Agent MCP

# Agent-MCP/mcp_template/mcp_server_src/core/config.py import logging import os import sys from pathlib import Path from typing import List, Optional # Version information VERSION = "2.0" GITHUB_REPO = "rinadelph/Agent-MCP" AUTHOR = "Luis Alejandro Rincon" GITHUB_URL = "https://github.com/rinadelph" # --- TUI Colors (ANSI Escape Codes) --- class TUIColors: HEADER = "\033[95m" # Light Magenta OKBLUE = "\033[94m" # Light Blue OKCYAN = "\033[96m" # Light Cyan OKGREEN = "\033[92m" # Light Green WARNING = "\033[93m" # Yellow FAIL = "\033[91m" # Red ENDC = "\033[0m" # Reset to default BOLD = "\033[1m" UNDERLINE = "\033[4m" DIM = "\033[2m" # Specific log level colors DEBUG = OKCYAN INFO = OKGREEN WARNING = WARNING ERROR = FAIL CRITICAL = BOLD + FAIL class ColorfulFormatter(logging.Formatter): """Custom formatter to add colors to log messages for console output.""" LOG_LEVEL_COLORS = { logging.DEBUG: TUIColors.DEBUG, logging.INFO: TUIColors.INFO, logging.WARNING: TUIColors.WARNING, logging.ERROR: TUIColors.ERROR, logging.CRITICAL: TUIColors.CRITICAL, } def format(self, record): color = self.LOG_LEVEL_COLORS.get(record.levelno, TUIColors.ENDC) record.levelname = ( f"{color}{record.levelname:<8}{TUIColors.ENDC}" # Pad levelname ) record.name = f"{TUIColors.OKBLUE}{record.name}{TUIColors.ENDC}" return super().format(record) # --- General Configuration --- DB_FILE_NAME: str = "mcp_state.db" # From main.py:39 # --- Logging Configuration --- LOG_FILE_NAME: str = "mcp_server.log" # Based on main.py:46 LOG_LEVEL: int = logging.INFO # From main.py:43 LOG_FORMAT_FILE: str = "%(asctime)s - %(name)s - %(levelname)s - %(message)s" LOG_FORMAT_CONSOLE: str = ( f"%(asctime)s - %(name)s - %(levelname)s - {TUIColors.DIM}%(message)s{TUIColors.ENDC}" # Dim message text ) CONSOLE_LOGGING_ENABLED = ( os.environ.get("MCP_DEBUG", "false").lower() == "true" ) # Enable console logging in debug mode def setup_logging(): """Configures global logging for the application.""" root_logger = logging.getLogger() # Get the root logger root_logger.setLevel(LOG_LEVEL) # Set level on the root logger # Clear any existing handlers on the root logger to avoid duplication for handler in root_logger.handlers[:]: root_logger.removeHandler(handler) # 1. File Handler (only in debug mode) debug_mode = os.environ.get("MCP_DEBUG", "false").lower() == "true" if debug_mode: file_formatter = logging.Formatter(LOG_FORMAT_FILE) file_handler = logging.FileHandler( LOG_FILE_NAME, mode="a", encoding="utf-8" ) # Append mode file_handler.setFormatter(file_formatter) root_logger.addHandler(file_handler) # 2. Console Handler (with colors, conditional) if CONSOLE_LOGGING_ENABLED: console_formatter = ColorfulFormatter( LOG_FORMAT_CONSOLE, datefmt="%H:%M:%S" ) # Simpler datefmt for console console_handler = logging.StreamHandler(sys.stdout) console_handler.setFormatter(console_formatter) # Filter out less important messages for console if desired # console_handler.setLevel(logging.INFO) # Example: only INFO and above for console root_logger.addHandler(console_handler) # Suppress overly verbose logs from specific libraries for both file and console logging.getLogger("watchfiles").setLevel(logging.WARNING) # Uvicorn access logs are handled by Uvicorn's config (access_log=False in cli.py) # but we can also try to manage its error logger if needed. logging.getLogger("uvicorn.error").setLevel(logging.WARNING) logging.getLogger("uvicorn").setLevel(logging.WARNING) # General uvicorn logger logging.getLogger("mcp.server.lowlevel.server").propagate = ( False # Prevent duplication if it logs directly ) def enable_console_logging(): """Enable console logging dynamically (used when debug mode is enabled).""" global CONSOLE_LOGGING_ENABLED CONSOLE_LOGGING_ENABLED = True # Re-setup logging to add file handler when debug mode is enabled setup_logging() root_logger = logging.getLogger() # Check if console handler already exists has_console_handler = any( isinstance(handler, logging.StreamHandler) and handler.stream == sys.stdout for handler in root_logger.handlers ) if not has_console_handler: console_formatter = ColorfulFormatter(LOG_FORMAT_CONSOLE, datefmt="%H:%M:%S") console_handler = logging.StreamHandler(sys.stdout) console_handler.setFormatter(console_formatter) # Set