Phase 7: Hoàn thiện Modular RAG Backend với FastAPI và Đa LLM Provider

.env.example

@@ -0,0 +1,49 @@
+# Microsoft Entra ID (Azure AD) Config
+TENANT_ID=your_tenant_id_here
+CLIENT_ID=your_client_id_here
+CLIENT_SECRET=your_client_secret_here
+
+# SharePoint Config
+SHAREPOINT_SITE_ID=your_sharepoint_site_id_here
+SHAREPOINT_DRIVE_ID="b!..."
+
+# ========================================
+# VLM (Vision-Language Model) CONFIGURATION
+# ========================================
+# The URL to your local/LAN VLM server (e.g. llama.cpp)
+VLM_ENDPOINT="http://10.202.50.3:8080/v1/chat/completions"
+# Set to low value (0.1) for exact extraction, higher (0.7) for more creativity
+VLM_TEMPERATURE="0.1"
+# Max tokens to generate per page
+VLM_MAX_TOKENS="2000"
+# Connection timeout in seconds
+VLM_TIMEOUT="120.0"
+
+# OpenSearch Config
+OPENSEARCH_HOST=localhost
+OPENSEARCH_PORT=9200
+OPENSEARCH_USER=admin
+OPENSEARCH_PASS=admin
+
+# ========================================
+# CHAT LLM CONFIGURATION (Phase 7)
+# ========================================
+# LLM_PROVIDER can be: 'gemini', 'groq', or 'local'
+LLM_PROVIDER=gemini
+
+# 1. Gemini Config (Default)
+GEMINI_API_KEY=your_gemini_api_key_here
+
+# 2. Groq Config
+GROQ_API_KEY=your_groq_api_key_here
+GROQ_MODEL=llama3-70b-8192
+
+# 3. Local Llama.cpp Config
+LOCAL_LLM_ENDPOINT="http://10.202.50.3:8081/v1/chat/completions"
+
+# (Legacy OpenAI - Can be removed if not used)
+OPENAI_API_KEY=your_openai_api_key_here
+
+# App Settings
+LOG_LEVEL=INFO
+ENVIRONMENT=development