Xplainable MCP Server

""" Autotrain service MCP tools. Auto-generated and maintained by the xplainable-client sync workflow. """ import logging from typing import Optional, List, Dict, Any from ..mcp_instance import mcp import xplainable_client.client.py_models.autotrain logger = logging.getLogger(__name__) # Import shared utilities from ..server import get_client # Autotrain Tools # ============================================ @mcp.tool() def autotrain_generate_labels(summary: xplainable_client.client.py_models.autotrain.DatasetSummary, team_id: Optional[str] = None, textgen_config: Optional[xplainable_client.client.py_models.autotrain.TextGenConfig] = None): """ Generate label recommendations for training. Args: summary: Dataset summary from summarize_dataset team_id: Team ID (uses session team_id if not provided) textgen_config: Text generation configuration Returns: List of label recommendations Raises: XplainableAPIError: If label generation fails Category: analysis """ try: client = get_client() result = client.autotrain.generate_labels(summary, team_id, textgen_config) logger.info(f"Executed autotrain.generate_labels") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_generate_labels: {e}") raise @mcp.tool() def autotrain_start_autotrain(model_name: str, model_description: str, summary: xplainable_client.client.py_models.autotrain.DatasetSummary, team_id: Optional[str] = None, textgen_config: Optional[xplainable_client.client.py_models.autotrain.TextGenConfig] = None): """ Start the autotrain process. Args: model_name: Name for the model model_description: Description of the model summary: Dataset summary from summarize_dataset team_id: Team ID (uses session team_id if not provided) textgen_config: Text generation configuration Returns: Training job information Raises: XplainableAPIError: If autotrain fails to start Category: write """ try: client = get_client() result = client.autotrain.start_autotrain(model_name, model_description, summary, team_id, textgen_config) logger.info(f"Executed autotrain.start_autotrain") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_start_autotrain: {e}") raise @mcp.tool() def autotrain_summarize_dataset(file_path: str, team_id: Optional[str] = None, textgen_config: Optional[xplainable_client.client.py_models.autotrain.TextGenConfig] = None): """ Summarize a dataset by uploading a file. Args: file_path: Path to the dataset file team_id: Team ID (uses session team_id if not provided) textgen_config: Text generation configuration Returns: Dataset summary and metadata Raises: FileNotFoundError: If file doesn't exist XplainableAPIError: If summarization fails Category: analysis """ try: client = get_client() result = client.autotrain.summarize_dataset(file_path, team_id, textgen_config) logger.info(f"Executed autotrain.summarize_dataset") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_summarize_dataset: {e}") raise @mcp.tool() def autotrain_generate_feature_engineering(summary: xplainable_client.client.py_models.autotrain.DatasetSummary, team_id: Optional[str] = None, n: int = 5, textgen_config: Optional[xplainable_client.client.py_models.autotrain.TextGenConfig] = None): """ Generate feature engineering recommendations. Args: summary: Dataset summary from summarize_dataset team_id: Team ID (uses session team_id if not provided) n: Number of recommendations to generate textgen_config: Text generation configuration Returns: List of feature engineering recommendations Raises: XplainableAPIError: If generation fails Category: analysis """ try: client = get_client() result = client.autotrain.generate_feature_engineering(summary, team_id, n, textgen_config) logger.info(f"Executed autotrain.generate_feature_engineering") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_generate_feature_engineering: {e}") raise @mcp.tool() def autotrain_generate_goals(summary: xplainable_client.client.py_models.autotrain.DatasetSummary, team_id: Optional[str] = None, n: int = 5, textgen_config: Optional[xplainable_client.client.py_models.autotrain.TextGenConfig] = None): """ Generate training goals based on dataset summary. Args: summary: Dataset summary from summarize_dataset team_id: Team ID (uses session team_id if not provided) n: Number of goals to generate textgen_config: Text generation configuration Returns: List of training goals Raises: XplainableAPIError: If goal generation fails Category: analysis """ try: client = get_client() result = client.autotrain.generate_goals(summary, team_id, n, textgen_config) logger.info(f"Executed autotrain.generate_goals") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_generate_goals: {e}") raise @mcp.tool() def autotrain_check_training_status(training_id: str, team_id: Optional[str] = None): """ Check the status of a training job. Args: training_id: Training job ID from start_autotrain team_id: Team ID (uses session team_id if not provided) Returns: Training status and progress information Raises: XplainableAPIError: If status check fails Category: read """ try: client = get_client() result = client.autotrain.check_training_status(training_id, team_id) logger.info(f"Executed autotrain.check_training_status") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_check_training_status: {e}") raise @mcp.tool() def autotrain_generate_insights(goal: Dict[str, Any], summary: xplainable_client.client.py_models.autotrain.DatasetSummary, team_id: Optional[str] = None, textgen_config: Optional[xplainable_client.client.py_models.autotrain.TextGenConfig] = None): """ Generate insights about the dataset. Args: goal: Analysis goal summary: Dataset summary team_id: Team ID (uses session team_id if not provided) textgen_config: Text generation configuration Returns: Generated insights and analysis Raises: XplainableAPIError: If insight generation fails Category: analysis """ try: client = get_client() result = client.autotrain.generate_insights(goal, summary, team_id, textgen_config) logger.info(f"Executed autotrain.generate_insights") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_generate_insights: {e}") raise @mcp.tool() def autotrain_visualize_data(summary: xplainable_client.client.py_models.autotrain.DatasetSummary, goal: Dict[str, Any], team_id: Optional[str] = None, library: str = 'plotly', textgen_config: Optional[xplainable_client.client.py_models.autotrain.TextGenConfig] = None): """ Generate data visualizations. Args: summary: Dataset summary goal: Visualization goal team_id: Team ID (uses session team_id if not provided) library: Visualization library (plotly, matplotlib, seaborn) textgen_config: Text generation configuration Returns: Visualization code and metadata Raises: XplainableAPIError: If visualization generation fails Category: analysis """ try: client = get_client() result = client.autotrain.visualize_data(summary, goal, team_id, library, textgen_config) logger.info(f"Executed autotrain.visualize_data") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_visualize_data: {e}") raise @mcp.tool() def autotrain_train_manual(label: str, model_name: str, model_description: str, preprocessor_id: str, version_id: str, team_id: Optional[str] = None, drop_columns: Optional[List[str]] = None): """ Train a model manually with specific parameters. Args: label: Target label column model_name: Name for the model model_description: Description of the model preprocessor_id: Preprocessor ID version_id: Preprocessor version ID team_id: Team ID (uses session team_id if not provided) drop_columns: Columns to drop Returns: Training job information Raises: XplainableAPIError: If training fails to start Category: write """ try: client = get_client() result = client.autotrain.train_manual(label, model_name, model_description, preprocessor_id, version_id, team_id, drop_columns) logger.info(f"Executed autotrain.train_manual") # Handle different return types if hasattr(result, 'model_dump'): return result.model_dump() elif isinstance(result, list) and result and hasattr(result[0], 'model_dump'): return [item.model_dump() for item in result] else: return result except Exception as e: logger.error(f"Error in autotrain_train_manual: {e}") raise