Update preprocessing.py

preprocessing.py

@@ -26,6 +26,7 @@ def prepare_miner_features(
     power=None,
     efficiency=None,
     electricity_rate=None,
+    release_date=None
 ):
     """
     Add miner-specific features - EXACTLY 14 features.
@@ -73,9 +74,16 @@ def prepare_miner_features(
     else:
         df["efficiency"] = specs["efficiency"]
 
-    # Calculate age_days (days since miner was released)
-    release_date = pd.to_datetime(specs["release_date"])
-    df["age_days"] = (df["date"] - release_date).dt.days
+    # ----- Age of the machine in days -----
+    # If the user supplied a release_date, use that; otherwise fall back
+    # to the miner spec's release_date (S19 Pro default).
+    if release_date is not None and str(release_date).strip() != "":
+        release_str = str(release_date).strip()
+    else:
+        release_str = specs.get("release_date", "2020-01-01")
+
+    release_dt = pd.to_datetime(release_str)
+    df["age_days"] = (df["date"] - release_dt).dt.days
 
     # Days since halving
     df["days_since_halving"] = df["date"].apply(get_days_since_halving)
@@ -114,6 +122,7 @@ def get_latest_sequence(
     power=None,
     efficiency=None,
     electricity_rate=None,
+    release_date=None
 ):
     """
     Get the most recent sequence for prediction - EXACTLY 14 features in CORRECT ORDER.
@@ -134,6 +143,7 @@ def get_latest_sequence(
         power=power,
         efficiency=efficiency,
         electricity_rate=electricity_rate,
+        release_date=release_date,
     )
 
     # CRITICAL: This order MUST match your training data CSV exactly!