938 lines
32 KiB
Python
938 lines
32 KiB
Python
"""
|
||
Governance Query Service — /governance 頁面 DB 查詢邏輯
|
||
======================================================
|
||
封裝 3 個 governance endpoint 的資料庫查詢。
|
||
Router 層禁直接存取 DB（leWOOOgo 積木化鐵律）。
|
||
|
||
函式清單:
|
||
query_governance_events(...) → GovernanceEventsResponse
|
||
query_governance_queue(...) → GovernanceQueueResponse
|
||
query_governance_summary(...) → GovernanceSummaryResponse
|
||
|
||
Graceful fallback 規則:
|
||
queue endpoint — governance_remediation_dispatch 表可能尚未建立（Track D 進行中）。
|
||
捕捉 sqlalchemy.exc.ProgrammingError（表不存在）後回傳 table_pending=True 的空列表，
|
||
確保 API 在表建立前不拋 500。
|
||
|
||
2026-05-02 ogt + Claude Sonnet 4.6 Asia/Taipei
|
||
"""
|
||
|
||
from __future__ import annotations
|
||
|
||
from datetime import datetime, timedelta, timezone
|
||
from typing import Any
|
||
|
||
import structlog
|
||
from sqlalchemy import String, bindparam, func, or_, select, text
|
||
from sqlalchemy.exc import ProgrammingError
|
||
|
||
from src.db.base import get_db_context
|
||
from src.db.models import AiGovernanceEvent, KnowledgeEntryRecord
|
||
from src.models.governance import (
|
||
DailyCount,
|
||
DispatchItem,
|
||
GovernanceEvent,
|
||
GovernanceEventsResponse,
|
||
GovernanceQueueResponse,
|
||
GovernanceSummaryResponse,
|
||
KnowledgeReviewDraftDedupeGroup,
|
||
KnowledgeReviewDraftDedupeResponse,
|
||
map_severity,
|
||
)
|
||
from src.models.knowledge import EntryStatus, EntryType
|
||
from src.utils.timezone import now_taipei
|
||
|
||
logger = structlog.get_logger(__name__)
|
||
|
||
# =============================================================================
|
||
# 常數
|
||
# =============================================================================
|
||
|
||
_TAIPEI = timezone(timedelta(hours=8))
|
||
|
||
|
||
# =============================================================================
|
||
# helpers
|
||
# =============================================================================
|
||
|
||
def _extract_impact(details: dict) -> str:
|
||
"""
|
||
從 details 抽摘要字串，≤80 字。
|
||
|
||
優先讀 details["impact"]（dict），取 status + 主要 metric 欄位。
|
||
fallback 到 details 頂層常見欄位。
|
||
"""
|
||
impact_block = details.get("impact")
|
||
if isinstance(impact_block, dict):
|
||
parts: list[str] = []
|
||
if "status" in impact_block:
|
||
parts.append(str(impact_block["status"]))
|
||
# 主要 metric 欄位優先順序
|
||
for key in ("metric", "value", "rate", "ratio", "score", "count"):
|
||
if key in impact_block:
|
||
parts.append(f"{key}={impact_block[key]}")
|
||
break
|
||
summary = " ".join(parts)
|
||
return summary[:80] if summary else ""
|
||
|
||
# fallback: 頂層常見欄位
|
||
for key in ("message", "reason", "summary", "description"):
|
||
val = details.get(key)
|
||
if isinstance(val, str) and val:
|
||
return val[:80]
|
||
|
||
# 最後 fallback: 把 details 第一個 string value 截取
|
||
for val in details.values():
|
||
if isinstance(val, str) and val:
|
||
return val[:80]
|
||
|
||
return ""
|
||
|
||
|
||
def _extract_remediation(details: dict) -> str | None:
|
||
"""
|
||
將治理事件 details.remediation 正規化為前端可顯示的短字串。
|
||
|
||
Production 事件已出現 dict 形態（例如 {"items": [...]}），API response
|
||
schema 則是字串。這裡做 read-side normalization，避免歷史資料讓
|
||
/governance events 變成 500。
|
||
"""
|
||
remediation = details.get("remediation")
|
||
if remediation is None:
|
||
return None
|
||
if isinstance(remediation, str):
|
||
return remediation[:160]
|
||
if isinstance(remediation, dict):
|
||
for key in ("summary", "message", "reason", "action"):
|
||
value = remediation.get(key)
|
||
if isinstance(value, str) and value:
