Cleanup: MB 2.0 Gap Analysis előtti állapot (adatok kizárva)

2026-02-23 09:44:02 +01:00
parent 5757754aae
commit 893f39fa15
74 changed files with 34239 additions and 2834 deletions
--- a/backend/app/workers/researcher_v2_1.py
+++ b/backend/app/workers/researcher_v2_1.py
@@ -0,0 +1,117 @@
+import asyncio
+import logging
+import warnings
+import os
+from sqlalchemy import select, update, and_, func, or_, case # Explicit case import
+from app.db.session import SessionLocal
+from app.models.vehicle_definitions import VehicleModelDefinition
+import httpx
+
+# 1. KRITIKUS JAVÍTÁS: A figyelmeztetések globális elnyomása az import előtt
+warnings.filterwarnings("ignore", category=RuntimeWarning, module='duckduckgo_search')
+from duckduckgo_search import DDGS
+
+# Logolás beállítása, hogy lássuk a haladást
+logging.basicConfig(level=logging.INFO, format='%(asctime)s [%(levelname)s] %(name)s: %(message)s')
+logger = logging.getLogger("Robot-Researcher-v2.1")
+
+class ResearcherBot:
+    def __init__(self):
+        self.batch_size = 15 
+        self.max_parallel_queries = 5
+
+    async def fetch_source(self, label, query):
+        """Egyedi forrás lekérése a DuckDuckGo-tól."""
+        try:
+            def search():
+                # Az újabb verziókban a DDGS() hívás így a legstabilabb
+                with DDGS() as ddgs:
+                    results = ddgs.text(query, max_results=3)
+                    return [r['body'] for r in results] if results else []
+            
+            results = await asyncio.to_thread(search)
+            
+            if not results:
+                return f"=== SOURCE: {label} | NO DATA FOUND ===\n\n"
+
+            content = f"=== SOURCE: {label} | QUERY: {query} ===\n"
+            content += "\n---\n".join(results)
+            content += "\n=== END SOURCE ===\n\n"
+            return content
+        except Exception as e:
+            logger.error(f"❌ Keresési hiba ({label}): {e}")
+            return f"=== SOURCE: {label} ERROR: {str(e)} ===\n\n"
+
+    async def research_vehicle(self, vehicle_id):
+        async with SessionLocal() as db:
+            res = await db.execute(select(VehicleModelDefinition).where(VehicleModelDefinition.id == vehicle_id))
+            v = res.scalar_one_or_none()
+            if not v: return
+            
+            make, model = v.make, v.marketing_name
+            # Jelöljük be, hogy a kutatás folyamatban van
+            await db.execute(update(VehicleModelDefinition).where(VehicleModelDefinition.id == vehicle_id).values(status='research_in_progress'))
+            await db.commit()
+
+        logger.info(f"🔎 Kutatás indul: {make} {model}")
+        
+        queries = [
+            ("TECH_SPECS", f"{make} {model} technical specifications engine power"),
+            ("MAINTENANCE", f"{make} {model} service manual oil capacity spark plug"),
+            ("TIRES_BRAKES", f"{make} {model} tire size brake pad type"),
+            ("FLUIDS", f"{make} {model} coolant quantity transmission oil")
+        ]
+        
+        tasks = [self.fetch_source(label, q) for label, q in queries]
+        search_results = await asyncio.gather(*tasks)
+        
+        full_context = "".join(search_results)
+        
+        async with SessionLocal() as db:
+            await db.execute(
+                update(VehicleModelDefinition)
+                .where(VehicleModelDefinition.id == vehicle_id)
+                .values(
+                    raw_search_context=full_context,
+                    status='awaiting_ai_synthesis', # Itt adjuk át a Robot 2.2-nek (Alchemist)
+                    updated_at=func.now()
+                )
+            )
+            await db.commit()
+        logger.info(f"✅ Kutatás kész, adat a tartályban: {make} {model}")
+
+    async def run(self):
+        logger.info("🚀 Robot 2.1 (Researcher) ONLINE")
+        while True:
+            async with SessionLocal() as db:
+                # 2. KRITIKUS JAVÍTÁS: func.case helyett az explicit case() használata
+                # Ez javítja a "TypeError: got an unexpected keyword argument 'else_'" hibát
+                priorities = case(
+                    (and_(VehicleModelDefinition.vehicle_type == 'car', 
+                          VehicleModelDefinition.make.in_(['SUZUKI', 'TOYOTA', 'SKODA', 'VOLKSWAGEN', 'OPEL'])), 1),
+                    (VehicleModelDefinition.vehicle_type == 'car', 2),
+                    (and_(VehicleModelDefinition.vehicle_type == 'motorcycle', 
+                          VehicleModelDefinition.make.in_(['HONDA', 'YAMAHA', 'SUZUKI', 'KAWASAKI'])), 3),
+                    else_=4
+                )
+
+                stmt = select(VehicleModelDefinition.id).where(
+                    or_(VehicleModelDefinition.status == 'unverified', VehicleModelDefinition.status == 'awaiting_research')
+                ).order_by(priorities).limit(self.batch_size)
+                
+                res = await db.execute(stmt)
+                ids = [r[0] for r in res.fetchall()]
+
+            if not ids:
+                logger.info("💤 Nincs több feldolgozandó feladat, pihenés...")
+                await asyncio.sleep(60)
+                continue
+
+            # Batch feldolgozás indítása párhuzamosan
+            await asyncio.gather(*[self.research_vehicle(rid) for rid in ids])
+            
+            # Rövid szünet a keresőmotorok kímélése érdekében
+            await asyncio.sleep(2) 
+
+if __name__ == "__main__":
+    asyncio.run(ResearcherBot().run())