Expose PChome rescore review metrics

2026-05-24 21:12:27 +08:00
parent 0fff09a15c
commit dcabebbcf2
15 changed files with 58 additions and 11 deletions
--- a/.env.example
+++ b/.env.example
@@ -74,6 +74,7 @@ ALERT_WEBHOOK_PASSWORD=your_secure_webhook_password_here
 AUTO_FIX_ENABLED=true

 # --- GitLab CI/CD ---
+GITLAB_ENABLED=false
 GITLAB_URL=http://192.168.0.110:8929
 GITLAB_TOKEN=your_gitlab_token_here
 GITLAB_PROJECT_ID=1
--- a/TODO_NEXT_STEPS.txt
+++ b/TODO_NEXT_STEPS.txt
@@ -4,6 +4,7 @@
 ================================================================================

 【已完成】
+   - V10.450 補 PChome 覆核 fast-count UI 語意與重算可採用指標：預設全量覆核頁跳過 exact count 時，模板會顯示「約」作為快取總數提示；搜尋、分類、單一狀態仍是精準總數。`fetch_competitor_coverage()` 同步輸出 `rescore_accepted_count`，讓 Dashboard、daily/growth 與 OpenClaw 摘要能把「重算可採用待審」從一般覆核隊列拆出來。
   - V10.449 修正 PChome 覆核 exact count 條件：只有預設「全部覆核、無搜尋、無分類」頁跳過 exact count；只要有搜尋詞、分類篩選或單一 review status，就保留精準總數，避免分頁資訊失準。
   - V10.448 讓 PChome 覆核「全部」頁跳過 exact count：`review_status=all` 使用 shared overview cache 的待處理總數作為分頁總數提示，只查當頁 50 筆；單一狀態分流仍保留 exact count，降低全量覆核頁互動成本。
   - V10.447 反轉 PChome 覆核頁查詢方向：review queue page 先從最新 `competitor_match_attempts` 的可覆核狀態縮小候選，再 join ACTIVE 商品與最新價，並用 `NOT EXISTS` 排除已有有效 identity_v2 正式價；避免每次「全部覆核」先掃全站 ACTIVE 商品。
--- a/config.py
+++ b/config.py
@@ -325,7 +325,7 @@ YOUTUBE_API_KEY = os.getenv('YOUTUBE_API_KEY', '')
 # ==========================================
 # 系統版本與路徑
 # ==========================================
-SYSTEM_VERSION = "V10.449"
+SYSTEM_VERSION = "V10.450"
 LOG_FILE_PATH = os.path.join(BASE_DIR, 'logs/system.log')
 public_url = PUBLIC_URL  # 用於模板顯示

--- a/docs/AI_INTELLIGENCE_MODULE_SOT.md
+++ b/docs/AI_INTELLIGENCE_MODULE_SOT.md
@@ -2,7 +2,7 @@

 > **最後更新**: 2026-05-24 (台北時間)
 > **狀態**: 🟢 四 AI Agent 自動化閉環已落地；LLM 路由紅線升級為 Ollama-first 三主機級聯，Gemini 備援預設關閉
-> **適用版本**: V10.449
+> **適用版本**: V10.450