|
||
return value[:160]
|
||
items = remediation.get("items")
|
||
if isinstance(items, list):
|
||
normalized = [str(item) for item in items if item is not None]
|
||
if normalized:
|
||
return "；".join(normalized[:3])[:160]
|
||
return str(remediation)[:160]
|
||
if isinstance(remediation, list):
|
||
normalized = [str(item) for item in remediation if item is not None]
|
||
return "；".join(normalized[:3])[:160] if normalized else None
|
||
return str(remediation)[:160]
|
||
|
||
|
||
def _to_governance_event(
|
||
row: AiGovernanceEvent,
|
||
*,
|
||
dispatch_ids: list[str] | None = None,
|
||
) -> GovernanceEvent:
|
||
details = row.details if isinstance(row.details, dict) else {}
|
||
return GovernanceEvent(
|
||
id=row.id,
|
||
event_type=row.event_type,
|
||
severity=map_severity(row.event_type),
|
||
triggered_at=row.triggered_at,
|
||
resolved=row.resolved,
|
||
resolved_at=row.resolved_at,
|
||
impact=_extract_impact(details),
|
||
details=details,
|
||
remediation=_extract_remediation(details),
|
||
dispatch_ids=_merge_dispatch_ids(dispatch_ids or [], details.get("dispatch_ids")),
|
||
)
|
||
|
||
|
||
def _merge_dispatch_ids(
|
||
db_dispatch_ids: list[str],
|
||
legacy_dispatch_ids: Any,
|
||
) -> list[str]:
|
||
"""合併 DB dispatch trail 與 legacy payload ids，DB truth-first。"""
|
||
merged: list[str] = []
|
||
for raw in [*db_dispatch_ids, *(legacy_dispatch_ids if isinstance(legacy_dispatch_ids, list) else [])]:
|
||
if raw is None:
|
||
continue
|
||
value = str(raw)
|
||
if value and value not in merged:
|
||
merged.append(value)
|
||
return merged
|
||
|
||
|
||
async def _load_dispatch_ids_for_events(event_ids: list[str]) -> dict[str, list[str]]:
|
||
"""從 governance_remediation_dispatch 讀取事件對應 dispatch ids。
|
||
|
||
events endpoint 必須能在 dispatch 表尚未建立的環境 graceful fallback，
|
||
因此這裡捕捉 ProgrammingError 並回空 dict。
|
||
"""
|
||
if not event_ids:
|
||
return {}
|
||
|
||
sql = text("""
|
||
SELECT
|
||
d.governance_event_id,
|
||
d.id
|
||
FROM governance_remediation_dispatch d
|
||
WHERE d.governance_event_id IN :event_ids
|
||
ORDER BY d.dispatched_at DESC
|
||
""").bindparams(bindparam("event_ids", expanding=True))
|
||
|
||
try:
|
||
async with get_db_context() as db:
|
||
result = await db.execute(sql, {"event_ids": event_ids})
|
||
rows = result.fetchall()
|
||
except ProgrammingError as exc:
|
||
logger.warning(
|
||
"governance_dispatch_ids_table_not_ready",
|
||
error=str(exc),
|
||
)
|
||
return {}
|
||
|
||
dispatch_ids_by_event: dict[str, list[str]] = {}
|
||
for row in rows:
|
||
dispatch_ids_by_event.setdefault(str(row.governance_event_id), []).append(str(row.id))
|
||
return dispatch_ids_by_event
|
||
|
||
|
||
# =============================================================================
|
||
# Endpoint 1: events
|
||
# =============================================================================
|
||
|
||
async def query_governance_events(
|
||
*,
|
||
event_ids: list[str] | None = None,
|
||
event_types: list[str] | None = None,
|
||
from_dt: datetime | None = None,
|
||
to_dt: datetime | None = None,
|
||
status: str | None = None, # "resolved" | "unresolved"
|
||
severity: str | None = None, # "critical" | "warning" | "info"
|
||
page: int = 1,
|
||
size: int = 20,
|
||
) -> GovernanceEventsResponse:
|
||
"""
|
||
查詢 ai_governance_events 表，支援多維度過濾與分頁。
|
||
|
||
severity 過濾在 Python 層完成（event_type 映射）；
|
||
其他過濾在 SQL 層完成（效能優先）。
|
||
"""
|
||
async with get_db_context() as db:
|
||
stmt = select(AiGovernanceEvent)
|
||
|
||
normalized_event_ids = [
|
||
event_id.strip()
|
||
