Upload runtime.py

runtime.py

@@ -1261,7 +1261,10 @@ def dashboard_payload() -> dict[str, Any]:
         _file_cache_key(TOMORROW_PREDICTION_HISTORY_PATH),
     )
     with _dashboard_payload_lock:
-        return copy.deepcopy(_dashboard_payload_cached(key))
+        payload = copy.deepcopy(_dashboard_payload_cached(key))
+    # Always rebuild track record (fresh Yahoo daily + calendar window), never serve from LRU cache.
+    payload["charts"]["track_record"] = build_prediction_track_record(sessions=10)
+    return payload
 
 
 def warm_dashboard_payload_cache() -> None:
@@ -1307,6 +1310,38 @@ def _load_forecaster_predictions_by_target() -> dict[date, dict[str, Any]]:
     return indexed
 
 
+def _load_track_record_daily_rows() -> pd.DataFrame:
+    """Daily OHLC for track record: live Yahoo pull merged over parquet history."""
+    frames: list[pd.DataFrame] = []
+    if NIFTY_1D_PATH.exists():
+        try:
+            frames.append(pd.read_parquet(NIFTY_1D_PATH))
+        except Exception:
+            pass
+    try:
+        yahoo_daily = fetch_yahoo_daily(period="3mo")
+        if not yahoo_daily.empty:
+            frames.append(yahoo_daily)
+            try:
+                append_parquet_rows(NIFTY_1D_PATH, yahoo_daily, ["date"])
+            except Exception:
+                pass
+    except Exception:
+        pass
+
+    if not frames:
+        return pd.DataFrame()
+
+    combined = pd.concat(frames, ignore_index=True)
+    combined["date"] = pd.to_datetime(combined["date"], errors="coerce").dt.normalize()
+    combined = combined.dropna(subset=["date"]).sort_values("date")
+    combined = combined.drop_duplicates(subset=["date"], keep="last")
+    combined = combined[
+        combined["close"].map(lambda value: np.isfinite(float(value)) if pd.notna(value) else False)
+    ].copy()
+    return combined.reset_index(drop=True)
+
+
 def _rolling_tomorrow_prediction(
     input_day: date,
     daily_rows: pd.DataFrame,
@@ -1330,20 +1365,7 @@ def build_prediction_track_record(
     fallback_prob = float(summary.get("latest_forecast_prob_up", 0.49900560447008563))
     forecaster_by_target = _load_forecaster_predictions_by_target()
 
-    end_session = _track_record_end_session()
-    latest_daily = latest_parquet_date(NIFTY_1D_PATH)
-    if latest_daily is None or latest_daily < end_session:
-        try:
-            refresh_daily_data()
-        except Exception:
-            pass
-
-    daily_rows = pd.read_parquet(NIFTY_1D_PATH)
-    daily_rows["date"] = pd.to_datetime(daily_rows["date"], errors="coerce").dt.normalize()
-    daily_rows = daily_rows.dropna(subset=["date"]).sort_values("date")
-    daily_rows = daily_rows[
-        daily_rows["close"].map(lambda value: np.isfinite(float(value)) if pd.notna(value) else False)
-    ].copy()
+    daily_rows = _load_track_record_daily_rows()
     if daily_rows.empty:
         return []
 
@@ -1353,6 +1375,16 @@ def build_prediction_track_record(
         if pd.notna(row["close"]) and np.isfinite(float(row["close"]))
     }
 
+    end_session = _track_record_end_session()
+    available_through = max(
+        (day for day in closes_by_date if day <= end_session),
+        default=None,
+    )
+    if available_through is None:
+        return []
+    if available_through < end_session:
+        end_session = available_through
+
     session_dates = last_n_trading_sessions(end_session, sessions)
 
     records: list[dict[str, Any]] = []
@@ -1485,7 +1517,6 @@ def _dashboard_payload_cached(key: tuple[tuple[str, int | None, int | None], ...
         "total_test_days": int(tomorrow_summary.get("n_test") or len(tomorrow_test) or 0),
         "models": model_metrics,
     }
-    track_record = build_prediction_track_record(sessions=10)
     return {
         "latest": t5_latest,
         "tomorrow_latest": tomorrow_latest,
@@ -1506,7 +1537,7 @@ def _dashboard_payload_cached(key: tuple[tuple[str, int | None, int | None], ...
             "tomorrow_recent_predictions": _json_ready_frame(tomorrow_recent),
             "tomorrow_history_predictions": _json_ready_frame(tomorrow_history.tail(80)),
             "tplus1_recent_predictions": _json_ready_frame(tplus1_test.tail(40)),
-            "track_record": track_record,
+            "track_record": [],
         },
         "data_status": {
             "nifty_1m_rows": int(len(pd.read_parquet(NIFTY_1M_PATH, columns=["date"]))),