 ---

@@ -81,6 +81,7 @@ SQL漏斗(~300筆)
 - 排程閉環：`run_pchome_match_backfill_task` 每日 10:30 執行，補抓 PChome 待比對商品、寫入歷史價格，再重算 `strategy='product_pick'` 清單。
 - PChome / MOMO 競價摘要出口 `services/competitor_intel_repository.py` 使用 30 分鐘共享快取（`COMPETITOR_INTEL_CACHE_TTL_SECONDS` 可調），避免 `/growth_analysis`、`/daily_sales`、PPT/AI 報表每次請求重跑昂貴覆蓋率與價差趨勢查詢；`run_competitor_price_feeder_task` 與 PChome backfill 完成後會主動清除快取。快取只包摘要輸出，不改 matcher 的高信心門檻與 identity_v2 準確性規則。
 - 商品看板第一屏：`/` 的 V2 看板直接以 `products`、`price_records`、`competitor_prices`、`competitor_match_attempts`、`competitor_match_reviews`、`ai_price_recommendations` 顯示比對覆蓋率、PChome 優勢、MOMO 威脅、AI 挑品、待比對優先清單與 PChome 覆核隊列；`filter=ai_picks` 可查看 50 品 AI 挑品列表，`filter=pchome_review` 可直接查看需人工處理的比價覆核 SKU，並以 DB 分頁支援 search/category/status 後的完整隊列，不得只截前 50 筆。覆核狀態篩選必須至少包含全部、需單位價、已排除、低信心、價格過期、找不到同款與人工閉環，讓人工可依 matcher 診斷類型分批處理。列內顯示候選 PChome 商品、候選價、match score、單位價換算摘要、人工動作與 matcher 診斷原因標籤（品牌不符、商品線不符、容量差異、組合差異、需單位價、價差極端等），不得只顯示籠統「待比對」。`/api/export/excel/pchome-review` 必須匯出同一套覆核隊列、人工處置、候選 PChome、單位價比較與原始診斷，讓人工覆核、簡報與後續 AI 分析共用同一份證據。`/api/pchome-review/<sku>/decision` 是人工閉環入口：`accept_identity` 才可把候選寫入 `competitor_prices` 與 `competitor_price_history` 並打上 `manual_review/manual_accept/identity_v2`；`reject_identity`、`unit_price_required` 與 `needs_research` 只寫 `competitor_match_reviews` 並追加 manual attempt，不得把不同販售組合或否決候選灌入正式價差。PChome feeder 後續搜尋同一候選時必須讀取 `competitor_match_reviews`：已否決候選寫 `manual_rejected` 並跳過正式寫入，且必須繼續評估下一個候選，不能讓已否決候選長期阻塞同 SKU；已標記單位價候選寫 `manual_unit_price_required`；已要求補搜尋候選寫 `manual_needs_research` 並停留在覆核隊列；已採用候選可保守補到最低門檻並保留 `manual_review/manual_accept` 標籤。搜尋候選池只有強同款分數達 `0.90` 才可提前停止，避免 0.76 灰區候選卡掉後續更精準搜尋詞。人工 `reject_identity`、`unit_price_required`、`needs_research` 若命中當前正式候選，必須將同候選 `competitor_prices` 過期，不得繼續顯示正式總價差。商品列表必須將 `manual_rejected`、`manual_unit_price_required`、`manual_needs_research` 顯示為明確人工閉環狀態，不可回落成籠統「待比對」。`fetch_competitor_coverage()` 必須輸出人工採用、人工否決、人工單位價與採用率，daily/growth/PPT 共用 payload 必須顯示人工閉環成效，避免只呈現待審數。商品看板深度快取同時寫入 `data/dashboard_full_cache.pkl`，供多個 Gunicorn worker 共用，避免部署後各 worker 重複重建 7,000+ 商品統計造成開頁變慢；所有資料異動與 AI 挑品重算都透過 `clear_dashboard_cache()` 同步清除記憶體與共享快取，手動重算 API 會立即預熱商品看板快取，避免第一位使用者承擔重建成本。
+- PChome re-score 回收線：`rescore_accepted_current` 只能表示最新版 matcher 判定「可人工採用」，不可直接寫入正式 `competitor_prices`；`fetch_competitor_coverage()` 必須輸出 `rescore_accepted_count`，Dashboard、daily/growth 與 OpenClaw 競品摘要都要把「重算可採用待審」獨立呈現，避免和一般低信心/單位價覆核混在一起。