for event_id in (event_ids or [])
|
||
if isinstance(event_id, str) and event_id.strip()
|
||
]
|
||
if normalized_event_ids:
|
||
stmt = stmt.where(AiGovernanceEvent.id.in_(normalized_event_ids))
|
||
|
||
if event_types:
|
||
stmt = stmt.where(AiGovernanceEvent.event_type.in_(event_types))
|
||
|
||
if from_dt is not None:
|
||
stmt = stmt.where(AiGovernanceEvent.triggered_at >= from_dt)
|
||
|
||
if to_dt is not None:
|
||
stmt = stmt.where(AiGovernanceEvent.triggered_at <= to_dt)
|
||
|
||
if status == "resolved":
|
||
stmt = stmt.where(AiGovernanceEvent.resolved.is_(True))
|
||
elif status == "unresolved":
|
||
stmt = stmt.where(AiGovernanceEvent.resolved.is_(False))
|
||
|
||
stmt = stmt.order_by(AiGovernanceEvent.triggered_at.desc())
|
||
|
||
# 取全部結果，severity 在 Python 層過濾（避免 DB 不認識 mapping 邏輯）
|
||
result = await db.execute(stmt)
|
||
all_rows = result.scalars().all()
|
||
|
||
event_rows_by_id = {str(r.id): r for r in all_rows}
|
||
events = [_to_governance_event(r) for r in all_rows]
|
||
|
||
# severity 過濾（Python 層）
|
||
if severity:
|
||
from src.models.governance import _CRITICAL_TYPES, _WARNING_TYPES
|
||
|
||
if severity == "critical":
|
||
events = [e for e in events if e.event_type in _CRITICAL_TYPES]
|
||
elif severity == "warning":
|
||
events = [e for e in events if e.event_type in _WARNING_TYPES]
|
||
elif severity == "info":
|
||
events = [
|
||
e for e in events
|
||
if e.event_type not in _CRITICAL_TYPES and e.event_type not in _WARNING_TYPES
|
||
]
|
||
|
||
total = len(events)
|
||
offset = (page - 1) * size
|
||
page_items = events[offset: offset + size]
|
||
dispatch_ids_by_event = await _load_dispatch_ids_for_events([e.id for e in page_items])
|
||
if dispatch_ids_by_event:
|
||
page_items = [
|
||
_to_governance_event(
|
||
event_rows_by_id[item.id],
|
||
dispatch_ids=dispatch_ids_by_event.get(item.id, []),
|
||
)
|
||
for item in page_items
|
||
]
|
||
|
||
return GovernanceEventsResponse(
|
||
items=page_items,
|
||
total=total,
|
||
page=page,
|
||
size=size,
|
||
)
|
||
|
||
|
||
# =============================================================================
|
||
# Endpoint 2: queue
|
||
# =============================================================================
|
||
|
||
async def query_governance_queue(
|
||
*,
|
||
dispatch_status: str = "pending",
|
||
event_types: list[str] | None = None,
|
||
page: int = 1,
|
||
size: int = 20,
|
||
) -> GovernanceQueueResponse:
|
||
"""
|
||
查詢 governance_remediation_dispatch 表。
|
||
|
||
Track D 進行中，表可能尚未建立。
|
||
捕捉 ProgrammingError → 回傳 table_pending=True 的空 response。
|
||
|
||
proposed_action 從 decision_context JSONB 抽取（Track D 完成後可改為真實 join）。
|
||
"""
|
||
try:
|
||
return await _query_dispatch_table(
|
||
dispatch_status=dispatch_status,
|
||
event_types=event_types,
|
||
page=page,
|
||
size=size,
|
||
)
|
||
except ProgrammingError as exc:
|
||
logger.warning(
|
||
"governance_dispatch_table_not_ready",
|
||
error=str(exc),
|
||
)
|
||
return GovernanceQueueResponse(
|
||
items=[],
|
||
total=0,
|
||
page=page,
|
||
size=size,
|
||
table_pending=True,
|
||
)
|
||
except ImportError as exc:
|
||
logger.warning(
|
||
"governance_dispatch_model_not_ready",
|
||
error=str(exc),
|
||
)
|
||
return GovernanceQueueResponse(
|
||
items=[],
|
||
total=0,
|
||
page=page,
|
||
size=size,
|
||
table_pending=True,
|
||
)
|
||
|
||
|
||
async def _query_dispatch_table(
|
||
*,
|
||
dispatch_status: str,
|
||
event_types: list[str] | None,
|
||
page: int,
|
||
size: int,
|
||
) -> GovernanceQueueResponse:
|
||
"""實際查詢 governance_remediation_dispatch 表（不含 graceful fallback）."""
