
],
|
|
}
|
|
quality_gate = {
|
|
"status": "passed",
|
|
"passed": True,
|
|
"failure_reasons": [],
|
|
"record_count": 9,
|
|
"consecutive_pass_count": 9,
|
|
"stable_monitoring_window": True,
|
|
"latest_monitor_generated_ts": "2026-05-11T08:47:56+00:00",
|
|
"latest_observed": observed,
|
|
"next_slice": run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE,
|
|
"guards_unchanged": trend["guards_unchanged"],
|
|
}
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_TREND_PATH.write_text(
|
|
json.dumps(trend),
|
|
encoding="utf-8",
|
|
)
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_QUALITY_GATE_PATH.write_text(
|
|
json.dumps(quality_gate),
|
|
encoding="utf-8",
|
|
)
|
|
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text(
|
|
"# ai-exchange Session\n\n"
|
|
"### automation_execution_layer_p7_monitor_maturity_rollup\n",
|
|
encoding="utf-8",
|
|
)
|
|
tasks_path.write_text(
|
|
"# ai-exchange Tasks\n\n"
|
|
"## In Progress\n\n"
|
|
f"- [ ] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE}`\n",
|
|
encoding="utf-8",
|
|
)
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_OUTPUT_JSON": "0"})
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
status_payload = json.loads(status_path.read_text(encoding="utf-8"))
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
|
|
self.assertEqual(summary["mode"], "analysis_decision_monitor_maturity_rollup")
|
|
self.assertEqual(summary["status"], "completed")
|
|
self.assertEqual(summary["rollup"]["state"], "automation_execution_layer_p7_monitor_maturity_rollup_closed")
|
|
self.assertEqual(summary["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_REMOTE_DRIFT_SLICE)
|
|
self.assertTrue(run_until_blocked.ANALYSIS_DECISION_MONITOR_MATURITY_ROLLUP_PATH.exists())
|
|
rollup_payload = json.loads(run_until_blocked.ANALYSIS_DECISION_MONITOR_MATURITY_ROLLUP_PATH.read_text(encoding="utf-8"))
|
|
self.assertEqual(rollup_payload["maturity_status"], "mature")
|
|
self.assertEqual(rollup_payload["remote_drift_check_status"], "pending")
|
|
self.assertIn("automation_execution_layer_p7_monitor_maturity_rollup_closed", session_text)
|
|
self.assertIn(run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_REMOTE_DRIFT_SLICE, tasks_text)
|
|
self.assertEqual(status_payload["mode"], "analysis_decision_monitor_maturity_rollup")
|
|
|
|
def test_p7_remote_drift_check_closes_and_routes_to_quality_review(self) -> None:
|
|
observed = {
|
|
"broker_status": "broker_state_ok",
|
|
"broker_fills": 0,
|
|
"runtime_health": "ok",
|
|
"market_gap": "market_data_gap_clear",
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
"decision_action": "restricted_decision_support",
|
|
"decision_operator_mode": "observe_only",
|
|
"decision_trust_classification": "blacklist",
|
|
"p6_governance_tick_status": "no_op_healthy",
|
|
}
|
|
guards = [
|
|
"execution_allowed",
|
|
"real_live_ready",
|
|
"risk_governor",
|
|
"trust_bucket",
|
|
"operator_confirmation_protocol",
|
|
]
|
|
local_rollup = {
|
|
"maturity_status": "mature",
|
|
"local": {
|
|
"quality_gate_status": "passed",
|
|
"record_count": 9,
|
|
"consecutive_pass_count": 9,
|
|
"stable_monitoring_window": True,
|
|
},
|
|
"latest_observed": observed,
|
|
"guards_unchanged": guards,
|
|
}
|
|
remote_quality = {
|
|
"status": "passed",
|
|
"passed": True,
|
|
"record_count": 12,
|
|
"consecutive_pass_count": 4,
|
|
"stable_monitoring_window": True,
|
|
"latest_observed": observed,
|
|
"guards_unchanged": guards,
|
|
}
|
|
remote_state = {
|
|
"status": "available",
|
|
"host": "192.168.3.90",
|
|
"container_quality_gate": remote_quality,
|
|
"container_trend": {"stable_sample_threshold": 3, "latest_observed": observed},
|
|
"host_quality_gate": remote_quality,
|
|
}
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_MATURITY_ROLLUP_PATH.write_text(
|
|
json.dumps(local_rollup),
|
|
encoding="utf-8",
|
|
)
|
|
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text(
|
|
"# ai-exchange Session\n\n"
|
|
"### automation_execution_layer_p7_remote_drift_check\n",
|
|
encoding="utf-8",
|
|
)
|
|
tasks_path.write_text(
|
|
"# ai-exchange Tasks\n\n"
|
|
"## In Progress\n\n"
|
|
f"- [ ] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_REMOTE_DRIFT_SLICE}`\n",
|
|
encoding="utf-8",
|
|
)
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
original_fetch_remote = run_until_blocked.fetch_remote_analysis_decision_monitor_state
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
