diff --git a/SESSION.md b/SESSION.md index 1390f68..4b57d24 100644 --- a/SESSION.md +++ b/SESSION.md @@ -2,7 +2,7 @@ ## Current Goal -Automation execution-layer P18 is closed locally. P18 adds a stable visual media evidence block so alert deliveries and workbench visual reviews can consume screenshot/media metadata through `media_attached`, `media_linked`, and `media_unavailable` states without breaking text-only channels. The default next action is P19 `Multi-Channel Alert Routing`. +Automation execution-layer P19 is closed locally. P19 adds structured channel routing and channel health observability so alert dispatch can distinguish enabled workbench/webhook paths from disabled WeCom/Email adapters with credential-missing reasons. The default next action is P20 `Operator Digest And Review Queue`. ## Current Branch @@ -14,8 +14,8 @@ Automation execution-layer P18 is closed locally. P18 adds a stable visual media ## Automation Status -- state: automation_execution_layer_p18_visual_evidence_media_dispatch_closed -- current_phase: P18 Visual Evidence Media Dispatch +- state: automation_execution_layer_p19_multi_channel_alert_routing_closed +- current_phase: P19 Multi-Channel Alert Routing - latest_p7_selector_choice: continue_periodic_analysis_decision_monitoring_with_maturity_rollup - latest_p14_slice: webhook_transport_retry - latest_p15_pxx_tasklist_status: ready @@ -37,7 +37,12 @@ Automation execution-layer P18 is closed locally. P18 adds a stable visual media - latest_visual_media_evidence_status: implemented - latest_visual_media_evidence_dispatch_status: visible - latest_visual_media_evidence_smoke_status: no_pending_alerts_safe_noop -- next_execution_layer_phase: P19 Multi-Channel Alert Routing +- latest_p19_slice: multi_channel_alert_routing +- latest_channel_router_status: implemented +- latest_channel_health_status: visible +- latest_disabled_adapter_status: visible_credential_missing +- latest_channel_routing_smoke_status: no_pending_alerts_safe_noop +- next_execution_layer_phase: P20 Operator Digest And Review Queue - latest_visual_review_script_status: idle - latest_runtime_visual_watch_task_status: ok - latest_visual_capture_cleanup_task_status: ok @@ -110,6 +115,9 @@ Automation execution-layer P18 is closed locally. P18 adds a stable visual media - Added `build_visual_media_evidence()` and `summarize_visual_media_evidence()` so visual reviews produce a stable media evidence block. - Extended alert dispatcher delivery payloads, dispatcher results, alert messages, alert quality summary, runtime `alert_dispatch` details, and `run_alert_dispatcher.py` with media evidence status. - Extended visual review API enrichment and workbench Visual Reviews with media evidence status and primary media path. +- Added structured channel route/capability helpers with workbench, webhook, WeCom, and email channel descriptions. +- Added channel health summary and disabled adapter metadata for WeCom/Email credential-missing states. +- Extended alert dispatcher payloads/results, delivery summary, runtime `alert_dispatch` details, dispatcher script output, and workbench delivery cards with per-channel observability. ## Current Working State @@ -131,6 +139,8 @@ Automation execution-layer P18 is closed locally. P18 adds a stable visual media - P17 visibility is implemented: decision alert payloads carry `visual_trigger_policy`/`visual_trigger_roi`, runtime exposes trigger/skip counts, visual review summaries count trigger types and review results, and the workbench displays trigger counters. - P18 visual media evidence dispatch is implemented: visual review context now becomes a stable `media_evidence` block with channel-aware `media_attached`, `media_linked`, or `media_unavailable` status. - P18 visibility is implemented: dispatcher payloads/results, delivery summary, runtime `alert_dispatch` details, script output, visual review API, and workbench panels expose media evidence status. +- P19 multi-channel alert routing is implemented: routing now returns `channel_route`, `channel_health`, channel capability metadata, and explicit disabled-adapter states for WeCom/Email when credentials/adapters are unavailable. +- P19 visibility is implemented: dispatcher result exposes `channel_summary` / `channel_health` / `channel_disabled`, delivery summary exposes `channel_status_counts`, runtime `alert_dispatch` details expose channel health, and the workbench `Alert Quality` panel shows channel counts. - Latest `scripts/run_visual_review.py` smoke found no pending reviews: attempted=0, reviewed=0, status=idle. - Latest capture smoke returned `status=ignored` when no watch run was supplied, which is the expected safe no-op. - Latest cleanup smoke deleted one expired temporary artifact through `scripts/run_visual_capture_cleanup.py`. @@ -140,7 +150,7 @@ Automation execution-layer P18 is closed locally. P18 adds a stable visual media - Latest `scripts/run_trading_runtime.py` smoke returned blockers=none and includes a healthy `visual_watch` runtime task. - Latest local runtime smoke confirmed `runtime_status.tasks.alert_dispatch.details` now exposes retry and silent-window counters