dslighting 1.3.9__py3-none-any.whl

dslighting/__init__.py

@@ -0,0 +1,170 @@
+"""
+DSLighting: Simplified API for Data Science Agent Automation
+
+A progressive API that provides sensible defaults with full control when needed.
+
+Quick Start:
+    >>> import dslighting
+    >>>
+    >>> # Simple usage
+    >>> data = dslighting.load_data("path/to/data")
+    >>> agent = dslighting.Agent()
+    >>> result = agent.run(data)
+    >>>
+    >>> # One-liner
+    >>> result = dslighting.run_agent("path/to/data")
+
+Advanced Usage:
+    >>> agent = dslighting.Agent(
+    ...     workflow="autokaggle",
+    ...     model="gpt-4o",
+    ...     temperature=0.5,
+    ...     max_iterations=10
+    ... )
+    >>> result = agent.run(data)
+
+For more information, see: https://github.com/usail-hkust/dslighting
+"""
+
+__version__ = "1.3.9"
+__author__ = "DSLighting Team"
+
+# Core API classes
+from dslighting.core.agent import Agent, AgentResult
+from dslighting.core.data_loader import DataLoader, LoadedData
+
+# Convenience functions
+def load_data(source, **kwargs):
+    """
+    Load and auto-detect data type.
+
+    This is a convenience function that creates a DataLoader and loads data.
+    For Kaggle/MLE-Bench competitions, it automatically extracts task_id from the path.
+
+    Args:
+        source: Data source (path, DataFrame, dict, etc.)
+        **kwargs: Additional parameters passed to DataLoader
+
+    Returns:
+        LoadedData with auto-detected task information and task_id
+
+    Examples:
+        >>> # Load from competition path (task_id auto-detected)
+        >>> data = dslighting.load_data("data/competitions/bike-sharing-demand")
+        >>> print(data.task_id)  # "bike-sharing-demand"
+        >>>
+        >>> agent = dslighting.Agent()
+        >>> result = agent.run(data)  # Auto-graded using task_id
+
+        >>> # Load DataFrame (no task_id)
+        >>> df = pd.read_csv("data.csv")
+        >>> data = dslighting.load_data(df)
+
+        >>> data = dslighting.load_data("data/house-prices")
+        >>> print(data.task_detection.task_type)
+        'kaggle'
+    """
+    loader = DataLoader()
+    return loader.load(source, **kwargs)
+
+
+def run_agent(data=None, task_id=None, data_dir=None, keep_workspace=False, keep_workspace_on_failure=True, **kwargs):
+    """
+    Quick one-liner: load data and run with defaults.
+
+    This function creates an Agent with the specified parameters and runs it on the data.
+
+    Args:
+        data: Optional data source (path, DataFrame, dict, etc.)
+        task_id: Task/Competition identifier (e.g., "bike-sharing-demand")
+        data_dir: Base data directory (default: "data/competitions")
+        keep_workspace: Keep workspace after completion (default: False)
+        keep_workspace_on_failure: Keep workspace on failure (default: True)
+        **kwargs: Parameters passed to Agent.__init__ and Agent.run
+
+    Returns:
+        AgentResult with output, metrics, and metadata
+
+    Examples:
+        >>> # Recommended: using task_id
+        >>> result = dslighting.run_agent(
+        ...     task_id="bike-sharing-demand",
+        ...     data_dir="data/competitions"
+        ... )
+        >>> print(f"Score: {result.score}, Cost: ${result.cost}")
+
+        >>> # Legacy: using data path
+        >>> result = dslighting.run_agent("data/titanic")
+
+        >>> # With custom parameters
+        >>> result = dslighting.run_agent(
+        ...     task_id="bike-sharing-demand",
+        ...     workflow="autokaggle",
+        ...     model="gpt-4o",
+        ...     keep_workspace=True  # Keep workspace for debugging
+        ... )
+    """
+    # Extract run-specific parameters if present
+    run_kwargs = {}
+    agent_params = {}
+
+    # Parameters that should go to run(), not __init__
+    run_only_params = {'task_id', 'data_dir', 'output_path', 'description'}
+
+    for key, value in kwargs.items():
+        if key in run_only_params:
+            run_kwargs[key] = value
+        else:
+            agent_params[key] = value
+
+    # Add explicit parameters to run_kwargs
+    if task_id is not None:
+        run_kwargs['task_id'] = task_id
+    if data_dir is not None:
+        run_kwargs['data_dir'] = data_dir
+
+    # Add workspace preservation parameters to agent
+    agent_params['keep_workspace'] = keep_workspace
+    agent_params['keep_workspace_on_failure'] = keep_workspace_on_failure
+
+    # Create agent and run
+    agent = Agent(**agent_params)
+    return agent.run(data, **run_kwargs)
+
+
+# Public API
+__all__ = [
+    # Version info
+    "__version__",
+    "__author__",
+
+    # Core classes
+    "Agent",
+    "AgentResult",
+    "DataLoader",
+    "LoadedData",
+
+    # Convenience functions
+    "load_data",
+    "run_agent",
+]
+
+
+# Import logging configuration
+try:
+    import logging
+    from rich.logging import RichHandler
+
+    # Set up rich logging if available
+    logging.basicConfig(
+        level="INFO",
+        format="%(asctime)s [%(levelname)s] %(message)s",
+        datefmt="%H:%M:%S",
+        handlers=[RichHandler(show_path=False)],
+    )
+except ImportError:
+    # Fallback to basic logging
+    logging.basicConfig(
+        level="INFO",
+        format="%(asctime)s [%(levelname)s] %(message)s",
+    )

dslighting/core/__init__.py

@@ -0,0 +1,13 @@
+"""
+Core modules for DSLighting simplified API.
+"""
+
+from dslighting.core.agent import Agent, AgentResult
+from dslighting.core.data_loader import DataLoader, LoadedData
+
+__all__ = [
+    "Agent",
+    "AgentResult",
+    "DataLoader",
+    "LoadedData",
+]