run_until_blocked.fetch_remote_analysis_decision_monitor_state = lambda *, env: remote_state
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_OUTPUT_JSON": "0"})
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
status_payload = json.loads(status_path.read_text(encoding="utf-8"))
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
run_until_blocked.fetch_remote_analysis_decision_monitor_state = original_fetch_remote
|
|
|
|
self.assertEqual(summary["mode"], "analysis_decision_monitor_remote_drift_check")
|
|
self.assertEqual(summary["status"], "completed")
|
|
self.assertEqual(summary["drift_check"]["drift_status"], "clear")
|
|
self.assertEqual(summary["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_QUALITY_REVIEW_SLICE)
|
|
self.assertTrue(run_until_blocked.ANALYSIS_DECISION_MONITOR_REMOTE_DRIFT_CHECK_PATH.exists())
|
|
self.assertIn("automation_execution_layer_p7_remote_drift_check_closed", session_text)
|
|
self.assertIn(run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_QUALITY_REVIEW_SLICE, tasks_text)
|
|
self.assertEqual(status_payload["mode"], "analysis_decision_monitor_remote_drift_check")
|
|
|
|
def test_p7_remote_drift_check_blocks_when_remote_quality_not_passed(self) -> None:
|
|
local_rollup = {
|
|
"maturity_status": "mature",
|
|
"local": {"quality_gate_status": "passed"},
|
|
"latest_observed": {"runtime_health": "ok"},
|
|
"guards_unchanged": ["execution_allowed"],
|
|
}
|
|
remote_state = {
|
|
"status": "available",
|
|
"container_quality_gate": {"status": "attention_required", "latest_observed": {"runtime_health": "noisy"}},
|
|
"container_trend": {"stable_sample_threshold": 3, "consecutive_pass_count": 0},
|
|
"host_quality_gate": {"status": "attention_required"},
|
|
}
|
|
|
|
drift = run_until_blocked.build_analysis_decision_monitor_remote_drift_check(
|
|
local_rollup=local_rollup,
|
|
remote_state=remote_state,
|
|
)
|
|
|
|
self.assertEqual(drift["status"], "blocked")
|
|
self.assertEqual(drift["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_REMOTE_DRIFT_REPAIR_SLICE)
|
|
self.assertIn("remote_container_quality_gate_not_passed", drift["failure_reasons"])
|
|
|
|
def test_p7_selectors_ignore_completed_milestone_mentions(self) -> None:
|
|
session_text = (
|
|
"# ai-exchange Session\n\n"
|
|
"- current_phase: Restricted Decision-Support Soft Rollout\n"
|
|
f"### {run_until_blocked.ANALYSIS_DECISION_MONITOR_WITH_MATURITY_SLICE}\n"
|
|
)
|
|
tasks_text = (
|
|
"# ai-exchange Tasks\n\n"
|
|
"## In Progress\n\n"
|
|
f"- [ ] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_WITH_MATURITY_SLICE}`\n\n"
|
|
"## Major Done Milestones\n\n"
|
|
f"- [x] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE}`\n"
|
|
f"- [x] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_REMOTE_DRIFT_SLICE}`\n"
|
|
f"- [x] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_QUALITY_REVIEW_SLICE}`\n"
|
|
f"- [x] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_SLICE}`\n"
|
|
)
|
|
|
|
self.assertFalse(
|
|
run_until_blocked.should_use_analysis_decision_monitor_p7_selector(
|
|
session_text=session_text,
|
|
tasks_text=tasks_text,
|
|
)
|
|
)
|
|
self.assertFalse(
|
|
run_until_blocked.should_use_analysis_decision_monitor_maturity_rollup(
|
|
session_text=session_text,
|
|
tasks_text=tasks_text,
|
|
)
|
|
)
|
|
self.assertFalse(
|
|
run_until_blocked.should_use_analysis_decision_monitor_remote_drift_check(
|
|
session_text=session_text,
|
|
tasks_text=tasks_text,
|
|
)
|
|
)
|
|
self.assertFalse(
|
|
run_until_blocked.should_use_analysis_decision_support_quality_review(
|
|
session_text=session_text,
|
|
tasks_text=tasks_text,
|
|
)
|
|
)
|
|
self.assertTrue(
|
|
run_until_blocked.should_use_analysis_decision_monitor(
|
|
session_text=session_text,
|
|
tasks_text=tasks_text,
|
|
)
|
|
)
|
|
|
|
def test_p7_quality_review_closes_and_routes_to_p7_selector(self) -> None:
|
|
observed = {
|
|
"broker_status": "broker_state_ok",
|
|
"broker_fills": 0,
|
|
"runtime_health": "ok",
|
|
"market_gap": "market_data_gap_clear",
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
"decision_action": "restricted_decision_support",
|
|
"decision_operator_mode": "observe_only",
|
|
"decision_trust_classification": "blacklist",
|
|
}
|
|
monitor_payload = {
|
|
"passed": True,
|
|
"generated_ts": "2026-05-11T09:15:00+00:00",
|
|
"phase": "analysis_decision_only",
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": observed,
|
|
