dragon-ml-toolbox 10.1.1__py3-none-any.whl → 14.2.0__py3-none-any.whl

ml_tools/ETL_engineering.py

@@ -1,8 +1,13 @@
 import polars as pl
 import re
+from pathlib import Path
 from typing import Literal, Union, Optional, Any, Callable, List, Dict, Tuple
+
+from .utilities import load_dataframe, save_dataframe_filename
+from .path_manager import make_fullpath
 from ._script_info import _script_info
 from ._logger import _LOGGER
+from .constants import CHEMICAL_ELEMENT_SYMBOLS
 
 
 __all__ = [
@@ -14,11 +19,15 @@ __all__ = [
     "KeywordDummifier",
     "NumberExtractor",
     "MultiNumberExtractor",
+    "TemperatureExtractor",
+    "MultiTemperatureExtractor",
     "RatioCalculator",
+    "TriRatioCalculator",
     "CategoryMapper",
     "RegexMapper",
     "ValueBinner",
-    "DateFeatureExtractor"
+    "DateFeatureExtractor",
+    "MolecularFormulaTransformer"
 ]
 
 ############ TRANSFORM MAIN ####################
@@ -42,17 +51,20 @@ class TransformationRecipe:
     def add(
         self,
         input_col_name: str,
-        output_col_names: Union[str, List[str]],
         transform: Union[str, Callable],
+        output_col_names: Optional[Union[str, List[str]]] = None
     ) -> "TransformationRecipe":
         """
         Adds a new transformation step to the recipe.
 
         Args:
-            input_col: The name of the column from the source DataFrame.
-            output_col: The desired name(s) for the output column(s).
-                        A string for a 1-to-1 mapping, or a list of strings
-                        for a 1-to-many mapping.
+            input_col_name: The name of the column from the source DataFrame.
+            output_col_names: The desired name(s) for the output column(s).
+                        - A string for a 1-to-1 mapping.
+                        - A list of strings for a 1-to-many mapping.
+                        - A string prefix for 1-to-many mapping.
+                        - If None, the input name is used for 1-to-1 transforms,
+                          or the transformer's default names are used for 1-to-many.
             transform: The transformation to apply: 
                 - Use "rename" for simple column renaming
                 - If callable, must accept a `pl.Series` as the only parameter and return either a `pl.Series` or `pl.DataFrame`.
@@ -72,10 +84,6 @@ class TransformationRecipe:
         elif not isinstance(transform, Callable):
             _LOGGER.error(f"'transform' must be a callable function or the string '{_RENAME}'.")
             raise TypeError()
-
-        if isinstance(output_col_names, list) and transform == _RENAME:
-            _LOGGER.error("A RENAME operation cannot have a list of output columns.")
-            raise ValueError()
         
         # --- Add Step ---
         step = {
@@ -99,7 +107,7 @@ class DataProcessor:
     """
     Transforms a Polars DataFrame based on a provided `TransformationRecipe` object.
     
-    Use the method `transform()`.
+    Use the methods `transform()` or `load_transform_save()`.
     """
     def __init__(self, recipe: TransformationRecipe):
         """
@@ -142,33 +150,53 @@ class DataProcessor:
                 result = transform_action(input_series)
 
                 if isinstance(result, pl.Series):
-                    if not isinstance(output_col_spec, str):
-                        _LOGGER.error(f"Function for '{input_col_name}' returned a Series but 'output_col' is not a string.")
+                    # Default to input name if spec is None
+                    output_name = output_col_spec if output_col_spec is not None else input_col_name
+                    
+                    if not isinstance(output_name, str):
+                        _LOGGER.error(f"Function for '{input_col_name}' returned a Series but 'output_col' must be a string or None.")
                         raise TypeError()
-                    processed_columns.append(result.alias(output_col_spec))
+                    processed_columns.append(result.alias(output_name))
                 
                 elif isinstance(result, pl.DataFrame):
-                    # 1. Handle list-based renaming
-                    if isinstance(output_col_spec, list):
+                    # 1. Handle None in output names
+                    if output_col_spec is None:
+                        # Use the column names generated by the transformer directly
+                        processed_columns.extend(result.get_columns())
+                    
+                    # 2. Handle list-based renaming
+                    elif isinstance(output_col_spec, list):
                         if len(result.columns) != len(output_col_spec):
                             _LOGGER.error(f"Mismatch in '{input_col_name}': function produced {len(result.columns)} columns, but recipe specifies {len(output_col_spec)} output names.")
                             raise ValueError()
                         
