refactor: optimize server setup (dev/prod mode), fix hybrid search columns, debug query logging

backend/.gitignore

@@ -2,3 +2,4 @@
 # Ignore embedded repo
 preference/
 
+query.txt

backend/Dockerfile

-# Python 3.11 slim (nhẹ, ít file)
 FROM python:3.11-slim
 
-# Thư mục làm việc
 WORKDIR /app
 
 ENV PYTHONUNBUFFERED=1 PYTHONDONTWRITEBYTECODE=1
 
-# Copy requirements rồi cài package
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy code
 COPY . .
 
-# Mở port (sẽ được override bởi docker-compose)
 EXPOSE 5004
 
-# Chạy uvicorn với auto workers (2 * CPU cores + 1)
-CMD ["sh", "-c", "uvicorn server:app --host 0.0.0.0 --port ${PORT:-5004} --workers $(( $(nproc) * 2 + 1 ))"]
\ No newline at end of file
+CMD ["python", "server.py"]
\ No newline at end of file

backend/config.py

@@ -37,6 +37,7 @@ __all__ = [
     "MONGODB_URI",
     "OPENAI_API_KEY",
     "PORT",
+    "ENV_MODE",
     "REDIS_HOST",
     "REDIS_PASSWORD",
     "REDIS_PORT",
@@ -74,6 +75,7 @@ JWT_ALGORITHM: str | None = os.getenv("JWT_ALGORITHM")
 
 # ====================== SERVER CONFIG ======================
 PORT: int = int(os.getenv("PORT", "5004"))
+ENV_MODE: str = os.getenv("ENV_MODE", "dev").lower()  # dev or prod
 FIRECRAWL_API_KEY: str | None = os.getenv("FIRECRAWL_API_KEY")
 
 # ====================== LANGFUSE CONFIGURATION (DEPRECATED) ======================

backend/search/hybrid_search.py

@@ -127,7 +127,6 @@ async def hybrid_search(
                 /*+ SET_VAR(ann_params='{{"ef_search":{ef_search}}}') */
                 internal_ref_code,
                 product_name,
-                description_text,
                 product_image_url,
                 product_web_url,
                 sale_price,
@@ -147,18 +146,12 @@ async def hybrid_search(
         SELECT 
             internal_ref_code,
             MAX_BY(product_name, similarity_score) as product_name,
-            MAX_BY(description_text, similarity_score) as description_text,
             MAX_BY(product_image_url, similarity_score) as product_image_url,
             MAX_BY(product_web_url, similarity_score) as product_web_url,
             MAX_BY(sale_price, similarity_score) as sale_price,
             MAX_BY(original_price, similarity_score) as original_price,
             MAX_BY(discount_amount, similarity_score) as discount_amount,
             GROUP_CONCAT(DISTINCT master_color ORDER BY master_color SEPARATOR ', ') as available_colors,
-            MAX_BY(season, similarity_score) as season,
-            MAX_BY(gender_by_product, similarity_score) as gender_by_product,
-            MAX_BY(age_by_product, similarity_score) as age_by_product,
-            MAX_BY(product_line_vn, similarity_score) as product_line_vn,
-            MAX_BY(style, similarity_score) as style,
             MAX(similarity_score) as similarity_score
         FROM semantic_candidates
         {where_clause}
@@ -169,6 +162,20 @@ async def hybrid_search(
         
         logger.info(f"SQL Query:\n{sql}")
 
+        # Ghi query ra file query.txt ở thư mục backend (parent của thư mục search)
+        try:
+            import os
+            # backend/search/hybrid_search.py -> backend/query.txt
+            current_dir = os.path.dirname(os.path.abspath(__file__))
+            backend_dir = os.path.dirname(current_dir)
+            query_file_path = os.path.join(backend_dir, "query.txt")
+            
+            with open(query_file_path, "w", encoding="utf-8") as f:
+                f.write(sql)
+            logger.info(f"Query saved to: {query_file_path}")
+        except Exception as e:
+            logger.error(f"Failed to write query to file: {e}")
+        
         # ═══════════════════════════════════════════════════════════════════
         # STEP 6: Execute query
         # ═══════════════════════════════════════════════════════════════════
@@ -230,7 +237,6 @@ async def semantic_only_search(
         season,
         gender_by_product,
         age_by_product,
-        product_line_vn,
         approx_cosine_similarity(vector, {v_str}) as similarity_score
     FROM shared_source.magento_product_dimension_with_text_embedding
     ORDER BY similarity_score DESC

backend/server.py

 """
-FastAPI main application - Contract AI Service
-
-Architecture:
-- REST API Routes: FastAPI routers for HTTP endpoints
-- SSE (Server-Sent Events): Real-time streaming for AI Contract and AI Legal
-
-Modules:
-- ai_contract: Contract generation and composition (SSE streaming)
-- ai_legal: Legal Q&A and search (SSE streaming)
-- conversation: Conversation history management
+FastAPI Server - Canifa Product Recommendation API
 """
 