"checks": [
|
|
{"code": "decision_restricted_support", "passed": True},
|
|
{"code": "decision_observe_only", "passed": True},
|
|
{"code": "decision_execution_not_allowed", "passed": True},
|
|
{"code": "review_packet_present", "passed": True},
|
|
{"code": "review_trust_score_present", "passed": True},
|
|
{"code": "review_research_summary_present", "passed": True},
|
|
{"code": "real_trade_sample_gate_deferred", "passed": True},
|
|
],
|
|
}
|
|
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text(
|
|
"# ai-exchange Session\n\n"
|
|
"### automation_execution_layer_p7_alert_decision_support_quality_review\n",
|
|
encoding="utf-8",
|
|
)
|
|
tasks_path.write_text(
|
|
"# ai-exchange Tasks\n\n"
|
|
"## In Progress\n\n"
|
|
f"- [ ] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_QUALITY_REVIEW_SLICE}`\n",
|
|
encoding="utf-8",
|
|
)
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
original_monitor = run_until_blocked.run_analysis_decision_monitor.run_monitor
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
run_until_blocked.run_analysis_decision_monitor.run_monitor = lambda **kwargs: monitor_payload
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_OUTPUT_JSON": "0"})
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
status_payload = json.loads(status_path.read_text(encoding="utf-8"))
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
run_until_blocked.run_analysis_decision_monitor.run_monitor = original_monitor
|
|
|
|
self.assertEqual(summary["mode"], "analysis_decision_support_quality_review")
|
|
self.assertEqual(summary["status"], "completed")
|
|
self.assertEqual(summary["quality_review"]["quality_status"], "healthy")
|
|
self.assertEqual(summary["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_SLICE)
|
|
self.assertTrue(run_until_blocked.ANALYSIS_DECISION_SUPPORT_QUALITY_REVIEW_PATH.exists())
|
|
review_payload = json.loads(run_until_blocked.ANALYSIS_DECISION_SUPPORT_QUALITY_REVIEW_PATH.read_text(encoding="utf-8"))
|
|
self.assertEqual(review_payload["state"], "automation_execution_layer_p7_alert_decision_support_quality_review_closed")
|
|
self.assertIn("automation_execution_layer_p7_alert_decision_support_quality_review_closed", session_text)
|
|
self.assertIn(run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_SLICE, tasks_text)
|
|
self.assertEqual(status_payload["mode"], "analysis_decision_support_quality_review")
|
|
|
|
def test_p7_quality_review_blocks_when_required_research_summary_missing(self) -> None:
|
|
monitor_payload = {
|
|
"passed": True,
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": {
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
"decision_action": "restricted_decision_support",
|
|
"decision_operator_mode": "observe_only",
|
|
"decision_trust_classification": "blacklist",
|
|
},
|
|
"checks": [
|
|
{"code": "decision_restricted_support", "passed": True},
|
|
{"code": "decision_observe_only", "passed": True},
|
|
{"code": "decision_execution_not_allowed", "passed": True},
|
|
{"code": "review_packet_present", "passed": True},
|
|
{"code": "review_trust_score_present", "passed": True},
|
|
{"code": "real_trade_sample_gate_deferred", "passed": True},
|
|
],
|
|
}
|
|
|
|
review = run_until_blocked.build_analysis_decision_support_quality_review(
|
|
monitor_payload=monitor_payload,
|
|
)
|
|
|
|
self.assertEqual(review["status"], "blocked")
|
|
self.assertEqual(review["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_SLICE)
|
|
self.assertIn("required_quality_checks_missing", review["failure_reasons"])
|
|
self.assertIn("review_research_summary_present", review["missing_required_checks"])
|
|
|
|
def test_p7_selector_stage_closes_and_routes_to_maturity_rollup(self) -> None:
|
|
guards = ["execution_allowed"]
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_TREND_PATH.write_text(
|
|
json.dumps(
|
|
{
|
|
"record_count": 10,
|
|
"consecutive_pass_count": 10,
|
|
"stable_sample_threshold": 3,
|
|
"stable_monitoring_window": True,
|
|
"latest_monitor_passed": True,
|
|
"latest_blockers": [],
|
|
"guards_unchanged": guards,
|
|
}
|
|
),
|
|
encoding="utf-8",
|
|
)
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_QUALITY_GATE_PATH.write_text(
|
|
json.dumps(
|
|
{
|
|
"status": "passed",
|
|
"passed": True,
|
|
"stable_monitoring_window": True,
|
|
"failure_reasons": [],
|
|
"guards_unchanged": guards,
|
|
}
|
|
),
|
|
encoding="utf-8",
|
|
)