                         renamed_df = result.rename(dict(zip(result.columns, output_col_spec)))
                         processed_columns.extend(renamed_df.get_columns())
-
-                    # 2. Handle a string prefix for AutoDummifier
+                    
+                    # 3. Global logic for adding a single prefix to all columns.
                     elif isinstance(output_col_spec, str):
                         prefix = output_col_spec
-                        # Replace the original name part with the desired prefix.
-                        new_names = {
-                            col: f"{prefix}{col[len(input_col_name):]}" for col in result.columns
-                        }
+                        new_names = {}
+                        
+                        for col in result.columns:
+                            # Case 1: Transformer's output column name contains the input name.
+                            # Action: Replace the input name with the desired prefix.
+                            # Example: input='color', output='color_red', prefix='spec' -> 'spec_red'
+                            if input_col_name in col:
+                                new_names[col] = col.replace(input_col_name, prefix, 1)
+                            
+                            # Case 2: Transformer's output is an independent name.
+                            # Action: Prepend the prefix to the output name.
+                            # Example: input='ratio', output='A_B', prefix='spec' -> 'spec_A_B'
+                            else:
+                                new_names[col] = f"{prefix}_{col}"
+                                
                         renamed_df = result.rename(new_names)
-                        processed_columns.extend(renamed_df.get_columns())
+                        processed_columns.extend(renamed_df.get_columns())    
+                    
                     
                     else:
-                        _LOGGER.error(f"Function for '{input_col_name}' returned a DataFrame, so 'output_col' must be a list of names or a string prefix.")
+                        _LOGGER.error(f"Function for '{input_col_name}' returned a DataFrame, so 'output_col' must be a list of names, a string prefix, or None.")
                         raise TypeError()
                 
                 else:
@@ -182,9 +210,28 @@ class DataProcessor:
         if not processed_columns:
             _LOGGER.error("The transformation resulted in an empty DataFrame.")
             return pl.DataFrame()
-            
+        
+        _LOGGER.info(f"Processed dataframe with {len(processed_columns)} columns.")
+
         return pl.DataFrame(processed_columns)
     
+    def load_transform_save(self, input_path: Union[str,Path], output_path: Union[str,Path]):
+        """
+        Convenience wrapper for the transform method that includes automatic dataframe loading and saving.
+        """
+        # Validate paths
+        in_path = make_fullpath(input_path, enforce="file")
+        out_path = make_fullpath(output_path, make=True, enforce="file")
+        
+        # load df
+        df, _ = load_dataframe(df_path=in_path, kind="polars", all_strings=True)
+        
+        # Process
+        df_processed = self.transform(df)
+        
+        # save processed df
+        save_dataframe_filename(df=df_processed, save_dir=out_path.parent, filename=out_path.name)
+        
     def __str__(self) -> str:
         """
         Provides a detailed, human-readable string representation of the
@@ -253,7 +300,7 @@ class BinaryTransformer:
             _LOGGER.error("Provide either 'true_keywords' or 'false_keywords', but not both.")
             raise ValueError()
         if true_keywords is None and false_keywords is None:
-            _LOGGER.error("You must provide either 'true_keywords' or 'false_keywords'.")
+            _LOGGER.error("Provide either 'true_keywords' or 'false_keywords'.")
             raise ValueError()
 
         # --- Configuration ---
@@ -285,16 +332,17 @@ class BinaryTransformer:
         Returns:
             pl.Series: A new Series of type UInt8 containing 1s and 0s.
         """
+        column_base_name = column.name
         # Create a boolean Series: True if any keyword is found, else False
         contains_keyword = column.str.contains(self.pattern)
 
