Release v0.1.0a11: API hardening, correctness fixes, beta prep

pyproject.toml

@@ -295,13 +295,6 @@ source = ["src/ml4t/engineer"]
 omit = [
     "*/tests/*",
     "*/__init__.py",
-    "*/bars/base.py",
-    "*/bars/imbalance.py",
-    "*/bars/run.py",
-    "*/bars/vectorized.py",
-    "*/labeling/barriers.py",
-    "*/labeling/core.py",
-    "*/features/microstructure/*",
 ]
 
 [tool.coverage.report]

src/ml4t/engineer/AGENT.md

@@ -45,11 +45,9 @@ Package-level navigation for feature engineering library.
 | `preprocessing_config.py` | Scaler config | `PreprocessingConfig`, `ScalerConfig` |
 | `base.py` | Base config classes | `BaseConfig` |
 
-## selection/ - Feature Selection
+## selection/ - Deprecated (Moved to ml4t-diagnostic)
 
-| File | Purpose |
-|------|---------|
-| `systematic.py` | Correlation-based, variance threshold, mutual info |
+FeatureSelector has moved to `ml4t.diagnostic.selection`.
 
 ## store/ - Storage
 

src/ml4t/engineer/api.py

@@ -78,20 +78,6 @@
     }
 )
 
-# Common default values for required parameters missing from metadata
-# Used as fallback when feature registration is incomplete
-COMMON_PARAM_DEFAULTS: dict[str, Any] = {
-    "period": 14,
-    "window": 20,
-    "lookback": 20,
-    "lag": 1,
-    "lags": [1],
-    "n": 5,
-    "bins": 10,
-    "features": ["close"],
-    "windows": [5, 10, 20],
-}
-
 
 def compute_features(
     data: pl.DataFrame | pl.LazyFrame,
@@ -175,6 +161,11 @@ def compute_features(
     - Circular dependencies are detected and raise ValueError
     - Parameters in config override default parameters from registry
     """
+    from ml4t.engineer.core.schemas import validate_ohlcv_schema
+
+    # Validate input schema (flexible: no asset_id required, flexible time column)
+    validate_ohlcv_schema(data, require_asset_id=False, allow_flexible_time=True)
+
     # Parse input to standardized format
     feature_specs = _parse_feature_input(features)
 
@@ -403,16 +394,12 @@ def _execute_feature(
             pass
         else:
             # Required parameter with no default and not in COLUMN_ARG_MAP
-            # This indicates incomplete metadata - try common defaults
-            if param_name in COMMON_PARAM_DEFAULTS:
-                keyword_params[param_name] = COMMON_PARAM_DEFAULTS[param_name]
-            else:
-                # Cannot proceed - need user to provide this parameter explicitly
-                raise ValueError(
-                    f"Feature '{feature_name}' requires parameter '{param_name}' but it's not "
-                    f"provided. Call with explicit parameters: "
-                    f'compute_features(df, [{{"name": "{feature_name}", "{param_name}": value}}])'
-                )
+            raise ValueError(
+                f"Feature '{feature_name}' requires parameter '{param_name}' but it's not "
+                f"provided in metadata.parameters or call params. "
+                f'Use: compute_features(df, [{{"name": "{feature_name}", '
+                f'"params": {{"{param_name}": value}}}}])'
+            )
 
     # Call the feature function
     try:

src/ml4t/engineer/bars/run.py

@@ -1,4 +1,3 @@
-# mypy: disable-error-code="misc,operator,assignment,arg-type"
 """Run bar sampler implementations.
 
 AFML-compliant run bars per López de Prado Chapter 2.3.

src/ml4t/engineer/bars/vectorized.py

@@ -1,4 +1,3 @@
-# mypy: disable-error-code="misc,operator,assignment,arg-type"
 """
 Vectorized bar samplers using Polars for high performance.
 

src/ml4t/engineer/config/base.py

@@ -1,4 +1,3 @@
-# mypy: disable-error-code="misc,no-any-return,type-arg"
 # ruff: noqa: E721
 """Base configuration classes and shared utilities.
 

src/ml4t/engineer/config/experiment.py

@@ -1,4 +1,3 @@
-# mypy: disable-error-code="misc,no-any-return"
 """Experiment configuration loading utilities.
 