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_MATURITY_ROLLUP_PATH.write_text(
|
|
json.dumps(
|
|
{
|
|
"state": "automation_execution_layer_p7_monitor_maturity_rollup_closed",
|
|
"status": "completed",
|
|
"maturity_status": "mature",
|
|
"guards_unchanged": guards,
|
|
}
|
|
),
|
|
encoding="utf-8",
|
|
)
|
|
run_until_blocked.ANALYSIS_DECISION_MONITOR_REMOTE_DRIFT_CHECK_PATH.write_text(
|
|
json.dumps(
|
|
{
|
|
"state": "automation_execution_layer_p7_remote_drift_check_closed",
|
|
"status": "completed",
|
|
"drift_status": "clear",
|
|
"failure_reasons": [],
|
|
"guards_unchanged": guards,
|
|
}
|
|
),
|
|
encoding="utf-8",
|
|
)
|
|
run_until_blocked.ANALYSIS_DECISION_SUPPORT_QUALITY_REVIEW_PATH.write_text(
|
|
json.dumps(
|
|
{
|
|
"state": "automation_execution_layer_p7_alert_decision_support_quality_review_closed",
|
|
"status": "completed",
|
|
"quality_status": "healthy",
|
|
"failure_reasons": [],
|
|
"guards_unchanged": guards,
|
|
}
|
|
),
|
|
encoding="utf-8",
|
|
)
|
|
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text(
|
|
"# ai-exchange Session\n\n"
|
|
f"### {run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_SLICE}\n",
|
|
encoding="utf-8",
|
|
)
|
|
tasks_path.write_text(
|
|
"# ai-exchange Tasks\n\n"
|
|
"## In Progress\n\n"
|
|
f"- [ ] `{run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_SLICE}`\n",
|
|
encoding="utf-8",
|
|
)
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_OUTPUT_JSON": "0"})
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
status_payload = json.loads(status_path.read_text(encoding="utf-8"))
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
|
|
self.assertEqual(summary["mode"], "analysis_decision_monitor_p7_selector")
|
|
self.assertEqual(summary["status"], "completed")
|
|
self.assertEqual(summary["selector"]["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_WITH_MATURITY_SLICE)
|
|
self.assertTrue(run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_PATH.exists())
|
|
self.assertIn("automation_execution_layer_p7_bounded_automation_slice_selector_closed", session_text)
|
|
self.assertIn(run_until_blocked.ANALYSIS_DECISION_MONITOR_WITH_MATURITY_SLICE, tasks_text)
|
|
self.assertEqual(status_payload["mode"], "analysis_decision_monitor_p7_selector")
|
|
|
|
def test_p7_selector_chooses_maturity_rollup_when_all_inputs_are_healthy(self) -> None:
|
|
observed = {
|
|
"broker_status": "broker_state_ok",
|
|
"broker_fills": 0,
|
|
"runtime_health": "ok",
|
|
"market_gap": "market_data_gap_clear",
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
"decision_action": "restricted_decision_support",
|
|
"decision_operator_mode": "observe_only",
|
|
"decision_trust_classification": "blacklist",
|
|
"p6_governance_tick_status": "no_op_healthy",
|
|
}
|
|
trend = {
|
|
"record_count": 10,
|
|
"consecutive_pass_count": 10,
|
|
"stable_sample_threshold": 3,
|
|
"stable_monitoring_window": True,
|
|
"latest_monitor_passed": True,
|
|
"latest_monitor_generated_ts": "2026-05-11T10:13:03+00:00",
|
|
"latest_blockers": [],
|
|
"latest_warnings": [],
|
|
"latest_observed": observed,
|
|
"next_slice": run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE,
|
|
"guards_unchanged": [
|
|
"execution_allowed",
|
|
"real_live_ready",
|
|
"risk_governor",
|
|
"trust_bucket",
|
|
"operator_confirmation_protocol",
|
|
],
|
|
}
|
|
quality_gate = {
|
|
"status": "passed",
|
|
"passed": True,
|
|
"failure_reasons": [],
|
|
"record_count": 10,
|
|
"consecutive_pass_count": 10,
|
|
"stable_monitoring_window": True,
|
|
"latest_monitor_generated_ts": "2026-05-11T10:13:03+00:00",
|
|
"latest_observed": observed,
|
|
"next_slice": run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_NEXT_SLICE,
|
|
"guards_unchanged": trend["guards_unchanged"],
|
|
}
|
|
maturity_rollup = {
|
|
"state": "automation_execution_layer_p7_monitor_maturity_rollup_closed",
|
|
"status": "completed",
|
|
"maturity_status": "mature",
|
|
"local": {
|
|
"quality_gate_status": "passed",
|
|
"record_count": 10,
|
|
"consecutive_pass_count": 10,
|
|
"stable_monitoring_window": True,
|
|
"quality_gate_passed": True,
|
|
},
|
|
"latest_observed": observed,
|
|
"remote_drift_check_status": "clear",
|
|
"guards_unchanged": trend["guards_unchanged"],
|
|
}
|
|
remote_drift_check = {
|
|
"state": "automation_execution_layer_p7_remote_drift_check_closed",
|
|
"status": "completed",
|
|
"drift_status": "clear",
|
|
"local": {
|
|
"maturity_status": "mature",
|
|