         # Apply logic and cast directly to integer type
         if self.mode == "true_mode":
             # True -> 1, False -> 0
-            return contains_keyword.cast(pl.UInt8)
+            return contains_keyword.cast(pl.UInt8).alias(column_base_name)
         else: # false_mode
             # We want the inverse: True -> 0, False -> 1
-            return (~contains_keyword).cast(pl.UInt8)
+            return (~contains_keyword).cast(pl.UInt8).alias(column_base_name)
 
 
 class AutoDummifier:
@@ -302,6 +350,15 @@ class AutoDummifier:
     A transformer that performs one-hot encoding on a categorical column,
     automatically detecting the unique categories from the data.
     """
+    def __init__(self, drop_first: bool = False):
+        """
+        Initializes the AutoDummifier.
+
+        Args:
+            drop_first (bool): If True, drops the first dummy column.
+        """
+        self.drop_first = drop_first
+
     def __call__(self, column: pl.Series) -> pl.DataFrame:
         """
         Executes the one-hot encoding logic.
@@ -314,8 +371,20 @@ class AutoDummifier:
                           Column names are auto-generated by Polars as
                           '{original_col_name}_{category_value}'.
         """
-        # Ensure the column is treated as a string before creating dummies
-        return column.cast(pl.Utf8).to_dummies()
+        # Store the original column name to construct the potential null column name
+        col_name = column.name
+
+        # Create the dummy variables from the series
+        dummies = column.cast(pl.Utf8).to_dummies(drop_first=self.drop_first)
+
+        # Define the name of the column that Polars creates for null values
+        null_col_name = f"{col_name}_null"
+
+        # Check if the null column exists and drop it if it does
+        if null_col_name in dummies.columns:
+            return dummies.drop(null_col_name)
+
+        return dummies
 
 
 class MultiBinaryDummifier:
@@ -332,7 +401,7 @@ class MultiBinaryDummifier:
             A list of strings, where each string is a keyword to search for. A separate
             binary column will be created for each keyword.
         case_insensitive (bool):
-            If True, keyword matching ignores case. Defaults to True.
+            If True, keyword matching ignores case.
     """
     def __init__(self, keywords: List[str], case_insensitive: bool = True):
         if not isinstance(keywords, list) or not all(isinstance(k, str) for k in keywords):
@@ -355,11 +424,12 @@ class MultiBinaryDummifier:
         Returns:
             pl.DataFrame: A DataFrame where each column corresponds to a keyword.
         """
+        column_base_name = column.name
         # Ensure the input is treated as a string, preserving nulls
         str_column = column.cast(pl.Utf8)
         
         output_expressions = []
-        for i, keyword in enumerate(self.keywords):
+        for keyword in self.keywords:
             # Escape keyword to treat it as a literal, not a regex pattern
             base_pattern = re.escape(keyword)
 
@@ -373,7 +443,7 @@ class MultiBinaryDummifier:
                 .when(str_column.str.contains(pattern))
                 .then(pl.lit(1, dtype=pl.UInt8))
                 .otherwise(pl.lit(0, dtype=pl.UInt8))
-                .alias(f"col_{i}") # Generic name for DataProcessor
+                .alias(f"{column_base_name}_{keyword}") # name for DataProcessor
             )
             output_expressions.append(expr)
 
@@ -417,6 +487,7 @@ class KeywordDummifier:
         Returns:
             pl.DataFrame: A DataFrame with one-hot encoded columns.
         """
+        column_base_name = column.name
         column = column.cast(pl.Utf8)
         
         categorize_expr = pl.when(pl.lit(False)).then(pl.lit(None, dtype=pl.Utf8))
@@ -435,22 +506,24 @@ class KeywordDummifier:
                 column.str.contains(pattern)
             ).then(pl.lit(name))
         
-        categorize_expr = categorize_expr.otherwise(None).alias("category")
+        dummy_name = 'dummy_category'
+        
+        categorize_expr = categorize_expr.otherwise(None).alias(dummy_name)
 
         temp_df = pl.select(categorize_expr)
-        df_with_dummies = temp_df.to_dummies(columns=["category"])
+        df_with_dummies = temp_df.to_dummies(columns=[dummy_name])
         
         final_columns = []
         for name in self.group_names:
-            dummy_col_name = f"category_{name}"
+            dummy_col_name = f"{dummy_name}_{name}"
             if dummy_col_name in df_with_dummies.columns:
-                # The alias here uses the group name as the temporary column name
+                # The alias here uses the group name as the final column name
                 final_columns.append(
-                    df_with_dummies.get_column(dummy_col_name).alias(name)
+                    df_with_dummies.get_column(dummy_col_name).alias(f"{column_base_name}_{name}")
                 )
             else:
                 # If a group had no matches, create a column of zeros
-                final_columns.append(pl.lit(0, dtype=pl.UInt8).alias(name))
+                final_columns.append(pl.lit(0, dtype=pl.UInt8).alias(f"{column_base_name}_{name}"))
 
         return pl.select(final_columns)
 