 This module provides helpers for loading complete experiment configurations
@@ -209,10 +208,22 @@ def save_experiment_config(
             exclude_none=True,
         )
 
+    # Convert tuples to lists for YAML safe_load compatibility
+    output = _tuples_to_lists(output)
+
     with open(path, "w") as f:
         yaml.dump(output, f, default_flow_style=False, sort_keys=False)
 
 
+def _tuples_to_lists(obj: Any) -> Any:
+    """Recursively convert tuples to lists for YAML-safe serialization."""
+    if isinstance(obj, dict):
+        return {k: _tuples_to_lists(v) for k, v in obj.items()}
+    elif isinstance(obj, tuple | list):
+        return [_tuples_to_lists(item) for item in obj]
+    return obj
+
+
 __all__ = [
     "ExperimentConfig",
     "load_experiment_config",

src/ml4t/engineer/config/feature_config.py

@@ -1,4 +1,3 @@
-# mypy: disable-error-code="misc,call-arg,arg-type"
 # ruff: noqa: UP006, UP045
 """Feature evaluation configuration (Modules A, B, C).
 

src/ml4t/engineer/config/labeling.py

@@ -1,4 +1,3 @@
-# mypy: disable-error-code="misc,no-any-return,arg-type"
 """Labeling configuration with Pydantic v2 serialization.
 
 This module provides config classes for ML labeling methods that extend BaseConfig,

src/ml4t/engineer/config/preprocessing_config.py

@@ -1,4 +1,3 @@
-# mypy: disable-error-code="misc,no-any-return,call-arg"
 """Preprocessing configuration with Pydantic v2 serialization.
 
 This module provides config classes for preprocessing (scalers) that extend BaseConfig,
@@ -289,15 +288,18 @@ def create_scaler(self) -> BaseScaler | None:
 
         if self.scaler == "standard":
             return StandardScaler(
+                columns=self.columns,
                 with_mean=self.with_mean,
                 with_std=self.with_std,
             )
         elif self.scaler == "minmax":
             return MinMaxScaler(
+                columns=self.columns,
                 feature_range=self.feature_range,
             )
         elif self.scaler == "robust":
             return RobustScaler(
+                columns=self.columns,
                 with_centering=self.with_centering,
                 with_scaling=self.with_scaling,
                 quantile_range=self.quantile_range,

src/ml4t/engineer/core/__init__.py

@@ -16,16 +16,11 @@
     DataError,
     DataSchemaError,
     DataValidationError,
-    ImplementationNotAvailableError,
-    IndicatorError,  # Deprecated
     InsufficientDataError,
     IntegrationError,
-    InvalidArgumentError,
     InvalidParameterError,
     # Base exception
-    QuantLabTAError,
-    # Backward compatibility aliases
-    TechnicalAnalysisError,
+    ML4TEngineerError,
     ValidationError,
 )
 from ml4t.engineer.core.registry import (
@@ -96,7 +91,7 @@
     "FeatureRegistry",
     "get_registry",
     # Exceptions - Base
-    "QuantLabTAError",
+    "ML4TEngineerError",
     # Exceptions - First-level (flat hierarchy)
     "ConfigurationError",
     "ValidationError",
@@ -107,11 +102,6 @@
     "ComputationError",
     "DataError",
     "IntegrationError",
-    # Exceptions - Backward compatibility aliases
-    "TechnicalAnalysisError",
-    "IndicatorError",  # Deprecated
-    "InvalidArgumentError",
-    "ImplementationNotAvailableError",
     # Validation
     "validate_lag",
     "validate_list_length",

src/ml4t/engineer/core/decorators.py

@@ -1,4 +1,3 @@
-# mypy: disable-error-code="misc,arg-type,assignment"
 """Feature registration decorators.
 
 Simple decorator-based registration for features with zero overhead.

src/ml4t/engineer/core/exceptions.py

@@ -5,7 +5,7 @@
 provide actionable error messages.
 
 Exception Hierarchy (Flat - D06 Pattern):
-    QuantLabTAError (base)
+    ML4TEngineerError (base)
     ├── ConfigurationError      # Configuration and setup errors
     ├── ValidationError         # Input validation failures (also inherits ValueError)
     ├── InvalidParameterError   # Invalid parameters to indicators
@@ -27,11 +27,10 @@
 
 from __future__ import annotations
 
-import warnings
 from typing import Any
 
 
-class QuantLabTAError(Exception):
+class ML4TEngineerError(Exception):
     """
     Base exception class for all ML4T Engineer errors.
 
@@ -44,7 +43,7 @@ class QuantLabTAError(Exception):
         cause: Original exception if error was wrapped
 
     Example:
-        >>> raise QuantLabTAError(
+        >>> raise ML4TEngineerError(
         ...     "Operation failed",
         ...     context={"operation": "compute_rsi", "reason": "insufficient_data"}
         ... )
@@ -98,7 +97,7 @@ def __repr__(self) -> str:
 # =============================================================================
 
 
-class ConfigurationError(QuantLabTAError):
+class ConfigurationError(ML4TEngineerError):
     """
     Configuration and setup errors.
 