"quality_gate_status": "passed",
|
|
"record_count": 10,
|
|
"consecutive_pass_count": 10,
|
|
"stable_monitoring_window": True,
|
|
},
|
|
"remote": {
|
|
"status": "available",
|
|
"host": "192.168.3.90",
|
|
"container_quality_gate_status": "passed",
|
|
"container_record_count": 12,
|
|
"container_consecutive_pass_count": 4,
|
|
"container_stable_monitoring_window": True,
|
|
"host_quality_gate_status": "passed",
|
|
"host_record_count": 12,
|
|
"host_consecutive_pass_count": 4,
|
|
"host_stable_monitoring_window": True,
|
|
},
|
|
"observed_diffs": {},
|
|
"latest_observed": {"local": observed, "remote": observed},
|
|
"guards_unchanged": trend["guards_unchanged"],
|
|
}
|
|
quality_review = {
|
|
"state": "automation_execution_layer_p7_alert_decision_support_quality_review_closed",
|
|
"status": "completed",
|
|
"quality_status": "healthy",
|
|
"current_phase": "Restricted Decision-Support Soft Rollout",
|
|
"current_slice": run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_QUALITY_REVIEW_SLICE,
|
|
"next_slice": run_until_blocked.ANALYSIS_DECISION_MONITOR_P7_SELECTOR_SLICE,
|
|
"failure_reasons": [],
|
|
"monitor_blockers": [],
|
|
"monitor_warnings": [],
|
|
"latest_observed": observed,
|
|
"operator_mode": "observe_only",
|
|
"decision_action": "restricted_decision_support",
|
|
"trust_classification": "blacklist",
|
|
"guards_unchanged": trend["guards_unchanged"],
|
|
}
|
|
|
|
selector = run_until_blocked.build_analysis_decision_monitor_p7_slice_selector(
|
|
trend=trend,
|
|
quality_gate=quality_gate,
|
|
maturity_rollup=maturity_rollup,
|
|
remote_drift_check=remote_drift_check,
|
|
quality_review=quality_review,
|
|
)
|
|
|
|
self.assertEqual(selector["status"], "completed")
|
|
self.assertEqual(selector["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_WITH_MATURITY_SLICE)
|
|
self.assertEqual(selector["selection_reason"], "all_monitoring_artifacts_healthy")
|
|
|
|
def test_p7_selector_routes_to_remote_drift_repair_when_drift_detected(self) -> None:
|
|
trend = {
|
|
"record_count": 10,
|
|
"consecutive_pass_count": 10,
|
|
"stable_sample_threshold": 3,
|
|
"stable_monitoring_window": True,
|
|
"latest_monitor_passed": True,
|
|
"latest_blockers": [],
|
|
"guards_unchanged": ["execution_allowed"],
|
|
}
|
|
quality_gate = {
|
|
"status": "passed",
|
|
"passed": True,
|
|
"stable_monitoring_window": True,
|
|
"failure_reasons": [],
|
|
"guards_unchanged": ["execution_allowed"],
|
|
}
|
|
maturity_rollup = {
|
|
"state": "automation_execution_layer_p7_monitor_maturity_rollup_closed",
|
|
"status": "completed",
|
|
"maturity_status": "mature",
|
|
"local": {"quality_gate_status": "passed"},
|
|
"guards_unchanged": ["execution_allowed"],
|
|
}
|
|
remote_drift_check = {
|
|
"state": "automation_execution_layer_p7_remote_drift_check_blocked",
|
|
"status": "blocked",
|
|
"drift_status": "drift_detected",
|
|
"failure_reasons": ["local_remote_observed_fields_drifted"],
|
|
"guards_unchanged": ["execution_allowed"],
|
|
}
|
|
quality_review = {
|
|
"state": "automation_execution_layer_p7_alert_decision_support_quality_review_closed",
|
|
"status": "completed",
|
|
"quality_status": "healthy",
|
|
"failure_reasons": [],
|
|
"guards_unchanged": ["execution_allowed"],
|
|
}
|
|
|
|
selector = run_until_blocked.build_analysis_decision_monitor_p7_slice_selector(
|
|
trend=trend,
|
|
quality_gate=quality_gate,
|
|
maturity_rollup=maturity_rollup,
|
|
remote_drift_check=remote_drift_check,
|
|
quality_review=quality_review,
|
|
)
|
|
|
|
self.assertEqual(selector["next_slice"], run_until_blocked.ANALYSIS_DECISION_MONITOR_REMOTE_DRIFT_REPAIR_SLICE)
|
|
self.assertEqual(selector["selection_reason"], "remote_drift_detected")
|
|
|
|
def test_analysis_decision_monitor_phase_blocks_on_monitor_blockers(self) -> None:
|
|
monitor_payload = {
|
|
"passed": False,
|
|
"generated_ts": "2026-05-11T04:46:00+00:00",
|
|
"phase": "analysis_decision_only",
|
|
"blockers": ["market_data_fresh"],
|
|
"warnings": [],
|
|
"observed": {
|
|
"real_trade_sample_gate_required": False,
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
},
|
|
}
|
|
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text(
|
|
"# ai-exchange Session\n\n"
|
|
"- state: analysis_decision_monitor_runner_deployed\n"
|
|
"- repair_next_action: run_analysis_decision_monitor_periodically_without_fill_gate\n",
|
|
encoding="utf-8",
|
|
)
|
|
tasks_path.write_text(
|
|
"# ai-exchange Tasks\n\n"
|
|