@@ -471,7 +544,7 @@ class NumberExtractor:
         round_digits (int | None):
             If the dtype is 'float', you can specify the number of decimal
             places to round the result to. This parameter is ignored if
-            dtype is 'int'. Defaults to None (no rounding).
+            dtype is 'int'.
     """
     def __init__(
         self,
@@ -519,6 +592,7 @@ class NumberExtractor:
         Returns:
             pl.Series: A new Series containing the extracted numbers.
         """
+        column_base_name = column.name
         # Extract the first (and only) capturing group
         extracted = column.str.extract(self.regex_pattern, 1)
         
@@ -529,7 +603,7 @@ class NumberExtractor:
         if self.dtype == "float" and self.round_digits is not None:
             return casted.round(self.round_digits)
             
-        return casted
+        return casted.alias(column_base_name)
 
 
 class MultiNumberExtractor:
@@ -590,12 +664,13 @@ class MultiNumberExtractor:
         """
         Executes the multi-number extraction logic. Preserves nulls from the input column.
         """
+        column_base_name = column.name
         output_expressions = []
         for i in range(self.num_outputs):
             # Define the core extraction logic for the i-th number
             extraction_expr = (
                 column.str.extract_all(self.regex_pattern)
-                .list.get(i)
+                .list.get(i, null_on_oob=True)
                 .cast(self.polars_dtype, strict=False)
             )
 