 | 角色 | 模型 | 主機 | 成本 | 每日限額 |
 |------|------|------|------|---------|
--- a/docs/memory/code_modularization_inventory_20260430.md
+++ b/docs/memory/code_modularization_inventory_20260430.md
@@ -50,6 +50,7 @@
 - 2026-05-21 追記：同步 browse.sh 診斷計畫寫入 `competitor_match_attempts` 後的 `services/competitor_price_feeder.py` 行數；此處只更新 inventory，不變更模組化決策。
 - 2026-05-24 追記：同步背景 PChome 近門檻身份回收與 focused identity 系列更新後的 `services/marketplace_product_matcher.py` 行數；此處只更新 inventory，不變更商品比對行為。
 - 2026-05-24 追記：同步 111 fallback circuit breaker、NemoTron 決策信封與 Telegram template governance 後的 `run_scheduler.py`、`services/ollama_service.py`、`services/nemoton_dispatcher_service.py`、`services/telegram_templates.py` 行數；此處只更新 inventory，不變更模組化決策。
+- 2026-05-24 追記：同步 PChome 覆核頁 fast-count、輕量 render 與重算可採用指標後的 `routes/dashboard_routes.py` 行數；此處只更新 inventory，不變更 dashboard 行為。

 ## 達到或超過 800 行檔案清單

@@ -63,7 +64,7 @@
 | 3681 | `routes/admin_observability_routes.py` | P0 觀測台巨型 Blueprint | `services/observability_query_service.py` / `services/observability_action_service.py` / route glue |
 | 1796 | `routes/ai_routes.py` | P1 AI Blueprint | route glue / AI orchestration service / prompt builders |
 | 2154 | `services/nemoton_dispatcher_service.py` | P1 NemoTron service | NIM client / tool-call parser / action dispatcher |
-| 2026 | `routes/dashboard_routes.py` | P1 Dashboard Blueprint | competitor decision overview / dashboard query service；首頁資料整併需抽 service |
+| 2535 | `routes/dashboard_routes.py` | P1 Dashboard Blueprint | competitor decision overview / dashboard query service；首頁資料整併需抽 service |
 | 1485 | `routes/vendor_routes.py` | P1 Vendor Blueprint | route glue / stockout mutation/email；V2 page query、stockout list/batches API query、vendor list/detail query 已抽到 `services/vendor_stockout_query_service.py` |
 | 1390 | `services/telegram_bot_service.py` | P1 Telegram service | command handlers / message formatters / bot client |
 | 1237 | `app.py` | P1 bootstrap | 保持只做 app setup；繼續往 app_factory / extension setup 抽；Phase 42 只做 metadata table name 對齊 |
--- a/docs/memory/history_logs.md
+++ b/docs/memory/history_logs.md
@@ -13,6 +13,7 @@
 ## 📅 詳細更新日誌 (考古存檔)

 ### 2026-05-24：PChome 近門檻身份回收第二輪
+- **V10.450 PChome 覆核 fast-count UI 語意與重算可採用指標**: 預設全量覆核頁跳過 exact count 時，模板會以「約」標記快取總數，避免操作員把快取總數誤認為即時計算；搜尋、分類與單一狀態分流仍保留精準總數。`fetch_competitor_coverage()` 同步輸出 `rescore_accepted_count`，Dashboard、daily/growth 與 OpenClaw 摘要會把「重算可採用待審」獨立顯示，不再只混在一般覆核隊列。
 - **V10.449 PChome 覆核 exact count 條件修正**: 只有預設「全部覆核、無搜尋、無分類」頁跳過 exact count；若使用搜尋詞、分類篩選或單一 review status，仍保留精準總數，避免操作員分頁資訊失準。
 - **V10.448 PChome 覆核全量頁跳過 exact count**: `review_status=all` 改用 shared overview cache 的待處理總數作為分頁總數提示，當頁只查 50 筆；單一狀態分流仍保留 exact count，避免每次操作全量覆核頁都為總筆數重掃整個 review queue。
 - **V10.447 PChome 覆核頁查詢方向反轉**: review queue page 改由最新 `competitor_match_attempts` 的可覆核狀態先縮小候選，再 join ACTIVE 商品與最新價，並用 `NOT EXISTS` 排除已有有效 `identity_v2` 正式 PChome 價格；避免「全部覆核」每次先掃全站 ACTIVE 商品後才過濾，提高核心比價覆核頁可操作性。
--- a/routes/dashboard_routes.py
+++ b/routes/dashboard_routes.py
@@ -676,6 +676,11 @@ def _merge_competitor_review_context(overview, review_context):
    overview.update({
        'review_queue_count': int(coverage.get('actionable_review_count') or len(review_queue) or 0),
        'unit_comparable_count': int(coverage.get('unit_comparable_count') or 0),
+        'rescore_accepted_count': int(
+            coverage.get('rescore_accepted_count')
+            or review_status_counts.get('rescore_accepted')
+            or 0
+        ),
        'review_status_counts': review_status_counts,
        'review_queue': review_queue[:3],
    })
@@ -819,6 +824,7 @@ def _load_competitor_decision_overview(session, latest_items=None):
        'pending_priority': [],
        'review_queue_count': 0,
        'unit_comparable_count': 0,
+        'rescore_accepted_count': 0,
        'review_queue': [],
    }