|
||
# 動態 import：Track D 完成前 ORM class 可能不存在
|
||
# 使用 raw SQL 降低 ORM 模型缺失的耦合風險
|
||
status_filter = (
|
||
"TRUE"
|
||
if dispatch_status == "all"
|
||
else "d.dispatch_status = CAST(:dispatch_status AS governance_dispatch_status)"
|
||
)
|
||
event_type_filter = (
|
||
"TRUE"
|
||
if not event_types
|
||
else "e.event_type::text IN :event_types"
|
||
)
|
||
params: dict[str, Any] = {}
|
||
if dispatch_status != "all":
|
||
params["dispatch_status"] = dispatch_status
|
||
if event_types:
|
||
params["event_types"] = event_types
|
||
|
||
sql = text(f"""
|
||
SELECT
|
||
d.id,
|
||
d.governance_event_id,
|
||
e.event_type,
|
||
d.dispatch_status,
|
||
d.executor_type,
|
||
d.decision_context,
|
||
d.playbook_id,
|
||
d.dispatched_at AS created_at,
|
||
d.dispatched_at,
|
||
d.started_at,
|
||
d.completed_at,
|
||
NULL::text AS operator_note
|
||
FROM governance_remediation_dispatch d
|
||
JOIN ai_governance_events e ON e.id = d.governance_event_id
|
||
WHERE {status_filter}
|
||
AND {event_type_filter}
|
||
ORDER BY d.dispatched_at DESC
|
||
""")
|
||
|
||
count_sql = text(f"""
|
||
SELECT count(*) AS cnt
|
||
FROM governance_remediation_dispatch d
|
||
JOIN ai_governance_events e ON e.id = d.governance_event_id
|
||
WHERE {status_filter}
|
||
AND {event_type_filter}
|
||
""")
|
||
|
||
if event_types:
|
||
sql = sql.bindparams(bindparam("event_types", expanding=True))
|
||
count_sql = count_sql.bindparams(bindparam("event_types", expanding=True))
|
||
|
||
async with get_db_context() as db:
|
||
count_row = await db.execute(count_sql, params)
|
||
total = int(count_row.scalar_one_or_none() or 0)
|
||
|
||
rows = await db.execute(sql, params)
|
||
all_rows = rows.fetchall()
|
||
|
||
offset = (page - 1) * size
|
||
page_rows = all_rows[offset: offset + size]
|
||
|
||
items: list[DispatchItem] = []
|
||
for row in page_rows:
|
||
decision_ctx: dict = (row.decision_context or {}) if hasattr(row, "decision_context") else {}
|
||
items.append(_to_dispatch_item(row, decision_ctx))
|
||
|
||
return GovernanceQueueResponse(
|
||
items=items,
|
||
total=total,
|
||
page=page,
|
||
size=size,
|
||
table_pending=False,
|
||
)
|
||
|
||
|
||
def _to_dispatch_item(row: Any, decision_ctx: dict) -> DispatchItem:
|
||
"""把 governance_remediation_dispatch SQL row 轉成 Work Items read model。"""
|
||
proposed_action = _extract_proposed_action(decision_ctx)
|
||
|
||
# playbook_trust: Track D 完成後改為 JOIN playbooks 表取 trust_score
|
||
# 現階段從 decision_context 取 mock 值
|
||
playbook_trust_raw = decision_ctx.get("playbook_trust")
|
||
try:
|
||
playbook_trust = float(playbook_trust_raw) if playbook_trust_raw is not None else None
|
||
except (TypeError, ValueError):
|
||
playbook_trust = None
|
||
|
||
return DispatchItem(
|
||
id=str(row.id),
|
||
governance_event_id=str(row.governance_event_id),
|
||
event_type=str(row.event_type),
|
||
dispatch_status=str(row.dispatch_status),
|
||
executor_type=str(row.executor_type) if row.executor_type else None,
|
||
proposed_action=proposed_action,
|
||
playbook_id=str(row.playbook_id) if row.playbook_id else None,
|
||
playbook_trust=playbook_trust,
|
||
created_at=row.created_at,
|
||
dispatched_at=row.dispatched_at,
|
||
started_at=row.started_at,
|
||
completed_at=row.completed_at,
|
||
operator_note=row.operator_note,
|
||
decision_path=_extract_decision_path(decision_ctx),
|
||
workflow_stage=_extract_workflow_stage(decision_ctx, str(row.dispatch_status)),
|
||
workflow_steps=_extract_workflow_steps(decision_ctx),
|
||
next_action=_extract_next_action(decision_ctx),
|
||
lead_agent=_extract_lead_agent(decision_ctx),
|
||
support_agents=_extract_support_agents(decision_ctx),
|
||