"- [ ] `run_analysis_decision_monitor_periodically_without_fill_gate`\n",
|
|
encoding="utf-8",
|
|
)
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
original_monitor = run_until_blocked.run_analysis_decision_monitor.run_monitor
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
run_until_blocked.run_analysis_decision_monitor.run_monitor = lambda **kwargs: monitor_payload
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={})
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
status_payload = json.loads(status_path.read_text(encoding="utf-8"))
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
run_until_blocked.run_analysis_decision_monitor.run_monitor = original_monitor
|
|
|
|
self.assertEqual(summary["status"], "blocked")
|
|
self.assertEqual(summary["next_slice"], "repair_analysis_decision_monitor_blockers")
|
|
self.assertEqual(summary["blocked_reason"], "analysis_decision_monitor_blockers:market_data_fresh")
|
|
self.assertIn("analysis_decision_monitor_blockers:market_data_fresh", tasks_text)
|
|
self.assertEqual(status_payload["state"], "blocked")
|
|
self.assertEqual(status_payload["blocked_reason"], "analysis_decision_monitor_blockers:market_data_fresh")
|
|
self.assertEqual(summary["monitor_history"]["record_count"], 1)
|
|
|
|
def test_analysis_decision_monitor_trend_marks_stable_window_after_three_passes(self) -> None:
|
|
records = [
|
|
{
|
|
"recorded_ts": "2026-05-11T00:00:00+00:00",
|
|
"monitor_generated_ts": "2026-05-11T00:00:00+00:00",
|
|
"status": "completed",
|
|
"monitor_passed": True,
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": {"broker_status": "broker_state_ok"},
|
|
"next_slice": "continue_periodic_analysis_decision_monitoring",
|
|
"guards_unchanged": ["execution_allowed"],
|
|
},
|
|
{
|
|
"recorded_ts": "2026-05-11T00:05:00+00:00",
|
|
"monitor_generated_ts": "2026-05-11T00:05:00+00:00",
|
|
"status": "completed",
|
|
"monitor_passed": True,
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": {"broker_status": "broker_state_ok"},
|
|
"next_slice": "continue_periodic_analysis_decision_monitoring",
|
|
"guards_unchanged": ["execution_allowed"],
|
|
},
|
|
{
|
|
"recorded_ts": "2026-05-11T00:10:00+00:00",
|
|
"monitor_generated_ts": "2026-05-11T00:10:00+00:00",
|
|
"status": "completed",
|
|
"monitor_passed": True,
|
|
"blockers": [],
|
|
"warnings": [],
|
|
"observed": {"broker_status": "broker_state_ok"},
|
|
"next_slice": "continue_periodic_analysis_decision_monitoring",
|
|
"guards_unchanged": ["execution_allowed"],
|
|
},
|
|
]
|
|
|
|
trend = run_until_blocked.build_analysis_decision_monitor_trend(records=records)
|
|
|
|
self.assertEqual(trend["record_count"], 3)
|
|
self.assertEqual(trend["consecutive_pass_count"], 3)
|
|
self.assertTrue(trend["stable_monitoring_window"])
|
|
self.assertEqual(trend["latest_observed"]["broker_status"], "broker_state_ok")
|
|
|
|
def test_analysis_decision_monitor_quality_gate_passes_for_stable_observe_only_window(self) -> None:
|
|
trend = {
|
|
"record_count": 3,
|
|
"consecutive_pass_count": 3,
|
|
"stable_monitoring_window": True,
|
|
"latest_monitor_passed": True,
|
|
"latest_blockers": [],
|
|
"latest_warnings": [],
|
|
"latest_monitor_generated_ts": "2026-05-11T00:10:00+00:00",
|
|
"latest_observed": {
|
|
"decision_action": "restricted_decision_support",
|
|
"decision_operator_mode": "observe_only",
|
|
"real_trade_sample_gate_required": False,
|
|
"real_trade_sample_gate_status": "deferred_analysis_decision_only",
|
|
"runtime_health": "ok",
|
|
"market_gap": "market_data_gap_clear",
|
|
"p6_governance_tick_status": "no_op_healthy",
|
|
},
|
|
"next_slice": "continue_periodic_analysis_decision_monitoring",
|
|
"guards_unchanged": [
|
|
"execution_allowed",
|
|
"real_live_ready",
|
|
"risk_governor",
|
|
"trust_bucket",
|
|
"operator_confirmation_protocol",
|
|
],
|
|
}
|
|
|
|
gate = run_until_blocked.build_analysis_decision_monitor_quality_gate(trend=trend)
|
|
|
|
self.assertTrue(gate["passed"])
|
|
self.assertEqual(gate["status"], "passed")
|
|
self.assertEqual(gate["failure_reasons"], [])
|
|
self.assertIn("execution_gate_unchanged", gate["notes"])
|
|
|
|
def test_advances_to_next_stage_and_updates_anchors(self) -> None:
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
session_path.write_text("# ai-exchange Session\n\n### `P0-2` 真实账户与订单只读同步\n", encoding="utf-8")
|
|
tasks_path.write_text("# ai-exchange Tasks\n\n## Done\n\n- [x] `P0-1` 实时市场真值层\n", encoding="utf-8")