with blockers=none. -## P18 Blockers +## P19 Blockers - none @@ -159,6 +169,7 @@ Automation execution-layer P18 is closed locally. P18 adds a stable visual media - P16 browser screenshot capture remains advisory evidence only and must not change execution permission, live readiness, trust bucket, risk governor, or operator confirmation. - P17 visual trigger routing remains advisory evidence routing only and must not change execution permission, live readiness, trust bucket, risk governor, or operator confirmation. - P18 media evidence remains delivery metadata/fallback routing only and must not change execution permission, live readiness, trust bucket, risk governor, or operator confirmation. +- P19 channel routing remains notification routing/observability only and must not change execution permission, live readiness, trust bucket, risk governor, or operator confirmation. - A final rerun of `scripts/run_alert_dispatcher.py` and `scripts/run_trading_runtime.py` hit local PostgreSQL connection timeout because nothing was listening on `127.0.0.1:5432`; the P14 code path is still covered by full unit tests and earlier successful smoke during implementation. - P15 dispatcher smoke succeeded with `suppressed_noise_policy`, `workbench_only`, and `policy_external_allowed` counters present. P15 runtime smoke reached `alert_dispatch` successfully, but the overall runtime returned external environment blockers for market stream disconnected and OKX read transport connection refused. - P16 direct browser smoke succeeded locally with `browser_capture_status=captured`. Script-level capture without a watch run still returns safe no-op `watch_run_missing`. @@ -168,11 +179,14 @@ Automation execution-layer P18 is closed locally. P18 adds a stable visual media - P18 dispatcher smoke returned attempted=0 and media counters present as safe no-op: `media_attached=0`, `media_linked=0`, `media_unavailable=0`. - P18 quality-summary API smoke returned 200 and exposed `delivery_summary.media_attached_count`, `media_linked_count`, and `media_unavailable_count`. - Latest P18 runtime smoke returned blockers=none and exposes `alert_dispatch.details.media_attached/media_linked/media_unavailable`. +- P19 dispatcher smoke returned attempted=0 and exposed `channel_health.enabled_channels=["workbench"]` plus disabled `webhook`, `wecom`, and `email` credential-missing states. +- P19 quality-summary API smoke returned 200 and exposed `delivery_summary.channel_status_counts`. +- Latest P19 runtime smoke returned blockers=none and exposes `alert_dispatch.details.channel_health` and `channel_summary`. - Real-live order submission remains intentionally disabled because the OKX order adapter is still not a real-live submitter. ## Exact Next Step -P18 is locally closed. Default next action is P19 `Multi-Channel Alert Routing`: extend channel capability/routing observability with disabled adapters and per-channel health/status counts, without changing trading gates. +P19 is locally closed. Default next action is P20 `Operator Digest And Review Queue`: aggregate recent alerts, visual reviews, feedback, suppression, retries, media, and channel health into an operator-facing digest/review queue. ## Validation Commands Used Recently @@ -192,4 +206,5 @@ D:\ai-exchange\.venv\Scripts\python.exe -m unittest discover -s tests/unit D:\ai-exchange\.venv\Scripts\python.exe scripts\run_visual_review.py D:\ai-exchange\.venv\Scripts\python.exe -m unittest tests.unit.test_visual_media_evidence_service tests.unit.test_alert_feedback_service tests.unit.test_alert_quality_summary_service tests.unit.test_trading_runtime_service tests.unit.test_api_and_main D:\ai-exchange\.venv\Scripts\python.exe scripts\run_alert_dispatcher.py +D:\ai-exchange\.venv\Scripts\python.exe -m unittest tests.unit.test_channel_router_service tests.unit.test_alert_feedback_service tests.unit.test_alert_quality_summary_service tests.unit.test_trading_runtime_service ``` diff --git a/TASKS.md b/TASKS.md index 92ffcc2..36ca65b 100644 --- a/TASKS.md +++ b/TASKS.md @@ -13,7 +13,7 @@ ## In Progress -- [x] P18 Visual Evidence Media Dispatch local closure complete +- [x] P19 Multi-Channel Alert Routing local closure complete ## Next Up @@ -29,14 +29,15 @@ - [x] P16 local closure complete - [x] P17 local closure complete - [x] P18 local closure complete -- [ ] Start P19 Multi-Channel Alert Routing +- [x] P19 local closure complete +- [ ] Start P20 Operator Digest And Review Queue - [ ] Continue periodic analysis-decision monitoring with maturity rollup in parallel - [ ] Treat browser screenshot capture and additional external media adapters as future enhancements ## Automation Status -- [x] state: automation_execution_layer_p18_visual_evidence_media_dispatch_closed -- [x] current_phase: P18 Visual Evidence Media Dispatch +- [x] state: automation_execution_layer_p19_multi_channel_alert_routing_closed +- [x] current_phase: P19 Multi-Channel Alert Routing - [x] latest_p7_selector_choice: continue_periodic_analysis_decision_monitoring_with_maturity_rollup - [x] latest_p14_slice: webhook_transport_retry - [x] latest_p15_pxx_tasklist_status: ready @@ -58,7 +59,12 @@ - [x] latest_visual_media_evidence_status: implemented - [x] latest_visual_media_evidence_dispatch_status: visible - [x] latest_visual_media_evidence_smoke_status: no_pending_alerts_safe_noop -- [x] next_execution_layer_phase: P19 Multi-Channel Alert Routing +- [x] latest_p19_slice: multi_channel_alert_routing +- [x] latest_channel_router_status: implemented +- [x] latest_channel_health_status: visible +- [x] latest_disabled_adapter_status: visible_credential_missing +- [x] latest_channel_routing_smoke_status: no_pending_alerts_safe_noop +- [x] next_execution_layer_phase: P20 Operator Digest And Review Queue - [x] latest_visual_review_script_status: idle - [x] latest_runtime_visual_watch_task_status: ok - [x] latest_visual_capture_cleanup_task_status: ok @@ -144,6 +150,11 @@ - [x] Added `media_attached`, `media_linked`, and `media_unavailable` counters to dispatcher, delivery summary, runtime `alert_dispatch`, and dispatcher script output - [x] Added media evidence status and primary path visibility to Visual Reviews API/workbench - [x] Added focused P18 media evidence, dispatcher, quality summary, runtime, and API tests +- [x] Added structured channel route/capability helpers for workbench, webhook, WeCom, and email +- [x] Added channel health summary with enabled/disabled/credential-missing channel visibility +- [x] Added disabled adapter states for WeCom and email when credentials or adapter support are unavailable +- [x] Added per-channel summary to dispatcher results, delivery summary, runtime `alert_dispatch`, dispatcher script output, and workbench Alert Quality +- [x] Added focused P19 channel router, disabled channel, summary, and runtime tests - [x] Added runtime `visual_watch` task - [x] Added focused service, script, API, and runtime tests - [x] Smoked the visual review script with no pending reviews @@ -163,6 +174,9 @@ - [x] Smoked P18 dispatcher safe no-op with media counters visible - [x] Smoked P18 quality-summary API with media counters visible - [x] Smoked P18 trading runtime with blockers=none and alert dispatch media counters visible +- [x] Smoked P19 dispatcher safe no-op with channel health visible +- [x] Smoked P19 quality-summary API with channel status counts visible +- [x] Smoked P19 trading runtime with blockers=none and alert dispatch channel health visible ## Blocked @@ -177,7 +191,7 @@ ## Future Enhancements Not Blocking P8 -- [ ] P19 Multi-Channel Alert Routing +- [x] none currently tracked outside the finite P15-PXX queue ## Deferred Until Real-Trading Phase @@ -217,7 +231,7 @@ - [x] `P16` Browser Screenshot Capture Adapter - [x] `P17` Visual Trigger ROI Router - [x] `P18` Visual Evidence Media Dispatch -- [ ] `P19` Multi-Channel Alert Routing +- [x] `P19` Multi-Channel Alert Routing - [ ] `P20` Operator Digest And Review Queue - [ ] `P21` Runtime Incident Journal - [ ] `P22` Maintenance And Retention Rollup diff --git a/docs/AI_ICT_盯盘告警开发任务清单_v1.md b/docs/AI_ICT_盯盘告警开发任务清单_v1.md index f17e060..9cd8db9 100644 --- a/docs/AI_ICT_盯盘告警开发任务清单_v1.md +++ b/docs/AI_ICT_盯盘告警开发任务清单_v1.md @@ -408,6 +408,7 @@ candidate -> decision alert -> dispatcher -> 外部消息 -> delivery record - 2026-05-12 P16 已完成 `browser screenshot capture adapter` 本地闭环：visual snapshot capture 现在在保留 HTML/JSON artifact 的同时，可通过本机 Edge/Chrome headless 生成 PNG screenshot，并把 browser capture status、HTML path、screenshot path 接入脚本、API 与 workbench Visual Reviews。 - 2026-05-12 P17 已完成 `visual trigger ROI router` 本地闭环：视觉触发策略现在区分 `triggered` / `skipped`，普通 observe/monitoring 状态不触发视觉复核，actionable、blocked、market degraded、broker/reconciliation degraded 与 council watch item 会输出可审计 trigger type / reason codes，并在 runtime、alert quality summary、脚本和 workbench 中暴露 trigger/skip/ROI 计数。 - 2026-05-12 P18 已完成 `visual evidence media dispatch` 本地闭环：visual review 现在会生成稳定 `media_evidence` block，外部文本通道不支持文件时降级为 `media_linked`，workbench 可消费本地截图路径时标记 `media_attached`，无视觉证据时标记 `media_unavailable`；dispatcher、delivery summary、runtime、脚本和 workbench 均可见媒体证据状态。 +- 2026-05-12 P19 已完成 `multi-channel alert routing` 本地闭环：channel router 现在输出 workbench/webhook/WeCom/Email capability 与 health summary，WeCom/Email 在凭据缺失或 adapter 未实现时会以 disabled adapter 形式返回 `credential_missing` / `adapter_not_implemented`，dispatcher、delivery summary、runtime、脚本和 workbench 均可见 per-channel 状态。 ### 验收 - 非重大错误不打断主循环 diff --git a/docs/AI_ICT_盯盘告警技术设计稿_v1.md b/docs/AI_ICT_盯盘告警技术设计稿_v1.md index 4e2743b..6cf1649 100644 --- a/docs/AI_ICT_盯盘告警技术设计稿_v1.md +++ b/docs/AI_ICT_盯盘告警技术设计稿_v1.md @@ -107,6 +107,8 @@ - `candidate_blocked` -> dashboard only - `runtime_degraded` -> Telegram + dashboard +2026-05-12 P19 已落地：当前 channel router 保持现有 workbench/webhook 行为，同时输出结构化 `channel_route`、channel capability 与 channel health summary。WeCom / Email 已预留为 disabled adapter；凭据缺失时暴露 `credential_missing`，凭据存在但 adapter 未实现时暴露 `adapter_not_implemented`，不会把未配置通道伪装为成功发送。dispatcher、delivery summary、runtime 与 workbench 均可查看 per-channel 状态。 + #### `src/services/alerts/delivery_policy_service.py` 职责： diff --git a/docs/AI_ICT_自动化执行层_P15-PXX闭环任务清单_v1.md b/docs/AI_ICT_自动化执行层_P15-PXX闭环任务清单_v1.md index f323eb4..934402c 100644 --- a/docs/AI_ICT_自动化执行层_P15-PXX闭环任务清单_v1.md +++ b/docs/AI_ICT_自动化执行层_P15-PXX闭环任务清单_v1.md @@ -189,6 +189,13 @@ - 凭据缺失时显示 `channel_disabled` 或 `credential_missing`，不伪装成功。 - 现有 webhook/workbench 行为不回归。 +### 当前进展 + +- 2026-05-12 P19 已完成 `Multi-Channel Alert Routing` 本地闭环：`ChannelRouterService` 现在输出结构化 `channel_route`、channel capability 与 channel health summary，覆盖 workbench、webhook、WeCom、Email。 +- 2026-05-12 P19 已将 WeCom/Email 预留为 disabled adapter：凭据缺失时显示 `credential_missing`，即使未来填入凭据但 adapter 未实现也会保持 `adapter_not_implemented`，不会伪装成成功发送。 +- 2026-05-12 P19 已在 dispatcher result、delivery payload、delivery summary、runtime `alert_dispatch` details、`run_alert_dispatcher.py` 与 workbench `Alert Quality` 中暴露 per-channel status、channel disabled count 与 channel health。 +- 2026-05-12 P19 本地闭环已验证：focused P19 tests 通过，`tests/unit` 全量 679 个测试通过，script static checks 与 `node --check src/web/app.js` 通过；dispatcher smoke 在无 pending alert 时安全 no-op 且 channel health 可见，quality-summary API 与 runtime smoke 均可见 channel counts/health 且 runtime blockers=none。 + --- ## 10. P20 Operator Digest And Review Queue diff --git a/scripts/run_alert_dispatcher.py b/scripts/run_alert_dispatcher.py index 5837217..0fcfa3a 100644 --- a/scripts/run_alert_dispatcher.py +++ b/scripts/run_alert_dispatcher.py @@ -36,6 +36,9 @@ def main(env=None) -> int: "Media attached / linked / unavailable: " f"{payload.get('media_attached', 0)} / {payload.get('media_linked', 0)} / {payload.get('media_unavailable', 0)}" ) + print(f"Channel disabled: {payload.get('channel_disabled', 0)}") + if payload.get("channel_summary"): + print(f"Channel summary: {json.dumps(payload.get('channel_summary'), ensure_ascii=False, sort_keys=True)}") items = payload.get("items") or [] return 0 if all(item.get("status") != "failed" for item in items) else 2 diff --git a/src/config/settings.py b/src/config/settings.py index 867098a..ed91236 100644 --- a/src/config/settings.py +++ b/src/config/settings.py @@ -121,6 +121,9 @@ class Settings(BaseSettings): alert_noise_suppressed_operator_modes: str = "silent,workbench_only" alert_default_channel: str = "workbench" alert_webhook_url: str = "" + alert_wecom_webhook_url: str = "" + alert_email_smtp_url: str = "" + alert_email_to: str = "" default_timezone: str = "UTC" session_timezone: str = "America/New_York" log_level: str = "INFO" diff --git a/src/services/alerts/alert_dispatcher_service.py b/src/services/alerts/alert_dispatcher_service.py index 558c671..7461c92 100644 --- a/src/services/alerts/alert_dispatcher_service.py +++ b/src/services/alerts/alert_dispatcher_service.py @@ -9,7 +9,7 @@ from src.config.settings import settings from src.repositories.alert_delivery_repository import AlertDeliveryRepository from src.repositories.decision_packet_repository import DecisionPacketRepository from src.services.feedback import AlertFeedbackService -from src.services.alerts.channel_router_service import select_alert_channel +from src.services.alerts.channel_router_service import build_channel_health_summary, select_alert_route from src.services.alerts.delivery_policy_service import evaluate_alert_delivery_policy from src.services.alerts.alert_silent_window_service import AlertSilentWindowService from src.services.alerts.message_template_service import build_alert_message @@ -74,6 +74,9 @@ class AlertDispatcherService: "media_attached": sum(1 for item in items if item.get("media_status") == "media_attached"), "media_linked": sum(1 for item in items if item.get("media_status") == "media_linked"), "media_unavailable": sum(1 for item in items if item.get("media_status") == "media_unavailable"), + "channel_summary": _build_dispatch_channel_summary(items), + "channel_health": build_channel_health_summary(), + "channel_disabled": sum(1 for item in items if item.get("reason") == "channel_disabled"), "items": items, "silent_window": silent_window, "suppressed_silent_window": sum(1 for item in items if item.get("reason") == "alert_silent_window_active"), @@ -84,7 +87,9 @@ class AlertDispatcherService: if not alert: return {"alert_id": alert_id, "status": "missing"} - channel = select_alert_channel(alert) + channel_route = select_alert_route(alert) + channel = str(channel_route.get("selected_channel") or "workbench") + channel_capability = dict(channel_route.get("capability") or {}) packet = None packet_id = alert.get("decision_packet_id") if packet_id: @@ -113,6 +118,38 @@ class AlertDispatcherService: reason = str(delivery_policy.get("reason") or "") attempted_ts = datetime.now(timezone.utc) + if not bool(channel_capability.get("enabled", True)): + reason = "channel_disabled" + self._safe_update_alert_status(alert_id=alert_id, status="skipped", channel=channel) + self._safe_insert_delivery( + alert_id=alert_id, + channel=channel, + status="skipped", + attempted_ts=attempted_ts, + error=reason, + payload={ + "reason": reason, + "message": message, + "dedupe_scope": dedupe_scope, + "visual": visual, + "media_evidence": media_evidence, + "delivery_policy": delivery_policy, + "channel_route": channel_route, + }, + ) + return { + "alert_id": alert_id, + "status": "skipped", + "dispatch_mode": "failed_retry" if str(alert.get("status") or "") == "failed" else "queued", + "channel": channel, + "reason": reason, + "channel_route": channel_route, + "delivery_policy": delivery_policy, + "media_status": media_evidence.get("status"), + "media_evidence": media_summary, + "visual": self._visual_result_summary(visual, media_evidence=media_evidence), + } + if str(alert.get("status") or "") == "failed" and not allow_failed_retry: reason = str(retry_state.get("reason") or "retry_not_allowed") if reason == "retry_backoff_active": @@ -124,6 +161,7 @@ class AlertDispatcherService: "reason": reason, "retry_attempt_count": retry_state.get("retry_attempt_count"), "next_retry_ts": retry_state.get("next_retry_ts"), + "channel_route": channel_route, "delivery_policy": delivery_policy, "media_status": media_evidence.get("status"), "media_evidence": media_summary, @@ -137,6 +175,7 @@ class AlertDispatcherService: "channel": channel, "reason": reason, "next_retry_ts": retry_state.get("next_retry_ts"), + "channel_route": channel_route, "delivery_policy": delivery_policy, "media_status": media_evidence.get("status"), "media_evidence": media_summary, @@ -156,6 +195,7 @@ class AlertDispatcherService: "visual": visual, "media_evidence": media_evidence, "delivery_policy": delivery_policy, + "channel_route": channel_route, "dispatch_mode": "failed_retry", "retry_attempt_count": retry_state.get("retry_attempt_count"), }, @@ -167,6 +207,7 @@ class AlertDispatcherService: "channel": channel, "reason": reason, "retry_attempt_count": retry_state.get("retry_attempt_count"), + "channel_route": channel_route, "delivery_policy": delivery_policy, "media_status": media_evidence.get("status"), "media_evidence": media_summary, @@ -181,7 +222,7 @@ class AlertDispatcherService: status="skipped", attempted_ts=attempted_ts, error=reason, - payload={"reason": reason, "message": message, "dedupe_scope": dedupe_scope, "visual": visual, "media_evidence": media_evidence, "delivery_policy": delivery_policy}, + payload={"reason": reason, "message": message, "dedupe_scope": dedupe_scope, "visual": visual, "media_evidence": media_evidence, "delivery_policy": delivery_policy, "channel_route": channel_route}, ) return { "alert_id": alert_id, @@ -189,6 +230,7 @@ class AlertDispatcherService: "dispatch_mode": "failed_retry" if str(alert.get("status") or "") == "failed" else "queued", "channel": channel, "reason": reason, + "channel_route": channel_route, "delivery_policy": delivery_policy, "media_status": media_evidence.get("status"), "media_evidence": media_summary, @@ -202,7 +244,7 @@ class AlertDispatcherService: channel=channel, status="skipped", attempted_ts=attempted_ts, - payload={"message": message, "reason": "workbench_only", "dedupe_scope": dedupe_scope, "visual": visual, "media_evidence": media_evidence, "delivery_policy": delivery_policy}, + payload={"message": message, "reason": "workbench_only", "dedupe_scope": dedupe_scope, "visual": visual, "media_evidence": media_evidence, "delivery_policy": delivery_policy, "channel_route": channel_route}, ) return { "alert_id": alert_id, @@ -210,6 +252,7 @@ class AlertDispatcherService: "dispatch_mode": "failed_retry" if str(alert.get("status") or "") == "failed" else "queued", "channel": channel, "reason": "workbench_only", + "channel_route": channel_route, "delivery_policy": delivery_policy, "media_status": media_evidence.get("status"), "media_evidence": media_summary, @@ -232,7 +275,7 @@ class AlertDispatcherService: attempted_ts=attempted_ts, delivered_ts=delivered_ts, external_message_id=result.get("external_message_id"), - payload={"message": message, "result": result, "dedupe_scope": dedupe_scope, "visual": visual, "media_evidence": media_evidence, "delivery_policy": delivery_policy}, + payload={"message": message, "result": result, "dedupe_scope": dedupe_scope, "visual": visual, "media_evidence": media_evidence, "delivery_policy": delivery_policy, "channel_route": channel_route}, ) return { "alert_id": alert_id, @@ -243,6 +286,7 @@ class AlertDispatcherService: "delivery_write_ok": delivery_write_ok, "retry_attempt_count": retry_state.get("retry_attempt_count"), "transport_attempt_count": int(result.get("attempt_count") or 1), + "channel_route": channel_route, "delivery_policy": delivery_policy, "media_status": media_evidence.get("status"), "media_evidence": media_summary, @@ -257,7 +301,7 @@ class AlertDispatcherService: status="failed", attempted_ts=attempted_ts, error=str(exc), - payload={"message": message, "dedupe_scope": dedupe_scope, "visual": visual, "media_evidence": media_evidence, "transport_attempt_count": transport_attempt_count, "delivery_policy": delivery_policy}, + payload={"message": message, "dedupe_scope": dedupe_scope, "visual": visual, "media_evidence": media_evidence, "transport_attempt_count": transport_attempt_count, "delivery_policy": delivery_policy, "channel_route": channel_route}, ) return { "alert_id": alert_id, @@ -269,6 +313,7 @@ class AlertDispatcherService: "delivery_write_ok": delivery_write_ok, "retry_attempt_count": retry_state.get("retry_attempt_count"), "transport_attempt_count": transport_attempt_count, + "channel_route": channel_route, "delivery_policy": delivery_policy, "media_status": media_evidence.get("status"), "media_evidence": media_summary, @@ -451,6 +496,30 @@ def _snapshot_artifact_path(visual: dict | None) -> str | None: return None +def _build_dispatch_channel_summary(items: list[dict]) -> dict: + summary: dict[str, dict] = {} + for item in list(items or []): + channel = str(item.get("channel") or "unknown") + channel_summary = summary.setdefault( + channel, + { + "attempted": 0, + "delivered": 0, + "failed": 0, + "skipped": 0, + "deferred": 0, + "channel_disabled": 0, + }, + ) + channel_summary["attempted"] += 1 + status = str(item.get("status") or "unknown") + if status in channel_summary: + channel_summary[status] += 1 + if item.get("reason") == "channel_disabled": + channel_summary["channel_disabled"] += 1 + return summary + + def _parse_ts(value) -> datetime | None: if value in (None, ""): return None diff --git a/src/services/alerts/alert_quality_summary_service.py b/src/services/alerts/alert_quality_summary_service.py index 8fd5287..570ac5b 100644 --- a/src/services/alerts/alert_quality_summary_service.py +++ b/src/services/alerts/alert_quality_summary_service.py @@ -136,6 +136,8 @@ def _build_delivery_summary(deliveries: list[dict]) -> dict: "media_attached_count": sum(1 for item in items if _media_status(item) == "media_attached"), "media_linked_count": sum(1 for item in items if _media_status(item) == "media_linked"), "media_unavailable_count": sum(1 for item in items if _media_status(item) == "media_unavailable"), + "channel_status_counts": _build_channel_status_counts(items), + "channel_disabled_count": sum(1 for item in items if str((item.get("payload") or {}).get("reason") or item.get("error") or "") == "channel_disabled"), "external_attempt_count": external_attempts, "success_rate": _rate(delivered, external_attempts), "latest_status": latest.get("status"), @@ -177,6 +179,16 @@ def _media_status(delivery: dict) -> str: return str(evidence.get("status") or "") +def _build_channel_status_counts(deliveries: list[dict]) -> dict: + counts: dict[str, dict[str, int]] = {} + for delivery in deliveries: + channel = str(delivery.get("channel") or "unknown") + status = str(delivery.get("status") or "unknown") + channel_counts = counts.setdefault(channel, {}) + channel_counts[status] = channel_counts.get(status, 0) + 1 + return counts + + def _build_visual_review_summary(rows: list[dict]) -> dict: items = list(rows or []) state_counts: dict[str, int] = {} diff --git a/src/services/alerts/channel_router_service.py b/src/services/alerts/channel_router_service.py index 697319c..8c9c5a3 100644 --- a/src/services/alerts/channel_router_service.py +++ b/src/services/alerts/channel_router_service.py @@ -3,14 +3,128 @@ from __future__ import annotations from src.config.settings import settings +CHANNEL_WORKBENCH = "workbench" +CHANNEL_WEBHOOK = "webhook" +CHANNEL_WECOM = "wecom" +CHANNEL_EMAIL = "email" + WEBHOOK_ALERT_TYPES = {"candidate_actionable", "ticket_ready", "runtime_degraded"} +KNOWN_CHANNELS = [CHANNEL_WORKBENCH, CHANNEL_WEBHOOK, CHANNEL_WECOM, CHANNEL_EMAIL] def select_alert_channel(alert: dict) -> str: + return select_alert_route(alert)["selected_channel"] + + +def select_alert_route(alert: dict) -> dict: alert_type = str(alert.get("alert_type") or "") action = str((alert.get("payload") or {}).get("action_recommendation") or "") if alert_type in WEBHOOK_ALERT_TYPES and settings.alert_webhook_url: - return "webhook" - if alert_type == "candidate_blocked" and action == "do_not_execute" and settings.alert_webhook_url: - return "webhook" - return settings.alert_default_channel + selected = CHANNEL_WEBHOOK + rule = "webhook_high_priority_alert" + elif alert_type == "candidate_blocked" and action == "do_not_execute" and settings.alert_webhook_url: + selected = CHANNEL_WEBHOOK + rule = "webhook_do_not_execute_block" + else: + selected = _normalize_channel(settings.alert_default_channel) + rule = "default_channel" + capability = get_channel_capability(selected) + return { + "selected_channel": selected, + "route_rule": rule, + "capability": capability, + "channel_health": build_channel_health_summary(), + } + + +def get_channel_capability(channel: str) -> dict: + normalized = _normalize_channel(channel) + if normalized == CHANNEL_WORKBENCH: + return { + "channel": CHANNEL_WORKBENCH, + "enabled": True, + "status": "enabled", + "reason": None, + "supports_text": True, + "supports_media_attachment": True, + "supports_media_link": True, + "adapter": "workbench", + } + if normalized == CHANNEL_WEBHOOK: + enabled = bool(settings.alert_webhook_url) + return _external_capability( + channel=CHANNEL_WEBHOOK, + enabled=enabled, + adapter="webhook", + credential_key="alert_webhook_url", + supports_media_attachment=False, + ) + if normalized == CHANNEL_WECOM: + has_credential = bool(settings.alert_wecom_webhook_url) + return _external_capability( + channel=CHANNEL_WECOM, + enabled=False, + adapter="wecom_disabled_adapter", + credential_key="alert_wecom_webhook_url", + supports_media_attachment=False, + disabled_reason="adapter_not_implemented" if has_credential else "credential_missing", + ) + if normalized == CHANNEL_EMAIL: + has_credential = bool(settings.alert_email_smtp_url and settings.alert_email_to) + missing_key = "alert_email_smtp_url_or_to" + return _external_capability( + channel=CHANNEL_EMAIL, + enabled=False, + adapter="email_disabled_adapter", + credential_key=missing_key, + supports_media_attachment=False, + disabled_reason="adapter_not_implemented" if has_credential else "credential_missing", + ) + return { + "channel": normalized, + "enabled": False, + "status": "disabled", + "reason": "unknown_channel", + "supports_text": False, + "supports_media_attachment": False, + "supports_media_link": False, + "adapter": "unknown", + } + + +def build_channel_health_summary(channels: list[str] | None = None) -> dict: + capabilities = [get_channel_capability(channel) for channel in list(channels or KNOWN_CHANNELS)] + disabled = [item for item in capabilities if not item.get("enabled")] + return { + "channels": {item["channel"]: item for item in capabilities}, + "enabled_channels": [item["channel"] for item in capabilities if item.get("enabled")], + "disabled_channels": [item["channel"] for item in disabled], + "credential_missing_channels": [item["channel"] for item in disabled if item.get("reason") == "credential_missing"], + "disabled_count": len(disabled), + } + + +def _external_capability( + *, + channel: str, + enabled: bool, + adapter: str, + credential_key: str, + supports_media_attachment: bool, + disabled_reason: str = "credential_missing", +) -> dict: + return { + "channel": channel, + "enabled": enabled, + "status": "enabled" if enabled else "disabled", + "reason": None if enabled else disabled_reason, + "credential_key": credential_key, + "supports_text": True, + "supports_media_attachment": supports_media_attachment, + "supports_media_link": True, + "adapter": adapter, + } + + +def _normalize_channel(channel: str) -> str: + return str(channel or CHANNEL_WORKBENCH).strip().lower() or CHANNEL_WORKBENCH diff --git a/src/services/runtime/trading_runtime_service.py b/src/services/runtime/trading_runtime_service.py index 10a0499..29712c7 100644 --- a/src/services/runtime/trading_runtime_service.py +++ b/src/services/runtime/trading_runtime_service.py @@ -549,6 +549,9 @@ class TradingRuntimeScheduler: "media_attached": int(payload.get("media_attached") or 0), "media_linked": int(payload.get("media_linked") or 0), "media_unavailable": int(payload.get("media_unavailable") or 0), + "channel_disabled": int(payload.get("channel_disabled") or 0), + "channel_summary": dict(payload.get("channel_summary") or {}), + "channel_health": dict(payload.get("channel_health") or {}), "delivered": sum(1 for item in items if item.get("status") == "delivered"), "skipped": sum(1 for item in items if item.get("status") == "skipped"), "failed": len(failed), diff --git a/src/web/app.js b/src/web/app.js index 92e5ea9..548688d 100644 --- a/src/web/app.js +++ b/src/web/app.js @@ -589,6 +589,8 @@ function renderAlertQuality(summary) { ['transport retry failed', String(delivery.transport_retry_failed_count ?? 0)], ['policy external allowed', String(delivery.policy_external_allowed_count ?? 0)], ['media attached / linked / unavailable', `${delivery.media_attached_count ?? 0} / ${delivery.media_linked_count ?? 0} / ${delivery.media_unavailable_count ?? 0}`], + ['channel disabled', String(delivery.channel_disabled_count ?? 0)], + ['channel status counts', formatJsonInline(delivery.channel_status_counts || {})], ['external send attempts', String(delivery.external_attempt_count ?? 0)], ['latest delivery', `${delivery.latest_status || 'n/a'} @ ${delivery.latest_channel || 'n/a'}`], ['latest attempted', formatTimestamp(delivery.latest_attempted_ts)], diff --git a/tests/unit/test_alert_feedback_service.py b/tests/unit/test_alert_feedback_service.py index e25552f..095403b 100644 --- a/tests/unit/test_alert_feedback_service.py +++ b/tests/unit/test_alert_feedback_service.py @@ -312,6 +312,41 @@ class AlertVisualDispatchTests(unittest.TestCase): "suppressed_noise_policy", ) + def test_dispatcher_records_disabled_channel_when_default_adapter_has_no_credentials(self) -> None: + decision_repository = StubDecisionRepository() + decision_repository.alerts[12] = { + "id": 12, + "decision_packet_id": 55, + "alert_type": "candidate_blocked", + "status": "queued", + "payload": {"action_recommendation": "wait_for_confirmations", "status": "blocked"}, + } + decision_repository.packets[55] = {"id": 55, "payload": {"symbol": "BTC-USDT"}} + delivery_repository = StubDeliveryRepository() + dispatcher = AlertDispatcherService( + decision_repository=decision_repository, + delivery_repository=delivery_repository, + visual_watch_service=StubVisualWatchService(), + feedback_service=StubFeedbackService(), + ) + + with patch("src.services.alerts.channel_router_service.settings.alert_default_channel", "email"), patch( + "src.services.alerts.channel_router_service.settings.alert_email_smtp_url", + "", + ), patch("src.services.alerts.channel_router_service.settings.alert_email_to", ""), patch.object( + AlertDispatcherService, + "_send", + ) as send_mock: + result = dispatcher.dispatch_alert(12) + + send_mock.assert_not_called() + self.assertEqual(result["status"], "skipped") + self.assertEqual(result["reason"], "channel_disabled") + self.assertEqual(result["channel"], "email") + self.assertEqual(result["channel_route"]["capability"]["reason"], "credential_missing") + self.assertEqual(delivery_repository.delivery_calls[0]["error"], "channel_disabled") + self.assertEqual(delivery_repository.delivery_calls[0]["payload"]["channel_route"]["selected_channel"], "email") + def test_dispatch_pending_retries_failed_alert_after_backoff(self) -> None: decision_repository = StubDecisionRepository() decision_repository.alerts[21] = { diff --git a/tests/unit/test_alert_quality_summary_service.py b/tests/unit/test_alert_quality_summary_service.py index 193f30d..c3b0895 100644 --- a/tests/unit/test_alert_quality_summary_service.py +++ b/tests/unit/test_alert_quality_summary_service.py @@ -83,6 +83,10 @@ class AlertQualitySummaryServiceTests(unittest.TestCase): self.assertEqual(summary["delivery_summary"]["media_attached_count"], 2) self.assertEqual(summary["delivery_summary"]["media_linked_count"], 2) self.assertEqual(summary["delivery_summary"]["media_unavailable_count"], 1) + self.assertEqual(summary["delivery_summary"]["channel_status_counts"]["webhook"]["delivered"], 1) + self.assertEqual(summary["delivery_summary"]["channel_status_counts"]["webhook"]["failed"], 1) + self.assertEqual(summary["delivery_summary"]["channel_status_counts"]["workbench"]["skipped"], 2) + self.assertEqual(summary["delivery_summary"]["channel_disabled_count"], 0) self.assertTrue(summary["silent_window"]["active"]) self.assertEqual(summary["delivery_summary"]["success_rate"], 0.5) self.assertEqual(summary["alert_feedback_summary"]["positive_count"], 1) diff --git a/tests/unit/test_channel_router_service.py b/tests/unit/test_channel_router_service.py new file mode 100644 index 0000000..351e8e8 --- /dev/null +++ b/tests/unit/test_channel_router_service.py @@ -0,0 +1,47 @@ +import unittest +from unittest.mock import patch + +from src.services.alerts.channel_router_service import ( + build_channel_health_summary, + get_channel_capability, + select_alert_route, +) + + +class ChannelRouterServiceTests(unittest.TestCase): + def test_high_priority_alert_routes_to_webhook_when_configured(self) -> None: + alert = {"alert_type": "candidate_actionable", "payload": {"action_recommendation": "confirm_ticket"}} + + with patch("src.services.alerts.channel_router_service.settings.alert_webhook_url", "https://example.test/hook"): + route = select_alert_route(alert) + + self.assertEqual(route["selected_channel"], "webhook") + self.assertTrue(route["capability"]["enabled"]) + self.assertEqual(route["route_rule"], "webhook_high_priority_alert") + + def test_missing_wecom_credentials_surface_disabled_adapter(self) -> None: + with patch("src.services.alerts.channel_router_service.settings.alert_wecom_webhook_url", ""): + capability = get_channel_capability("wecom") + + self.assertFalse(capability["enabled"]) + self.assertEqual(capability["reason"], "credential_missing") + self.assertEqual(capability["adapter"], "wecom_disabled_adapter") + + def test_channel_health_reports_disabled_external_channels(self) -> None: + with patch("src.services.alerts.channel_router_service.settings.alert_webhook_url", ""), patch( + "src.services.alerts.channel_router_service.settings.alert_wecom_webhook_url", + "", + ), patch("src.services.alerts.channel_router_service.settings.alert_email_smtp_url", ""), patch( + "src.services.alerts.channel_router_service.settings.alert_email_to", + "", + ): + summary = build_channel_health_summary() + + self.assertIn("workbench", summary["enabled_channels"]) + self.assertIn("webhook", summary["credential_missing_channels"]) + self.assertIn("wecom", summary["credential_missing_channels"]) + self.assertIn("email", summary["credential_missing_channels"]) + + +if __name__ == "__main__": + unittest.main() diff --git a/tests/unit/test_trading_runtime_service.py b/tests/unit/test_trading_runtime_service.py index f0c9598..feea733 100644 --- a/tests/unit/test_trading_runtime_service.py +++ b/tests/unit/test_trading_runtime_service.py @@ -136,6 +136,9 @@ class FakeAlertDispatcherService: "media_attached": 0, "media_linked": 0, "media_unavailable": 0, + "channel_disabled": 0, + "channel_summary": {}, + "channel_health": {"enabled_channels": ["workbench"], "disabled_channels": ["webhook", "wecom", "email"]}, } @@ -243,6 +246,8 @@ class TradingRuntimeSchedulerTests(unittest.TestCase): self.assertFalse(payload["tasks"]["alert_dispatch"]["details"]["silent_window_active"]) self.assertEqual(payload["tasks"]["alert_dispatch"]["details"]["suppressed_noise_policy"], 0) self.assertEqual(payload["tasks"]["alert_dispatch"]["details"]["media_unavailable"], 0) + self.assertEqual(payload["tasks"]["alert_dispatch"]["details"]["channel_disabled"], 0) + self.assertIn("enabled_channels", payload["tasks"]["alert_dispatch"]["details"]["channel_health"]) self.assertEqual(payload["runtime_watchdog"]["stalled_task_count"], 0) self.assertEqual(payload["latest_broker_state_by_symbol"]["BTC-USDT"]["account"]["equity"], 1000.0) self.assertTrue(payload["tasks"]["watch"]["next_run_ts"])