human_owner=_extract_human_owner(decision_ctx),
|
||
kb_draft_entry_id=_extract_kb_draft_entry_id(decision_ctx),
|
||
worker_status=_extract_worker_status(decision_ctx),
|
||
dry_run_plan_fingerprint=_extract_dry_run_plan_fingerprint(decision_ctx),
|
||
archived_count=_extract_archived_count(decision_ctx),
|
||
stale_ratio_snapshot=_extract_stale_ratio_snapshot(decision_ctx),
|
||
)
|
||
|
||
|
||
def _extract_proposed_action(decision_ctx: dict) -> str:
|
||
"""
|
||
從 decision_context JSONB 抽取 proposed_action，≤120 字。
|
||
|
||
Track D 完成後此函式可改為從真實欄位讀取。
|
||
"""
|
||
for key in (
|
||
"proposed_action",
|
||
"suggested_action",
|
||
"next_action",
|
||
"action",
|
||
"suggestion",
|
||
"description",
|
||
"summary",
|
||
):
|
||
val = decision_ctx.get(key)
|
||
if isinstance(val, str) and val:
|
||
return val[:120]
|
||
return "（待補充）"
|
||
|
||
|
||
def _extract_decision_path(decision_ctx: dict) -> str | None:
|
||
val = decision_ctx.get("decision_path")
|
||
return val[:80] if isinstance(val, str) and val else None
|
||
|
||
|
||
def _extract_next_action(decision_ctx: dict) -> str | None:
|
||
for key in ("next_action", "suggested_action", "proposed_action"):
|
||
val = decision_ctx.get(key)
|
||
if isinstance(val, str) and val:
|
||
return val[:120]
|
||
workflow = decision_ctx.get("workflow")
|
||
if isinstance(workflow, dict):
|
||
val = workflow.get("next_action")
|
||
if isinstance(val, str) and val:
|
||
return val[:120]
|
||
return None
|
||
|
||
|
||
def _extract_workflow_stage(decision_ctx: dict, dispatch_status: str) -> str | None:
|
||
workflow = decision_ctx.get("workflow")
|
||
if isinstance(workflow, dict):
|
||
stages = workflow.get("stage_by_dispatch_status")
|
||
if isinstance(stages, dict):
|
||
stage = stages.get(dispatch_status)
|
||
if isinstance(stage, str) and stage:
|
||
return stage[:120]
|
||
current = workflow.get("current_stage")
|
||
if isinstance(current, str) and current:
|
||
return current[:120]
|
||
|
||
return {
|
||
"pending": "queued_for_review",
|
||
"dispatched": "dispatched",
|
||
"executing": "executing",
|
||
"succeeded": "completed",
|
||
"failed": "failed",
|
||
"skipped": "skipped",
|
||
"cancelled": "cancelled",
|
||
}.get(dispatch_status)
|
||
|
||
|
||
def _extract_workflow_steps(decision_ctx: dict) -> list[str]:
|
||
workflow = decision_ctx.get("workflow")
|
||
if not isinstance(workflow, dict):
|
||
return []
|
||
steps = workflow.get("steps")
|
||
if not isinstance(steps, list):
|
||
return []
|
||
return [str(step)[:120] for step in steps if step is not None][:8]
|
||
|
||
|
||
def _extract_ownership(decision_ctx: dict) -> dict:
|
||
ownership = decision_ctx.get("ownership")
|
||
if isinstance(ownership, dict):
|
||
return ownership
|
||
extra = decision_ctx.get("extra")
|
||
if isinstance(extra, dict) and isinstance(extra.get("ownership"), dict):
|
||
return extra["ownership"]
|
||
return {}
|
||
|
||
|
||
def _extract_lead_agent(decision_ctx: dict) -> str | None:
|
||
val = _extract_ownership(decision_ctx).get("lead_agent")
|
||
return val[:80] if isinstance(val, str) and val else None
|
||
|
||
|
||
def _extract_support_agents(decision_ctx: dict) -> list[str]:
|
||
raw = _extract_ownership(decision_ctx).get("support_agents")
|
||
if not isinstance(raw, list):
|
||
return []
|
||
return [str(item)[:160] for item in raw if item is not None][:6]
|
||
|
||
|
||
def _extract_human_owner(decision_ctx: dict) -> str | None:
|
||
val = _extract_ownership(decision_ctx).get("human_owner")
|
||
return val[:120] if isinstance(val, str) and val else None
|
||
|
||
|
||
def _extract_kb_draft_entry_id(decision_ctx: dict) -> str | None:
|
||
"""Expose Hermes KM review draft id for Work Items owner review."""