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_runner = run_until_blocked.run_trading_delivery_rehearsal
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.run_trading_delivery_rehearsal = lambda **kwargs: {
|
|
"mode": "rehearsal_completed",
|
|
"passed": True,
|
|
"continuation_decision": "advance",
|
|
"promotion_decision": "promote_candidate",
|
|
"blockers": [],
|
|
}
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_MAX_ROUNDS": "1"})
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.run_trading_delivery_rehearsal = original_runner
|
|
|
|
self.assertEqual(summary["status"], "stopped")
|
|
self.assertEqual(summary["completed_stages"], ["P0-1", "P0-2"])
|
|
self.assertEqual(summary["current_stage"], "P0-3")
|
|
self.assertIn("### `P0-3` 盘中持续调度与运行监督", session_text)
|
|
self.assertIn("- [x] `P0-2` 真实账户与订单只读同步", tasks_text)
|
|
self.assertIn("- [ ] `P0-3` 盘中持续调度与运行监督", tasks_text)
|
|
|
|
def test_stops_when_rehearsal_reports_blocked(self) -> None:
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
session_path.write_text("# ai-exchange Session\n\n### `P0-2` 真实账户与订单只读同步\n", encoding="utf-8")
|
|
tasks_path.write_text("# ai-exchange Tasks\n\n## Done\n\n", encoding="utf-8")
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_runner = run_until_blocked.run_trading_delivery_rehearsal
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.run_trading_delivery_rehearsal = lambda **kwargs: {
|
|
"mode": "autonomous_cycle_failed",
|
|
"passed": False,
|
|
"continuation_decision": "blocked",
|
|
"promotion_decision": "hold",
|
|
"blockers": ["缺凭据:broker_readonly_api_key"],
|
|
}
|
|
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_MAX_ROUNDS": "2"})
|
|
tasks_text = tasks_path.read_text(encoding="utf-8")
|
|
session_text = session_path.read_text(encoding="utf-8")
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.run_trading_delivery_rehearsal = original_runner
|
|
|
|
self.assertEqual(summary["status"], "blocked")
|
|
self.assertEqual(summary["blocked_reason"], "缺凭据:broker_readonly_api_key")
|
|
self.assertEqual(summary["current_stage"], "P0-2")
|
|
self.assertIn("缺凭据:broker_readonly_api_key", tasks_text)
|
|
self.assertIn("当前自动化已阻塞：缺凭据:broker_readonly_api_key", session_text)
|
|
|
|
def test_json_output_is_machine_readable(self) -> None:
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
session_path.write_text("# ai-exchange Session\n\n### `P0-2` 真实账户与订单只读同步\n", encoding="utf-8")
|
|
tasks_path.write_text("# ai-exchange Tasks\n\n## Done\n\n", encoding="utf-8")
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_runner = run_until_blocked.run_trading_delivery_rehearsal
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.run_trading_delivery_rehearsal = lambda **kwargs: {
|
|
"mode": "autonomous_cycle_failed",
|
|
"passed": False,
|
|
"continuation_decision": "blocked",
|
|
"promotion_decision": "hold",
|
|
"blockers": ["缺凭据:broker_readonly_api_key"],
|
|
}
|
|
output = io.StringIO()
|
|
with contextlib.redirect_stdout(output):
|
|
summary = run_until_blocked.run_until_blocked(env={"RUN_UNTIL_BLOCKED_OUTPUT_JSON": "1"})
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.run_trading_delivery_rehearsal = original_runner
|
|
|
|
payload = json.loads(output.getvalue())
|
|
self.assertEqual(payload["status"], "blocked")
|
|
self.assertEqual(payload["blocked_reason"], "缺凭据:broker_readonly_api_key")
|
|
self.assertEqual(summary["status"], "blocked")
|
|
|
|
def test_run_until_blocked_writes_runner_status_file(self) -> None:
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
session_path = Path(tmp) / "SESSION.md"
|
|
tasks_path = Path(tmp) / "TASKS.md"
|
|
status_path = Path(tmp) / "automation_status.json"
|
|
session_path.write_text("# ai-exchange Session\n\n### `P0-2` 真实账户与订单只读同步\n", encoding="utf-8")
|
|
tasks_path.write_text("# ai-exchange Tasks\n\n## Done\n\n", encoding="utf-8")
|
|
|
|
original_session = run_until_blocked.SESSION_PATH
|
|
original_tasks = run_until_blocked.TASKS_PATH
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
original_runner = run_until_blocked.run_trading_delivery_rehearsal
|
|
try:
|
|
run_until_blocked.SESSION_PATH = session_path
|
|
run_until_blocked.TASKS_PATH = tasks_path
|
|
run_until_blocked.RUNNER_STATUS_PATH = status_path
|
|
run_until_blocked.run_trading_delivery_rehearsal = lambda **kwargs: {
|
|
"mode": "autonomous_cycle_failed",
|
|
"passed": False,
|
|
"continuation_decision": "blocked",