@@ -609,24 +684,214 @@ class MultiNumberExtractor:
                 pl.when(column.is_not_null())
                 .then(extraction_expr)
                 .otherwise(None)
-                .alias(f"col_{i}") # Name the final output expression
+                .alias(f"{column_base_name}_{i}") # Name the final output expression
+            )
+            
+            output_expressions.append(final_expr)
+        
+        return pl.select(output_expressions)
+
+
+class TemperatureExtractor:
+    """
+    Extracts temperature values from a string column.
+
+    This transformer assumes that the source temperature values are in Celsius.
+    It can extract a single value using a specific regex or find all numbers in
+    a string and calculate their average. It also supports converting the final
+    Celsius value to Kelvin or Rankine.
+
+    Args:
+        regex_pattern (str):
+            The regex to find a single temperature. MUST contain exactly one
+            capturing group `(...)`. This is ignored if `average_mode` is True.
+        average_mode (bool):
+            If True, extracts all numbers from the string and returns their average.
+            This overrides the `regex_pattern` with a generic number-finding regex.
+        convert (str | None):
+            If "K", converts the final Celsius value to Kelvin.
+            If "R", converts the final Celsius value to Rankine.
+            If None (default), the value remains in Celsius.
+    """
+    def __init__(
+        self,
+        regex_pattern: str = r"(\d+\.?\d*)",
+        average_mode: bool = False,
+        convert: Optional[Literal["K", "R"]] = None,
+    ):
+        # --- Store configuration ---
+        self.average_mode = average_mode
+        self.convert = convert
+        self.regex_pattern = regex_pattern
+        
+        # Generic pattern for average mode, defined once for efficiency.
+        self._avg_mode_pattern = r"(\d+\.?\d*)"
+
+        # --- Validation ---
+        if not self.average_mode:
+            try:
+                if re.compile(self.regex_pattern).groups != 1:
+                    _LOGGER.error("'regex_pattern' must contain exactly one capturing group '(...)' for single extraction mode.")
+                    raise ValueError()
+            except re.error as e:
+                _LOGGER.error(f"Invalid regex pattern provided: {e}")
+                raise ValueError()
+
+        if self.convert is not None and self.convert not in ["K", "R"]:
+            _LOGGER.error("'convert' must be either 'K' (Kelvin) or 'R' (Rankine).")
+            raise ValueError()
+
+    def __call__(self, column: pl.Series) -> pl.Series:
+        """
+        Applies the temperature extraction and conversion logic.
+
+        Args:
+            column (pl.Series): The input Polars Series with string data.
+
+        Returns:
+            pl.Series: A new Series containing the final temperature values as floats.
+        """
+        column_base_name = column.name
+        # --- Step 1: Extract number(s) to get a Celsius value expression ---
+        if self.average_mode:
+            # Extract all numbers and compute their mean. Polars' list.mean()
+            # handles the casting to float automatically.
+            celsius_expr = (
+                column.str.extract_all(self._avg_mode_pattern)
+                .list.eval(pl.element().cast(pl.Float64, strict=False))
+                .list.mean()
+            )
+        else:
+            # Extract a single number using the specified pattern.
+            # Cast to Float64, with non-matches becoming null.
+            celsius_expr = column.str.extract(self.regex_pattern, 1).cast(pl.Float64, strict=False)
+
+        # --- Step 2: Apply conversion if specified ---
+        if self.convert == "K":
+            # Celsius to Kelvin: C + 273.15
+            final_expr = celsius_expr + 273.15
+        elif self.convert == "R":
+            # Celsius to Rankine: (C * 9/5) + 491.67
+            final_expr = (celsius_expr * 1.8) + 491.67
+        else:
+            # No conversion needed
+            final_expr = celsius_expr
+
+        # --- Step 3: Round the result and return as a Series ---
+        # The select().to_series() pattern is a robust way to execute an
+        # expression and guarantee a Series is returned.
+        return pl.select(final_expr.round(2)).to_series().alias(column_base_name)
+
+
+class MultiTemperatureExtractor:
+    """
+    Extracts multiple temperature values from a single string column into
+    several new columns, assuming the source values are in Celsius.
+
+    This one-to-many transformer is designed for cases where multiple readings
+    are packed into one field, like "Min: 10C, Max: 25C".
+
+    Args:
+        num_outputs (int):
+            The number of numeric columns to create.
+        regex_pattern (str):
+            The regex to find all numbers. Must contain exactly one capturing
+            group around the number part (e.g., r"(-?\\d+\\.?\\d*)").
+        convert (str | None):
+            If "K", converts the final Celsius values to Kelvin.
+            If "R", converts the final Celsius values to Rankine.
+            If None (default), the values remain in Celsius.
+        fill_value (int | float | None):
+            A value to use if a temperature is not found at a given position.
+            For example, if `num_outputs=3` and only two temperatures are
+            found, the third column will be filled with this value. If None,
+            it will be filled with null.
+    """
+    def __init__(
+        self,
+        num_outputs: int,
+        regex_pattern: str = r"(\d+\.?\d*)",
+        convert: Optional[Literal["K", "R"]] = None,
+        fill_value: Optional[Union[int, float]] = None
+    ):
+        # --- Validation ---
+        if not isinstance(num_outputs, int) or num_outputs <= 0:
+            _LOGGER.error("'num_outputs' must be a positive integer.")
+            raise ValueError()
+        
+        try:
+            if re.compile(regex_pattern).groups != 1:
+                _LOGGER.error("'regex_pattern' must contain exactly one capturing group '(...)'.")
+                raise ValueError()
+        except re.error as e:
+            _LOGGER.error(f"Invalid regex pattern provided: {e}")
+            raise ValueError()
+
+        if convert is not None and convert not in ["K", "R"]:
+            _LOGGER.error("'convert' must be either 'K' (Kelvin) or 'R' (Rankine).")
+            raise ValueError()
+
+        # --- Store configuration ---
+        self.num_outputs = num_outputs
+        self.regex_pattern = regex_pattern
+        self.convert = convert
+        self.fill_value = fill_value
+
+    def __call__(self, column: pl.Series) -> pl.DataFrame:
+        """
+        Applies the multi-temperature extraction and conversion logic.
+        """
+        column_base_name = column.name
+        output_expressions = []
+        for i in range(self.num_outputs):
+            # --- Step 1: Extract the i-th number as a Celsius value ---
+            celsius_expr = (
+                column.str.extract_all(self.regex_pattern)
+                .list.get(i, null_on_oob=True)
+                .cast(pl.Float64, strict=False)
+            )
+
+            # --- Step 2: Apply conversion if specified ---
+            if self.convert == "K":
+                # Celsius to Kelvin: C + 273.15
+                converted_expr = celsius_expr + 273.15
+            elif self.convert == "R":
+                # Celsius to Rankine: (C * 9/5) + 491.67
+                converted_expr = (celsius_expr * 1.8) + 491.67
+            else:
+                # No conversion needed
+                converted_expr = celsius_expr
+            
+            # --- Step 3: Apply fill value and handle original nulls ---
+            final_expr = converted_expr.round(2)
+            if self.fill_value is not None:
+                final_expr = final_expr.fill_null(self.fill_value)
+
+            # Ensure that if the original row was null, all outputs are null
+            final_expr = (
+                pl.when(column.is_not_null())
+                .then(final_expr)
+                .otherwise(None)
+                .alias(f"{column_base_name}_{i}") # Temporary name for DataProcessor
             )
             
