feat: Robot ecosystem v1.2.6 - Google Search RAG & Master-Merge logic stabilized

2026-02-17 22:44:57 +00:00
parent 2def6b2201
commit b11b9bce87
25 changed files with 3192 additions and 789 deletions
--- a/backend/app/workers/technical_enricher.py.bak
+++ b/backend/app/workers/technical_enricher.py.bak
@@ -0,0 +1,102 @@
+import asyncio
+import httpx
+import logging
+import os  # <--- EZ HIÁNYZOTT!
+import datetime
+from sqlalchemy import select
+from sqlalchemy.exc import IntegrityError
+from app.db.session import SessionLocal
+from app.models.vehicle_definitions import VehicleModelDefinition
+from app.services.ai_service import AIService
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger("Robot-v1.2.4-Fixed")
+
+class TechEnricher:
+    API_URL = "https://opendata.rdw.nl/resource/kyri-nuah.json"
+    RDW_TOKEN = os.getenv("RDW_APP_TOKEN")
+    HEADERS = {"X-App-Token": RDW_TOKEN} if RDW_TOKEN else {}
+
+    @classmethod
+    async def fetch_rdw_tech_data(cls, make, model):
+        """Hibatűrő RDW lekérdezés tisztított paraméterekkel."""
+        clean_model = str(model).strip().upper()
+        params = {"merk": make.upper(), "handelsbenaming": clean_model, "$limit": 1}
+        async with httpx.AsyncClient(headers=cls.HEADERS) as client:
+            try:
+                resp = await client.get(cls.API_URL, params=params, timeout=15)
+                if resp.status_code == 200 and resp.json():
+                    return resp.json()[0]
+                return None
+            except Exception:
+                return None
+
+    @classmethod
+    async def run(cls):
+        logger.info("🚀 Master Enricher INDUL (Atomi mentés üzemmód)...")
+        
+        # 1. Csak az ID-kat kérjük le, hogy ne tartsuk nyitva a tranzakciót feleslegesen
+        async with SessionLocal() as main_db:
+            stmt = select(VehicleModelDefinition.id).where(
+                VehicleModelDefinition.status == "unverified"
+            ).limit(50)
+            res = await main_db.execute(stmt)
+            ids = res.scalars().all()
+
+        if not ids:
+            logger.info("😴 Nincs dúsítandó adat.")
+            return
+
+        # 2. Egyesével dolgozzuk fel a rekordokat saját session-ben
+        for m_id in ids:
+            async with SessionLocal() as db:
+                try:
+                    master = await db.get(VehicleModelDefinition, m_id)
+                    if not master:
+                        continue
+
+                    logger.info(f"🧪 Feldolgozás: {master.make} {master.marketing_name} (ID: {m_id})")
+                    data_found = False
+
+                    # A: RDW fázis
+                    rdw_data = await cls.fetch_rdw_tech_data(master.make, master.marketing_name)
+                    if rdw_data:
+                        master.engine_capacity = int(float(rdw_data.get("cilinderinhoud", 0))) or None
+                        master.power_kw = int(float(rdw_data.get("netto_maximum_vermogen_kw", 0))) or None
+                        data_found = True
+
+                    # B: AI fázis (ha hiányzik adat vagy pontosítani kell)
+                    if not data_found or master.engine_capacity is None:
+                        ai_data = await AIService.get_clean_vehicle_data(
+                            master.make, master.marketing_name, master.vehicle_type
+                        )
+                        if ai_data:
+                            master.marketing_name = ai_data.get("marketing_name", master.marketing_name)
+                            master.technical_code = ai_data.get("technical_code") or master.technical_code or "N/A"
+                            master.engine_capacity = ai_data.get("ccm") or master.engine_capacity
+                            master.power_kw = ai_data.get("kw") or master.power_kw
+                            master.specifications = ai_data.get("maintenance", {})
+                            data_found = True
+
+                    # C: Mentés és véglegesítés
+                    if data_found:
+                        master.status = "ai_enriched"
+                        master.updated_at = datetime.datetime.now()
+                        await db.commit() # AZONNALI COMMIT A LEMEZRE
+                        logger.info(f"✅ Sikeresen mentve: {master.marketing_name} (CCM: {master.engine_capacity})")
+                    else:
+                        logger.warning(f"⚠️ Nem találtam adatot az ID {m_id} esetében.")
+
+                except IntegrityError:
+                    await db.rollback()
+                    logger.warning(f"🚫 Duplikáció vagy Constraint hiba (ID: {m_id}). Kihagyva.")
+                except Exception as e:
+                    await db.rollback()
+                    logger.error(f"❌ Váratlan hiba az ID {m_id} esetében: {e}")
+                finally:
+                    await db.close()
+
+        logger.info("🏁 50-es batch feldolgozva.")
+
+if __name__ == "__main__":
+    asyncio.run(TechEnricher.run())