@@ -1688,6 +1694,7 @@ def _load_cached_competitor_overview_for_review(now_taipei, review_queue, review
    if not overview.get('review_queue'):
        overview['review_queue'] = list(review_queue[:3])
    overview.setdefault('unit_comparable_count', 0)
+    overview.setdefault('rescore_accepted_count', 0)
    return overview


@@ -1727,6 +1734,7 @@ def _render_pchome_review_dashboard(
    )
    review_queue = review_page.get('items') or []
    review_queue_total = int(review_page.get('total') or 0)
+    review_total_is_estimated = False
    if review_queue_total < 0:
        review_queue_total = int(
            (overview_hint.get('review_status_counts') or {}).get('all')
@@ -1734,6 +1742,7 @@ def _render_pchome_review_dashboard(
            or len(review_queue)
            or 0
        )
+        review_total_is_estimated = True
    review_queue_map = {
        str(row.get('sku') or ''): row
        for row in review_queue
@@ -1774,6 +1783,7 @@ def _render_pchome_review_dashboard(
        current_category=category_filter,
        current_filter=filter_type,
        current_review_status=review_status,
+        review_total_is_estimated=review_total_is_estimated,
        review_status_options=review_status_options,
        search_query=search_query,
        current_sort=sort_by,
--- a/services/competitor_intel_repository.py
+++ b/services/competitor_intel_repository.py
@@ -403,7 +403,7 @@ def _cached_payload(cache_key: str, producer, ttl_seconds: int = COMPETITOR_INTE

 def fetch_competitor_coverage(engine) -> dict:
    return _cached_payload(
-        f"coverage:v4:floor={PCHOME_MATCH_SCORE_FLOOR}:manual_reviews=1",
+        f"coverage:v5:floor={PCHOME_MATCH_SCORE_FLOOR}:manual_reviews=1:rescore=1",
        lambda: _fetch_competitor_coverage_uncached(engine),
    )