 import os
 import logging
 
-# Configure Logging
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.staticfiles import StaticFiles
+from fastapi.responses import RedirectResponse
+
+from config import (
+    PORT,
+    ENV_MODE,
+    LANGSMITH_API_KEY,
+    LANGSMITH_ENDPOINT,
+    LANGSMITH_PROJECT,
+    LANGSMITH_TRACING,
+)
+from api.recommend_text import router as recommend_text_router
+
+# Logging
 logging.basicConfig(
     level=logging.INFO,
     format="%(asctime)s [%(levelname)s] %(name)s: %(message)s",
@@ -22,35 +28,20 @@ logging.basicConfig(
 )
 logger = logging.getLogger(__name__)
 
-from config import LANGSMITH_API_KEY, LANGSMITH_ENDPOINT, LANGSMITH_PROJECT, LANGSMITH_TRACING
-
-# Ensure LangSmith Env Vars are set for the process
+# LangSmith environment
 os.environ["LANGSMITH_TRACING"] = LANGSMITH_TRACING
 os.environ["LANGSMITH_ENDPOINT"] = LANGSMITH_ENDPOINT
 os.environ["LANGSMITH_API_KEY"] = LANGSMITH_API_KEY
 os.environ["LANGSMITH_PROJECT"] = LANGSMITH_PROJECT
 
-
-import uvicorn
-from fastapi import FastAPI
-from fastapi.middleware.cors import CORSMiddleware
-from fastapi.staticfiles import StaticFiles  # Import StaticFiles
-
-# from api.recommend_image import router as recommend_image_router
-from api.recommend_text import router as recommend_text_router
-
-from common.middleware import ClerkAuthMiddleware
-from config import PORT
-
+# FastAPI app
 app = FastAPI(
-    title="Contract AI Service",
-    description="API for Contract AI Service",
+    title="Canifa Product Recommendation API",
+    description="API for product recommendation using hybrid search",
     version="1.0.0",
 )
 
-print("✅ Clerk Authentication middleware DISABLED (for testing)")
-
-# Add CORS middleware
+# CORS
 app.add_middleware(
     CORSMiddleware,
     allow_origins=["*"],
@@ -59,51 +50,42 @@ app.add_middleware(
     allow_headers=["*"],
 )
 
-# app.include_router(recommend_image_router)
+# Routes
 app.include_router(recommend_text_router)
 
-try:
-    static_dir = os.path.join(os.path.dirname(__file__), "static")
-    if not os.path.exists(static_dir):
+# Static files
+static_dir = os.path.join(os.path.dirname(__file__), "static")
+if not os.path.exists(static_dir):
     os.makedirs(static_dir)
-    app.mount("/static", StaticFiles(directory=static_dir, html=True), name="static")
-    print(f"✅ Static files mounted at /static (Dir: {static_dir})")
-except Exception as e:
-    print(f"⚠️ Failed to mount static files: {e}")
+app.mount("/static", StaticFiles(directory=static_dir, html=True), name="static")
 
-from fastapi.responses import RedirectResponse
 
 @app.get("/")
 async def root():
     return RedirectResponse(url="/static/index.html")
 
+
 if __name__ == "__main__":
-    print("=" * 60)
-    print("🚀 Contract AI Service Starting...")
-    print("=" * 60)
-    print(f"📡 REST API: http://localhost:{PORT}")
-    print(f"📡 SSE Streaming: http://localhost:{PORT}/api/ai-contract/chat")
-    print(f"📡 Test Chatbot: http://localhost:{PORT}/static/index.html")
-    print(f"📚 API Docs: http://localhost:{PORT}/docs")
-    print("=" * 60)
-
-    # ENABLE_RELOAD = os.getenv("ENABLE_RELOAD", "false").lower() in ("true", "1", "yes")
-    ENABLE_RELOAD = True
-    print("⚠️  Hot reload: FORCED ON (Dev Mode)")
-    
-    reload_dirs = [
-        "ai_contract",
-        "conversation",
-        "common",
-        "api", # Watch api folder
-        "agent" # Watch agent folder
-    ]
+    import uvicorn
+    
+    is_dev = ENV_MODE != "prod"
     
+    logger.info(f"Starting server on port {PORT} (ENV_MODE={ENV_MODE})")
+    
+    if is_dev:
+        uvicorn.run(
+            "server:app",
+            host="0.0.0.0",
+            port=PORT,
+            reload=True,
+            reload_dirs=["."],
+            log_level="info",
+        )
+    else:
         uvicorn.run(
             "server:app",
             host="0.0.0.0",
             port=PORT,
-        reload=ENABLE_RELOAD,
-        reload_dirs=reload_dirs,
+            workers=os.cpu_count() * 2 + 1,
             log_level="info",
         )