|
||
workflow = decision_ctx.get("workflow")
|
||
if isinstance(workflow, dict):
|
||
val = workflow.get("kb_draft_entry_id")
|
||
if isinstance(val, str) and val:
|
||
return val[:120]
|
||
|
||
worker_result = decision_ctx.get("worker_result")
|
||
if isinstance(worker_result, dict):
|
||
val = worker_result.get("km_draft_entry_id")
|
||
if isinstance(val, str) and val:
|
||
return val[:120]
|
||
|
||
return None
|
||
|
||
|
||
def _extract_worker_status(decision_ctx: dict) -> str | None:
|
||
worker_result = decision_ctx.get("worker_result")
|
||
if not isinstance(worker_result, dict):
|
||
return None
|
||
val = worker_result.get("status")
|
||
return val[:80] if isinstance(val, str) and val else None
|
||
|
||
|
||
def _extract_dry_run_plan_fingerprint(decision_ctx: dict) -> str | None:
|
||
for source in (
|
||
decision_ctx,
|
||
decision_ctx.get("workflow"),
|
||
):
|
||
if not isinstance(source, dict):
|
||
continue
|
||
val = source.get("dry_run_plan_fingerprint")
|
||
if isinstance(val, str) and val:
|
||
return val[:80]
|
||
return None
|
||
|
||
|
||
def _extract_archived_count(decision_ctx: dict) -> int | None:
|
||
raw = decision_ctx.get("archived_count")
|
||
if isinstance(raw, int):
|
||
return max(raw, 0)
|
||
archived = decision_ctx.get("archived_entry_ids")
|
||
if isinstance(archived, list):
|
||
return len(archived)
|
||
workflow = decision_ctx.get("workflow")
|
||
if isinstance(workflow, dict):
|
||
archived = workflow.get("archived_entry_ids")
|
||
if isinstance(archived, list):
|
||
return len(archived)
|
||
return None
|
||
|
||
|
||
def _extract_stale_ratio_snapshot(decision_ctx: dict) -> dict | None:
|
||
for source in (
|
||
decision_ctx,
|
||
decision_ctx.get("workflow"),
|
||
):
|
||
if not isinstance(source, dict):
|
||
continue
|
||
snapshot = source.get("stale_ratio_snapshot")
|
||
if isinstance(snapshot, dict):
|
||
return {
|
||
key: snapshot.get(key)
|
||
for key in (
|
||
"stale_count",
|
||
"total_count",
|
||
"stale_ratio",
|
||
"threshold",
|
||
"stale_days",
|
||
)
|
||
if key in snapshot
|
||
}
|
||
return None
|
||
|
||
|
||
# =============================================================================
|
||
# Endpoint 2B: KM review draft dedupe
|
||
# =============================================================================
|
||
|
||
async def query_km_review_draft_dedupe(
|
||
*,
|
||
limit: int = 100,
|
||
) -> KnowledgeReviewDraftDedupeResponse:
|
||
"""產生 Hermes KM healthcheck review drafts 的 read-only 去重計畫。"""
|
||
rows = await _load_km_healthcheck_review_drafts(limit=limit)
|
||
event_ids = [
|
||
event_id
|
||
for row in rows
|
||
if (event_id := _extract_governance_event_id_from_tags(row.get("tags")))
|
||
]
|
||
preferred = await _load_preferred_km_draft_ids_by_event(event_ids)
|
||
archive_history = await _load_km_archive_history_by_event(event_ids)
|
||
groups = _build_km_review_draft_dedupe_groups(rows, preferred, archive_history)
|
||
|
||
return KnowledgeReviewDraftDedupeResponse(
|
||
total_review_drafts=len(rows),
|
||
event_group_total=len(groups),
|
||
duplicate_draft_total=sum(group.duplicate_count for group in groups),
|
||
groups=groups,
|
||
generated_at=now_taipei(),
|
||
)
|
||
|
||
|
||
async def _load_km_healthcheck_review_drafts(limit: int) -> list[dict[str, Any]]:
|
||
"""讀取 Hermes 產生、等待 owner review 的 KM healthcheck 草稿。"""
|
||
async with get_db_context() as db:
|
||
stmt = (
|
||
select(KnowledgeEntryRecord)
|
||
.where(
|
||
KnowledgeEntryRecord.entry_type == EntryType.AUTO_RUNBOOK,
|
||
KnowledgeEntryRecord.status == EntryStatus.REVIEW,
|
||
or_(
|
||
KnowledgeEntryRecord.title.ilike("%KM healthcheck%"),
|
||
KnowledgeEntryRecord.content.ilike("%KM healthcheck%"),
|
||
KnowledgeEntryRecord.tags.cast(String).ilike("%workflow:kb_growth_healthcheck%"),
|
||
),
|
||
)
|
||
.order_by(KnowledgeEntryRecord.updated_at.desc())
|
||
.limit(limit)
|
||
)
|
||
result = await db.execute(stmt)
|
||
records = result.scalars().all()
|
||
|
||
return [
|
||
{
|
||
"id": str(record.id),
|
||
"title": str(record.title),
|
||
"status": str(record.status.value if hasattr(record.status, "value") else record.status),