@@ -422,6 +422,7 @@ def _fetch_competitor_coverage_uncached(engine) -> dict:
            "match_rate": 0,
            "attempt_status": {},
            "unit_comparable_count": 0,
+            "rescore_accepted_count": 0,
            "actionable_review_count": 0,
            "manual_review_summary": manual_review_summary,
            "manual_review_total": manual_review_summary["total"],
@@ -509,6 +510,7 @@ def _fetch_competitor_coverage_uncached(engine) -> dict:
        for row in rows
    }
    unit_count = sum(statuses.get(status, 0) for status in UNIT_COMPARABLE_STATUSES)
+    rescore_accepted_count = int(statuses.get("rescore_accepted_current") or 0)
    actionable_count = sum(statuses.get(status, 0) for status in ACTIONABLE_ATTEMPT_STATUSES)
    return {
        "active_with_price": active,
@@ -517,6 +519,7 @@ def _fetch_competitor_coverage_uncached(engine) -> dict:
        "match_rate": round(valid / max(active, 1) * 100, 1),
        "attempt_status": statuses,
        "unit_comparable_count": unit_count,
+        "rescore_accepted_count": rescore_accepted_count,
        "actionable_review_count": actionable_count,
        "manual_review_summary": manual_review_summary,
        "manual_review_total": manual_review_summary["total"],
--- a/services/openclaw_strategist_service.py
+++ b/services/openclaw_strategist_service.py
@@ -592,7 +592,8 @@ def _fetch_competitor_summary() -> Dict[str, Any]:
                    )
                    SELECT
                        SUM(CASE WHEN attempt_status IN ('unit_comparable', 'refresh_unit_comparable') THEN 1 ELSE 0 END) AS unit_comparable_count,
-                        SUM(CASE WHEN attempt_status IN ('unit_comparable', 'refresh_unit_comparable', 'identity_veto', 'low_score', 'refresh_low_score', 'recoverable_low_score', 'true_low_confidence', 'protected_existing_match', 'expired_match', 'no_result', 'refresh_no_result') THEN 1 ELSE 0 END) AS review_queue_count
+                        SUM(CASE WHEN attempt_status = 'rescore_accepted_current' THEN 1 ELSE 0 END) AS rescore_accepted_count,
+                        SUM(CASE WHEN attempt_status IN ('rescore_accepted_current', 'unit_comparable', 'refresh_unit_comparable', 'identity_veto', 'low_score', 'refresh_low_score', 'recoverable_low_score', 'true_low_confidence', 'protected_existing_match', 'expired_match', 'no_result', 'refresh_no_result') THEN 1 ELSE 0 END) AS review_queue_count
                    FROM latest_attempt
                """)).fetchone()
            return {
@@ -601,7 +602,8 @@ def _fetch_competitor_summary() -> Dict[str, Any]:
                "undercut_count": int(row[2] or 0),
                "premium_count": int(row[3] or 0),
                "unit_comparable_count": int((attempt_row[0] if attempt_row else 0) or 0),
-                "review_queue_count": int((attempt_row[1] if attempt_row else 0) or 0),
+                "rescore_accepted_count": int((attempt_row[1] if attempt_row else 0) or 0),
+                "review_queue_count": int((attempt_row[2] if attempt_row else 0) or 0),
            }
        return {}
    except Exception as e:
@@ -1487,6 +1489,7 @@ def generate_weekly_strategy_report(
  被競品削價數：{competitor_summary.get('undercut_count', 0)} 個
  我方具優勢數：{competitor_summary.get('premium_count', 0)} 個
  需單位價覆核：{competitor_summary.get('unit_comparable_count', 0)} 個
+  重算可採用待審：{competitor_summary.get('rescore_accepted_count', 0)} 個

 TOP 威脅品項（近48h Hermes 偵測）：
 {_format_threats(threats)}
@@ -1767,6 +1770,7 @@ def _legacy_full_gemini_daily_report() -> dict:
  被削價風險：{competitor_summary.get('undercut_count', 0)} 個（價差超過10%）
  平均價差：{competitor_summary.get('avg_gap_pct', 0):+.1f}%
  單位價/身份覆核隊列：{competitor_summary.get('review_queue_count', 0)} 個
+  重算可採用待審：{competitor_summary.get('rescore_accepted_count', 0)} 個

 請按以下結構輸出（使用 HTML <b> 標題）：

@@ -1944,6 +1948,7 @@ def generate_monthly_report() -> dict:
  月均價差：{competitor_summary.get('avg_gap_pct', 0):+.1f}%
  被削價風險SKU：{competitor_summary.get('undercut_count', 0)} 個
  需單位價覆核SKU：{competitor_summary.get('unit_comparable_count', 0)} 個
+  重算可採用待審SKU：{competitor_summary.get('rescore_accepted_count', 0)} 個