|
|
"promotion_decision": "hold",
|
|
"blockers": ["缺凭据:broker_readonly_api_key"],
|
|
}
|
|
run_until_blocked.run_until_blocked(env={})
|
|
payload = json.loads(status_path.read_text(encoding="utf-8"))
|
|
finally:
|
|
run_until_blocked.SESSION_PATH = original_session
|
|
run_until_blocked.TASKS_PATH = original_tasks
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
run_until_blocked.run_trading_delivery_rehearsal = original_runner
|
|
|
|
self.assertEqual(payload["state"], "blocked")
|
|
self.assertEqual(payload["current_stage"], "P0-2")
|
|
self.assertEqual(payload["blocked_reason"], "缺凭据:broker_readonly_api_key")
|
|
|
|
def test_supervisor_pauses_on_code_fix_blocker(self) -> None:
|
|
sleeper = Mock()
|
|
original_run_until_blocked = run_until_blocked.run_until_blocked
|
|
original_fingerprint = run_until_blocked.compute_resume_fingerprint
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
try:
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
run_until_blocked.RUNNER_STATUS_PATH = Path(tmp) / "automation_status.json"
|
|
run_until_blocked.run_until_blocked = lambda **kwargs: {
|
|
"status": "blocked",
|
|
"started_ts": "2026-05-07T00:00:00+00:00",
|
|
"current_stage": "P0-2",
|
|
"completed_stages": ["P0-1"],
|
|
"blocked_reason": "autonomous_mode:pipeline_failed",
|
|
"rounds": [{"round": 1}],
|
|
}
|
|
run_until_blocked.compute_resume_fingerprint = lambda: "same"
|
|
result = run_until_blocked.run_supervised_until_stopped(
|
|
env={"RUN_UNTIL_BLOCKED_LOOP_SLEEP_SECONDS": "1", "RUN_UNTIL_BLOCKED_SUPERVISOR_CYCLES": "2"},
|
|
sleeper=sleeper,
|
|
)
|
|
finally:
|
|
run_until_blocked.run_until_blocked = original_run_until_blocked
|
|
run_until_blocked.compute_resume_fingerprint = original_fingerprint
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
|
|
sleeper.assert_called_once_with(1)
|
|
self.assertEqual(result["status"], "blocked")
|
|
|
|
def test_supervisor_resumes_after_fingerprint_change(self) -> None:
|
|
summaries = [
|
|
{
|
|
"status": "blocked",
|
|
"started_ts": "2026-05-07T00:00:00+00:00",
|
|
"current_stage": "P0-2",
|
|
"completed_stages": ["P0-1"],
|
|
"blocked_reason": "autonomous_mode:pipeline_failed",
|
|
"rounds": [{"round": 1}],
|
|
},
|
|
{
|
|
"status": "completed",
|
|
"started_ts": "2026-05-07T00:10:00+00:00",
|
|
"current_stage": None,
|
|
"completed_stages": ["P0-1", "P0-2"],
|
|
"blocked_reason": None,
|
|
"rounds": [{"round": 1}],
|
|
},
|
|
]
|
|
fingerprints = ["a", "a", "b", "b"]
|
|
sleeper = Mock()
|
|
original_run_until_blocked = run_until_blocked.run_until_blocked
|
|
original_fingerprint = run_until_blocked.compute_resume_fingerprint
|
|
original_status = run_until_blocked.RUNNER_STATUS_PATH
|
|
try:
|
|
with tempfile.TemporaryDirectory() as tmp:
|
|
run_until_blocked.RUNNER_STATUS_PATH = Path(tmp) / "automation_status.json"
|
|
run_until_blocked.run_until_blocked = lambda **kwargs: summaries.pop(0)
|
|
run_until_blocked.compute_resume_fingerprint = lambda: fingerprints.pop(0)
|
|
result = run_until_blocked.run_supervised_until_stopped(
|
|
env={"RUN_UNTIL_BLOCKED_LOOP_SLEEP_SECONDS": "1", "RUN_UNTIL_BLOCKED_SUPERVISOR_CYCLES": "2"},
|
|
sleeper=sleeper,
|
|
)
|
|
finally:
|
|
run_until_blocked.run_until_blocked = original_run_until_blocked
|
|
run_until_blocked.compute_resume_fingerprint = original_fingerprint
|
|
run_until_blocked.RUNNER_STATUS_PATH = original_status
|
|
|
|
sleeper.assert_called_once_with(1)
|
|
self.assertEqual(result["status"], "completed")
|
|
self.assertEqual(result["completed_stages"], ["P0-1", "P0-2"])
|
|
|
|
def test_main_uses_supervisor_mode_when_enabled(self) -> None:
|
|
original_run_until_blocked = run_until_blocked.run_until_blocked
|
|
original_supervisor = run_until_blocked.run_supervised_until_stopped
|
|
original_environ = dict(os.environ)
|
|
try:
|
|
os.environ["RUN_UNTIL_BLOCKED_SUPERVISOR"] = "1"
|
|
run_until_blocked.run_until_blocked = lambda **kwargs: (_ for _ in ()).throw(AssertionError("should not call one-shot runner"))
|
|
run_until_blocked.run_supervised_until_stopped = lambda **kwargs: {"status": "completed"}
|
|
exit_code = run_until_blocked.main()
|
|
finally:
|
|
os.environ.clear()
|
|
os.environ.update(original_environ)
|
|
run_until_blocked.run_until_blocked = original_run_until_blocked
|
|
run_until_blocked.run_supervised_until_stopped = original_supervisor
|
|
|
|
self.assertEqual(exit_code, 0)
|
|
|
|
if __name__ == "__main__":
|
|
unittest.main()
|