|
||
"tags": list(record.tags or []),
|
||
"created_by": record.created_by,
|
||
"created_at": record.created_at,
|
||
"updated_at": record.updated_at,
|
||
}
|
||
for record in records
|
||
]
|
||
|
||
|
||
async def _load_preferred_km_draft_ids_by_event(
|
||
event_ids: list[str],
|
||
) -> dict[str, str]:
|
||
"""從 dispatch worker_result 讀取 event 對應的 canonical KM draft id。"""
|
||
if not event_ids:
|
||
return {}
|
||
|
||
unique_event_ids = list(dict.fromkeys(event_ids))
|
||
sql = text("""
|
||
SELECT
|
||
d.governance_event_id,
|
||
d.decision_context
|
||
FROM governance_remediation_dispatch d
|
||
WHERE d.governance_event_id IN :event_ids
|
||
AND d.executor_type = 'hermes_kb_growth_healthcheck'
|
||
AND d.dispatch_status::text = 'succeeded'
|
||
ORDER BY d.completed_at DESC NULLS LAST, d.dispatched_at DESC
|
||
""").bindparams(bindparam("event_ids", expanding=True))
|
||
|
||
try:
|
||
async with get_db_context() as db:
|
||
result = await db.execute(sql, {"event_ids": unique_event_ids})
|
||
rows = result.fetchall()
|
||
except ProgrammingError as exc:
|
||
logger.warning(
|
||
"km_review_dedupe_dispatch_table_not_ready",
|
||
error=str(exc),
|
||
)
|
||
return {}
|
||
|
||
preferred: dict[str, str] = {}
|
||
for row in rows:
|
||
event_id = str(row.governance_event_id)
|
||
if event_id in preferred:
|
||
continue
|
||
decision_ctx = row.decision_context or {}
|
||
if not isinstance(decision_ctx, dict):
|
||
continue
|
||
draft_id = _extract_kb_draft_entry_id(decision_ctx)
|
||
if draft_id:
|
||
preferred[event_id] = draft_id
|
||
return preferred
|
||
|
||
|
||
async def _load_km_archive_history_by_event(
|
||
event_ids: list[str],
|
||
) -> dict[str, list[DispatchItem]]:
|
||
"""讀取 KM duplicate archive / stale ratio recheck 的 terminal audit trail。"""
|
||
if not event_ids:
|
||
return {}
|
||
|
||
unique_event_ids = list(dict.fromkeys(event_ids))
|
||
sql = text("""
|
||
SELECT
|
||
d.id,
|
||
d.governance_event_id,
|
||
e.event_type,
|
||
d.dispatch_status,
|
||
d.executor_type,
|
||
d.decision_context,
|
||
d.playbook_id,
|
||
d.dispatched_at AS created_at,
|
||
d.dispatched_at,
|
||
d.started_at,
|
||
d.completed_at,
|
||
NULL::text AS operator_note
|
||
FROM governance_remediation_dispatch d
|
||
JOIN ai_governance_events e ON e.id = d.governance_event_id
|
||
WHERE d.governance_event_id IN :event_ids
|
||
AND d.executor_type IN (
|
||
'hermes_km_review_dedupe_owner_archive',
|
||
'hermes_km_stale_ratio_recheck'
|
||
)
|
||
ORDER BY
|
||
d.governance_event_id,
|
||
d.completed_at DESC NULLS LAST,
|
||
d.started_at DESC NULLS LAST,
|
||
d.dispatched_at DESC
|
||
""").bindparams(bindparam("event_ids", expanding=True))
|
||
|
||
try:
|
||
async with get_db_context() as db:
|
||
result = await db.execute(sql, {"event_ids": unique_event_ids})
|
||
rows = result.fetchall()
|
||
except ProgrammingError as exc:
|
||
logger.warning(
|
||
"km_review_dedupe_archive_history_table_not_ready",
|
||
error=str(exc),
|
||
)
|
||
return {}
|
||
|
||
history: dict[str, list[DispatchItem]] = {}
|
||
for row in rows:
|
||
event_id = str(row.governance_event_id)
|
||
bucket = history.setdefault(event_id, [])
|
||
if len(bucket) >= 3:
|
||
continue
|
||
decision_ctx: dict = (row.decision_context or {}) if hasattr(row, "decision_context") else {}
|
||
bucket.append(_to_dispatch_item(row, decision_ctx))
|
||
|
||
return history
|
||
|
||
|
||
def _extract_governance_event_id_from_tags(tags: Any) -> str | None:
|
||
if not isinstance(tags, list):
|
||
return None
|
||
for raw in tags:
|
||
tag = str(raw)
|
||
if tag.startswith("governance_event:"):
|
||
event_id = tag.replace("governance_event:", "", 1).strip()
|
||
return event_id or None
|
||
return None
|
||
|
||
|
||
def _build_km_review_draft_dedupe_groups(
|
||
rows: list[dict[str, Any]],
|
||
preferred_draft_ids_by_event: dict[str, str] | None = None,
|
||
archive_history_by_event: dict[str, list[DispatchItem]] | None = None,
|
||
) -> list[KnowledgeReviewDraftDedupeGroup]:
|
||
"""把 KM review drafts 依 governance_event tag 分組並產生 owner action。"""
|
||
preferred_draft_ids_by_event = preferred_draft_ids_by_event or {}