 【價格變動概況】
  本月調價次數：{price_trend_data.get('price_changes', 0)} 次
--- a/templates/daily_sales.html
+++ b/templates/daily_sales.html
@@ -359,6 +359,10 @@
              <span>需單位價覆核</span>
              <strong class="momo-mono">{{ comp_coverage.unit_comparable_count | default(0) | number_format }}</strong>
            </div>
+            <div>
+              <span>重算可採用待審</span>
+              <strong class="momo-mono">{{ comp_coverage.rescore_accepted_count | default(0) | number_format }}</strong>
+            </div>
            <div>
              <span>人工採用</span>
              <strong class="momo-mono">{{ comp_coverage.manual_accept_count | default(0) | number_format }}</strong>
--- a/templates/dashboard_v2.html
+++ b/templates/dashboard_v2.html
@@ -42,7 +42,8 @@
                <div class="dashboard-kpi-label momo-mono">比價覆核</div>
                <div class="dashboard-kpi-value momo-mono is-warning">{{ overview.review_queue_count | default(0) | number_format }}</div>
                <div class="dashboard-kpi-sub momo-mono">
-                    <a class="dashboard-kpi-sub-link" href="{{ url_for('dashboard.index', filter='pchome_review', category=current_category, q=search_query, sort_by='pchome_review', order='desc') }}">需單位價覆核 {{ overview.unit_comparable_count | default(0) | number_format }}</a>
+                    <a class="dashboard-kpi-sub-link" href="{{ url_for('dashboard.index', filter='pchome_review', category=current_category, q=search_query, review_status='rescore_accepted', sort_by='pchome_review', order='desc') }}">重算可採用 {{ overview.rescore_accepted_count | default(0) | number_format }}</a>
+                    · <a class="dashboard-kpi-sub-link" href="{{ url_for('dashboard.index', filter='pchome_review', category=current_category, q=search_query, review_status='unit_comparable', sort_by='pchome_review', order='desc') }}">需單位價 {{ overview.unit_comparable_count | default(0) | number_format }}</a>
                    · 待補抓 {{ overview.pending_match_count | default(0) | number_format }}
                </div>
            </div>
@@ -61,7 +62,7 @@
                <div class="dashboard-backfill-label momo-mono">PCHOME MATCH BACKFILL</div>
                <div class="dashboard-backfill-title">待比對補抓產線</div>
                <div class="dashboard-backfill-meta momo-mono">
-                    待補抓 {{ overview.pending_match_count | default(0) | number_format }} · 覆核 {{ overview.review_queue_count | default(0) | number_format }} · 單位價 {{ overview.unit_comparable_count | default(0) | number_format }}
+                    待補抓 {{ overview.pending_match_count | default(0) | number_format }} · 覆核 {{ overview.review_queue_count | default(0) | number_format }} · 重算可採用 {{ overview.rescore_accepted_count | default(0) | number_format }} · 單位價 {{ overview.unit_comparable_count | default(0) | number_format }}
                </div>
            </div>
            <div class="dashboard-backfill-progress" aria-hidden="true">
@@ -300,7 +301,7 @@
                    {% if current_filter == 'ai_picks' %}
                        {{ total_items | number_format }} / {{ ai_pick_list_limit }} 品
                    {% elif current_filter == 'pchome_review' %}
-                        {{ total_items | number_format }} / {{ overview.review_queue_count | default(0) | number_format }} 待處理
+                        {% if review_total_is_estimated %}約 {% endif %}{{ total_items | number_format }} / {{ overview.review_queue_count | default(0) | number_format }} 待處理
                    {% else %}
                        {{ total_items | number_format }} 筆
                    {% endif %}
--- a/templates/growth_analysis.html
+++ b/templates/growth_analysis.html
@@ -149,6 +149,8 @@
            <strong class="momo-mono">{{ coverage.pending | default(0) | number_format }}</strong>
            <span>需單位價覆核</span>
            <strong class="momo-mono">{{ coverage.unit_comparable_count | default(0) | number_format }}</strong>
+            <span>重算可採用待審</span>
+            <strong class="momo-mono">{{ coverage.rescore_accepted_count | default(0) | number_format }}</strong>
            <span>人工採用</span>
            <strong class="momo-mono">{{ coverage.manual_accept_count | default(0) | number_format }}</strong>
            <span>人工否決</span>
--- a/tests/test_competitor_intel_cache.py
+++ b/tests/test_competitor_intel_cache.py
@@ -75,7 +75,8 @@ def test_competitor_coverage_counts_only_active_product_intersection():
        "def _fetch_manual_review_summary", 1
    )[0]