@@ -119,11 +118,11 @@ class ConfigurationError(QuantLabTAError):
     pass
 
 
-class ValidationError(QuantLabTAError, ValueError):
+class ValidationError(ML4TEngineerError, ValueError):
     """
     Input validation failures.
 
-    Inherits from both QuantLabTAError (for library-specific catching)
+    Inherits from both ML4TEngineerError (for library-specific catching)
     and ValueError (for standard Python parameter error handling).
 
     Raised when:
@@ -143,7 +142,7 @@ class ValidationError(QuantLabTAError, ValueError):
     pass
 
 
-class InvalidParameterError(QuantLabTAError, ValueError):
+class InvalidParameterError(ML4TEngineerError, ValueError):
     """
     Invalid parameters provided to indicators.
 
@@ -167,7 +166,7 @@ class InvalidParameterError(QuantLabTAError, ValueError):
     pass
 
 
-class DataValidationError(QuantLabTAError):
+class DataValidationError(ML4TEngineerError):
     """
     Data validation failures.
 
@@ -188,7 +187,7 @@ class DataValidationError(QuantLabTAError):
     pass
 
 
-class DataSchemaError(QuantLabTAError):
+class DataSchemaError(ML4TEngineerError):
     """
     Schema validation failures.
 
@@ -209,7 +208,7 @@ class DataSchemaError(QuantLabTAError):
     pass
 
 
-class InsufficientDataError(QuantLabTAError):
+class InsufficientDataError(ML4TEngineerError):
     """
     Insufficient data for calculation.
 
@@ -229,7 +228,7 @@ class InsufficientDataError(QuantLabTAError):
     pass
 
 
-class ComputationError(QuantLabTAError):
+class ComputationError(ML4TEngineerError):
     """
     Calculation and numerical errors.
 
@@ -250,7 +249,7 @@ class ComputationError(QuantLabTAError):
     pass
 
 
-class DataError(QuantLabTAError):
+class DataError(ML4TEngineerError):
     """
     Data access and format errors.
 
@@ -271,7 +270,7 @@ class DataError(QuantLabTAError):
     pass
 
 
-class IntegrationError(QuantLabTAError):
+class IntegrationError(ML4TEngineerError):
     """
     External library integration errors.
 
@@ -292,50 +291,13 @@ class IntegrationError(QuantLabTAError):
     pass
 
 
-# =============================================================================
-# Backward Compatibility Aliases
-# =============================================================================
-
-
-# Primary alias for historical naming
-TechnicalAnalysisError = QuantLabTAError
-
-
-class _DeprecatedIndicatorError(QuantLabTAError):
-    """Deprecated intermediate class - use QuantLabTAError or specific errors."""
-
-    def __init__(
-        self,
-        message: str,
-        context: dict[str, Any] | None = None,
-        cause: Exception | None = None,
-    ):
-        warnings.warn(
-            "IndicatorError is deprecated. Use QuantLabTAError or a specific "
-            "exception type (InsufficientDataError, ComputationError) instead.",
-            DeprecationWarning,
-            stacklevel=2,
-        )
-        super().__init__(message, context, cause)
-
-
-# IndicatorError was an intermediate class - deprecated but kept for backward compat
-IndicatorError = _DeprecatedIndicatorError
-
-# InvalidArgumentError is an alias for InvalidParameterError
-InvalidArgumentError = InvalidParameterError
-
-# ImplementationNotAvailableError maps to IntegrationError
-ImplementationNotAvailableError = IntegrationError
-
-
 # =============================================================================
 # Public API
 # =============================================================================
 
 __all__ = [
     # Base exception
-    "QuantLabTAError",
+    "ML4TEngineerError",
     # First-level exceptions (flat hierarchy)
     "ConfigurationError",
     "ValidationError",
@@ -346,9 +308,4 @@ def __init__(
     "ComputationError",
     "DataError",
     "IntegrationError",
-    # Backward compatibility aliases
-    "TechnicalAnalysisError",
-    "IndicatorError",  # Deprecated
-    "InvalidArgumentError",
-    "ImplementationNotAvailableError",
 ]

src/ml4t/engineer/core/registry.py

@@ -59,7 +59,7 @@ class FeatureMetadata:
     """
 
     name: str
-    func: Callable[..., pl.DataFrame | pl.LazyFrame]
+    func: Callable[..., pl.Expr | dict[str, pl.Expr] | pl.DataFrame | pl.LazyFrame]
     category: str
     description: str
     formula: str = ""