|
||
archive_history_by_event = archive_history_by_event or {}
|
||
grouped: dict[str, list[dict[str, Any]]] = {}
|
||
for row in rows:
|
||
event_id = _extract_governance_event_id_from_tags(row.get("tags"))
|
||
if not event_id:
|
||
continue
|
||
grouped.setdefault(event_id, []).append(row)
|
||
|
||
groups: list[KnowledgeReviewDraftDedupeGroup] = []
|
||
for event_id, entries in grouped.items():
|
||
sorted_entries = sorted(
|
||
entries,
|
||
key=lambda item: str(item.get("updated_at") or item.get("created_at") or ""),
|
||
reverse=True,
|
||
)
|
||
preferred_id = preferred_draft_ids_by_event.get(event_id)
|
||
canonical = next(
|
||
(entry for entry in sorted_entries if entry.get("id") == preferred_id),
|
||
sorted_entries[0],
|
||
)
|
||
duplicate_ids = [
|
||
str(entry["id"])
|
||
for entry in sorted_entries
|
||
if entry.get("id") != canonical.get("id")
|
||
]
|
||
preferred_source = (
|
||
"dispatch_context"
|
||
if preferred_id and canonical.get("id") == preferred_id
|
||
else "latest_review_draft"
|
||
)
|
||
groups.append(KnowledgeReviewDraftDedupeGroup(
|
||
governance_event_id=event_id,
|
||
canonical_entry_id=str(canonical["id"]),
|
||
canonical_title=str(canonical.get("title") or ""),
|
||
canonical_updated_at=canonical.get("updated_at"),
|
||
preferred_source=preferred_source,
|
||
duplicate_entry_ids=duplicate_ids,
|
||
duplicate_count=len(duplicate_ids),
|
||
total_entries=len(sorted_entries),
|
||
suggested_action="owner_review_canonical_then_archive_duplicates",
|
||
owner_action="review_canonical_and_archive_duplicate_drafts",
|
||
writes_on_read=False,
|
||
can_archive_without_owner_approval=False,
|
||
archive_history=archive_history_by_event.get(event_id, []),
|
||
))
|
||
|
||
return sorted(
|
||
groups,
|
||
key=lambda group: (
|
||
group.duplicate_count,
|
||
group.canonical_updated_at.isoformat() if group.canonical_updated_at else "",
|
||
),
|
||
reverse=True,
|
||
)
|
||
|
||
|
||
# =============================================================================
|
||
# Endpoint 3: summary
|
||
# =============================================================================
|
||
|
||
async def query_governance_summary(*, days: int = 30) -> GovernanceSummaryResponse:
|
||
"""
|
||
過去 N 天 SLO 違反時序統計 + compliance_rate。
|
||
|
||
compliance_rate = 1 - unresolved / total（total=0 時回 1.0）
|
||
"""
|
||
since = now_taipei() - timedelta(days=days)
|
||
|
||
async with get_db_context() as db:
|
||
# 總數 & 未解決數
|
||
count_stmt = select(
|
||
func.count().label("total"),
|
||
func.count().filter(AiGovernanceEvent.resolved.is_(False)).label("unresolved"),
|
||
).where(AiGovernanceEvent.triggered_at >= since)
|
||
|
||
count_row = await db.execute(count_stmt)
|
||
counts = count_row.one()
|
||
total_events = int(counts.total)
|
||
unresolved_count = int(counts.unresolved)
|
||
|
||
# 每日計數（DATE_TRUNC 在 Postgres 端執行）
|
||
daily_sql = text("""
|
||
SELECT
|
||
DATE_TRUNC('day', triggered_at AT TIME ZONE 'Asia/Taipei')::date AS day,
|
||
event_type,
|
||
count(*) AS cnt
|
||
FROM ai_governance_events
|
||
WHERE triggered_at >= :since
|
||
GROUP BY day, event_type
|
||
ORDER BY day ASC
|
||
""")
|
||
daily_result = await db.execute(daily_sql, {"since": since})
|
||
daily_rows = daily_result.fetchall()
|
||
|
||
# 彙整每日資料
|
||
daily_map: dict[str, dict[str, int]] = {}
|
||
for row in daily_rows:
|
||
day_str = row.day.strftime("%Y-%m-%d") if hasattr(row.day, "strftime") else str(row.day)
|
||
if day_str not in daily_map:
|
||
daily_map[day_str] = {}
|
||
daily_map[day_str][row.event_type] = int(row.cnt)
|
||
|
||
daily_counts = [
|
||
DailyCount(
|
||
date=day_str,
|
||
total=sum(by_type.values()),
|
||
by_type=by_type,
|
||
)
|
||
for day_str, by_type in sorted(daily_map.items())
|
||
]
|
||
|
||
if total_events == 0:
|
||
compliance_rate = 1.0
|
||
else:
|
||
compliance_rate = round(1.0 - unresolved_count / total_events, 4)
|
||
|
||
return GovernanceSummaryResponse(
|
||
compliance_rate=compliance_rate,
|
||
total_events=total_events,
|
||
unresolved_count=unresolved_count,
|
||
daily_counts=daily_counts,
|
||
)
|