-    assert "coverage:v4" in source
+    assert "coverage:v5" in source
+    assert "rescore_accepted_count" in coverage_source
    assert "(SELECT COUNT(*) FROM valid_competitor) AS valid_matches" not in coverage_source
    assert "FROM latest_momo lm\n             JOIN valid_competitor vc ON vc.sku = lm.sku" in coverage_source
    assert "FROM products p\n            JOIN LATERAL" in coverage_source
@@ -101,6 +102,7 @@ def test_competitor_review_queue_is_canonical_unit_price_handoff():
    assert "def fetch_competitor_review_queue" in source
    assert "\"review_queue\": fetch_competitor_review_queue" in source
    assert "\"unit_comparable_count\"" in source
+    assert "\"rescore_accepted_count\"" in source
    assert "manual_review_summary" in source
    assert "manual_accept_count" in source
    assert "manual_reject_count" in source
@@ -119,9 +121,11 @@ def test_competitor_review_queue_is_canonical_unit_price_handoff():
    assert "人工單位價" in daily_template
    assert "competitor_intel.review_queue" in daily_template
    assert "coverage.unit_comparable_count" in growth_template
+    assert "coverage.rescore_accepted_count" in growth_template
    assert "coverage.manual_accept_count" in growth_template
    assert "coverage.manual_reject_count" in growth_template
    assert "coverage.manual_unit_price_count" in growth_template
+    assert "comp_coverage.rescore_accepted_count" in daily_template


 def test_competitor_review_reasons_prefer_json_payload_labels():
--- a/tests/test_frontend_v2_assets.py
+++ b/tests/test_frontend_v2_assets.py
@@ -142,6 +142,8 @@ def test_dashboard_v2_is_production_default_and_uses_real_dashboard_data():
    assert "or bool(search_query)" in route_source
    assert "or bool(_normalize_dashboard_category_filter(category_filter))" in route_source
    assert "count_total=count_total" in route_source
+    assert "review_total_is_estimated = True" in route_source
+    assert "review_total_is_estimated=review_total_is_estimated" in route_source
    assert "只替 PChome 覆核當頁建立商品列" in route_source
    assert "_load_competitor_decision_overview(session, unique_items)" in route_source
    assert "item_map = {}" in route_source
@@ -150,6 +152,7 @@ def test_dashboard_v2_is_production_default_and_uses_real_dashboard_data():
    assert "pending_match_count" in route_source
    assert "review_queue_count" in route_source
    assert "unit_comparable_count" in route_source
+    assert "rescore_accepted_count" in route_source
    assert "filter_type == 'pchome_review'" in route_source
    assert "total_items = review_queue_total" in route_source
    assert "REVIEW_STATUS_OPTIONS" in route_source
@@ -169,7 +172,9 @@ def test_dashboard_v2_is_production_default_and_uses_real_dashboard_data():
    assert "overview.top_momo_threats" in dashboard
    assert "overview.pending_priority" in dashboard
    assert "overview.review_queue" in dashboard
-    assert "需單位價覆核" in dashboard
+    assert "需單位價 {{ overview.unit_comparable_count" in dashboard
+    assert "重算可採用 {{ overview.rescore_accepted_count" in dashboard
+    assert "{% if review_total_is_estimated %}約 {% endif %}" in dashboard
    assert "filter='ai_picks'" in dashboard
    assert "filter='pchome_review'" in dashboard
    assert "review_status=option.key" in dashboard
--- a/tests/test_openclaw_daily_template.py
+++ b/tests/test_openclaw_daily_template.py
@@ -19,6 +19,7 @@ Operation Ollama-First v5.0 / Phase 3 / A8 — 日報模板路由測試
 import os
 import logging
 from datetime import date, datetime
+from pathlib import Path
 from unittest.mock import patch, MagicMock

 import pytest
@@ -150,6 +151,13 @@ class TestKPIComputation:
        assert threats[0]["gap_pct"] == 0
        assert "metadata_json decode failed" in caplog.text

+    def test_competitor_summary_tracks_rescore_review_bucket(self):
+        source = (Path(__file__).resolve().parents[1] / "services/openclaw_strategist_service.py").read_text(encoding="utf-8")
+
+        assert "rescore_accepted_count" in source
+        assert "attempt_status = 'rescore_accepted_current'" in source
+        assert "重算可採用待審" in source
+

 # ═══════════════════════════════════════════════════════════════════════════
 # T4+T5 — Template 渲染與缺欄位優雅降級