flexible assest id handling

app.py

@@ -14,7 +14,7 @@ service = AnalysisService(config)
 async def analyze_sensor_data(data: SensorData):
     try:
         logger.info(f"Processing request with {len(data.vdc1)} voltage and {len(data.idc1)} current data points")
-        ml_output, agent_output = service.analyze(data.vdc1, data.idc1, data.api_key)
+        ml_output, agent_output = service.analyze(data.vdc1, data.idc1, data.api_key, data.asset_id)
         return AnalysisResponse(ml_output=ml_output, agent_output=agent_output)
     except HTTPException:
         raise

ml/inference.py

@@ -111,7 +111,7 @@ class MLEngine:
             "confidence": confidence
         }
 
-    def predict_from_raw(self, raw_df: pd.DataFrame):
+    def predict_from_raw(self, raw_df: pd.DataFrame, asset_id: str = None):
         logger.info("ML analysis start")
         df = build_features(raw_df, self.window)
         df = df[self.feature_cols].dropna()
@@ -122,9 +122,15 @@ class MLEngine:
         )
         anomaly_lstm, health = self._compute_anomalies(df_scaled)
         predictions = self._make_predictions(df_scaled, anomaly_lstm, health)
+        
+        # Use provided asset_id or generate default
+        if asset_id is None:
+            import uuid
+            asset_id = f"Solar_Panel_{str(uuid.uuid4())[:8]}"
+        
         logger.info("ML analysis end")
         return {
-            "asset_id": "PV_INVERTER_001",
+            "asset_id": asset_id,
             "failure_probability": round(predictions["failure_prob"], 2),
             "expected_ttf_days": round(predictions["ttf_days"], 1),
             "expected_rul_days": round(predictions["rul_days"], 1),

src/models.py

@@ -4,6 +4,7 @@ class SensorData(BaseModel):
     vdc1: list[float]
     idc1: list[float]
     api_key: str = None  # Optional Google API key for LLM features
+    asset_id: str = None  # Optional asset identifier
 
 class AnalysisResponse(BaseModel):
     ml_output: dict

src/services.py

@@ -12,12 +12,12 @@ class AnalysisService:
         self.config = config
         self.ml_engine = MLEngine()
 
-    def analyze(self, vdc1: list, idc1: list, api_key: str) -> tuple:
+    def analyze(self, vdc1: list, idc1: list, api_key: str, asset_id: str = None) -> tuple:
         """Analyze sensor data and return ML and agent outputs."""
         logger.info(f"Complete analysis start - processing {len(vdc1)} data points")
         validate_sensor_data(vdc1, idc1)
         raw_df = prepare_dataframe(vdc1, idc1)
-        ml_output = self.ml_engine.predict_from_raw(raw_df)
+        ml_output = self.ml_engine.predict_from_raw(raw_df, asset_id)
         agent_output = self.get_agent_output(api_key, ml_output)
         
         logger.info("Complete analysis end")