logging level to DEBUG for more verbose output console_handler.setLevel(logging.DEBUG) root_logger.addHandler(console_handler) # Also set root logger to DEBUG level root_logger.setLevel(logging.DEBUG) # Initialize logging when this module is imported setup_logging() logger = logging.getLogger("mcp_server") # Main application logger # --- Agent Appearance --- AGENT_COLORS: List[str] = ( [ # From main.py:160-164 (Note: original had 160-165, but list ends on 164) "#FF5733", "#33FF57", "#3357FF", "#FF33A1", "#A133FF", "#33FFA1", "#FFBD33", "#33FFBD", "#BD33FF", "#FF3333", "#33FF33", "#3333FF", "#FF8C00", "#00CED1", "#9400D3", "#FF1493", "#7FFF00", "#1E90FF", ] ) # --- OpenAI Model Configuration --- # Advanced mode flag - set by CLI ADVANCED_EMBEDDINGS: bool = False # Default to simple mode # Auto-indexing control - set by CLI DISABLE_AUTO_INDEXING: bool = False # Default to automatic indexing # Original/Simple mode configuration (default) - restored to original values SIMPLE_EMBEDDING_MODEL: str = ( "text-embedding-3-large" # Original embedding model (unchanged) ) SIMPLE_EMBEDDING_DIMENSION: int = 1536 # Increased from 1024 for better performance # Advanced mode configuration - new enhanced mode ADVANCED_EMBEDDING_MODEL: str = "text-embedding-3-large" # From main.py:178 ADVANCED_EMBEDDING_DIMENSION: int = ( 3072 # Full dimension for text-embedding-3-large for better code understanding ) # Dynamic configuration based on mode EMBEDDING_MODEL: str = ( ADVANCED_EMBEDDING_MODEL if ADVANCED_EMBEDDINGS else SIMPLE_EMBEDDING_MODEL ) EMBEDDING_DIMENSION: int = ( ADVANCED_EMBEDDING_DIMENSION if ADVANCED_EMBEDDINGS else SIMPLE_EMBEDDING_DIMENSION ) CHAT_MODEL: str = ( "gpt-4.1-2025-04-14" # From main.py:179 (Ensure this matches your desired model) ) TASK_ANALYSIS_MODEL: str = ( "gpt-4.1-2025-04-14" # Same model for consistent task placement analysis ) MAX_EMBEDDING_BATCH_SIZE: int = 100 # From main.py:181 MAX_CONTEXT_TOKENS: int = 1000000 # GPT-4.1 has 1M token context window TASK_ANALYSIS_MAX_TOKENS: int = ( 1000000 # Same 1M token context window for task analysis ) # --- Project Directory Helpers --- # These rely on an environment variable "MCP_PROJECT_DIR" being set, # typically by the CLI entry point (previously in main.py:1953, will be in cli.py). def get_project_dir() -> Path: """Gets the resolved absolute path to the project directory.""" project_dir_str = os.environ.get("MCP_PROJECT_DIR") if not project_dir_str: # This case should ideally be handled at startup by the CLI, # ensuring MCP_PROJECT_DIR is always set. logger.error("CRITICAL: MCP_PROJECT_DIR environment variable is not set.") # Fallback to current directory, but this is likely not intended for normal operation. return Path(".").resolve() return Path(project_dir_str).resolve() def get_agent_dir() -> Path: """Gets the path to the .agent directory within the project directory.""" return get_project_dir() / ".agent" def get_db_path() -> Path: """Gets the full path to the SQLite database file.""" return get_agent_dir() / DB_FILE_NAME # --- Environment Variable Check (Optional but good practice) --- OPENAI_API_KEY_ENV: Optional[str] = os.environ.get("OPENAI_API_KEY") # From main.py:174 # Debug print statement removed for clean console output if not OPENAI_API_KEY_ENV: logger.error( "CRITICAL: OPENAI_API_KEY not found in environment variables. Please set it in your .env file or environment." ) # Depending on strictness, you might want to raise an exception or sys.exit(1) here # For now, just logging, as the openai_service.py will handle the client init failure. # --- Task Placement Configuration (System 8) --- ENABLE_TASK_PLACEMENT_RAG: bool = ( os.getenv("ENABLE_TASK_PLACEMENT_RAG", "true").lower() == "true" ) TASK_DUPLICATION_THRESHOLD: float = float( os.getenv("TASK_DUPLICATION_THRESHOLD", "0.8") ) ALLOW_RAG_OVERRIDE: bool = os.getenv("ALLOW_RAG_OVERRIDE", "true").lower() == "true" TASK_PLACEMENT_RAG_TIMEOUT: int = int( os.getenv("TASK_PLACEMENT_RAG_TIMEOUT", "5") ) # seconds # Log that configuration is loaded (optional) logger.info("Core configuration loaded (with colorful logging setup).") # Example of how other modules will use this logger: # from mcp_server_src.core.config import logger # logger.info("This is a log message from another module.")