             output_expressions.append(final_expr)
         
+        # Execute all expressions at once for performance
         return pl.select(output_expressions)
 
 
 class RatioCalculator:
     """
     A transformer that parses a string ratio (e.g., "40:5" or "30/2") and
-    computes the result of the division. It gracefully handles strings that
-    do not match the pattern by returning null.
+    computes the result of the division. Includes robust handling for
+    zeros and single numbers.
     """
     def __init__(
         self,
-        # Default pattern includes the full-width colon '：'
-        regex_pattern: str = r"(\d+\.?\d*)\s*[:：/]\s*(\d+\.?\d*)"
+        regex_pattern: str = r"(\d+\.?\d*)\s*[:：/]\s*(\d+\.?\d*)",
+        handle_zeros: bool = False,
+        handle_single_number: bool = False
     ):
         # --- Robust Validation ---
         try:
@@ -642,24 +907,119 @@ class RatioCalculator:
             raise ValueError()
 
         self.regex_pattern = regex_pattern
+        self.handle_zeros = handle_zeros
+        self.handle_single_number = handle_single_number
 
     def __call__(self, column: pl.Series) -> pl.Series:
         """
-        Applies the ratio calculation logic to the input column.
-        This version uses .str.extract() for maximum stability.
+        Applies the ratio calculation logic to the input column. Uses .str.extract() for maximum stability and includes optional handling for zeros and single numbers.
         """
+        column_base_name = column.name
         # Extract numerator (group 1) and denominator (group 2) separately.
         numerator_expr = column.str.extract(self.regex_pattern, 1).cast(pl.Float64, strict=False)
         denominator_expr = column.str.extract(self.regex_pattern, 2).cast(pl.Float64, strict=False)
 
-        # Calculate the ratio, handling division by zero.
-        final_expr = pl.when(denominator_expr != 0).then(
-            numerator_expr / denominator_expr
-        ).otherwise(
-            None # Handles both null denominators and division by zero
-        )
+        # --- Logic for Requirement A: Special zero handling ---
+        if self.handle_zeros:
+            ratio_expr = (
+                pl.when(numerator_expr.is_not_null() & denominator_expr.is_not_null())
+                .then(
+                    pl.when((numerator_expr == 0) & (denominator_expr == 0)).then(pl.lit(0.0))
+                    .when((numerator_expr != 0) & (denominator_expr == 0)).then(numerator_expr)
+                    .when((numerator_expr == 0) & (denominator_expr != 0)).then(denominator_expr)
+                    .otherwise(numerator_expr / denominator_expr)  # Default: both are non-zero
+                )
+            )
+        else:
+            # Original logic
+            ratio_expr = pl.when(denominator_expr != 0).then(
+                numerator_expr / denominator_expr
+            ).otherwise(
+                None # Handles null denominators and division by zero
+            )
+
+        # --- Logic for Requirement B: Handle single numbers as a fallback ---
+        if self.handle_single_number:
+            # Regex to match a string that is ONLY a valid float/int
+            single_number_regex = r"^\d+\.?\d*$"
+            single_number_expr = (
+                pl.when(column.str.contains(single_number_regex))
+                .then(column.cast(pl.Float64, strict=False))
+                .otherwise(None)
+            )
+            # If ratio_expr is null, try to fill it with single_number_expr
+            final_expr = ratio_expr.fill_null(single_number_expr)
+        else:
+            final_expr = ratio_expr
+
+        return pl.select(final_expr.round(4)).to_series().alias(column_base_name)
+
+
+class TriRatioCalculator:
+    """
+    A transformer that handles three-part ("A:B:C") ratios, enforcing a strict output structure.
+
+    - Three-part ratios produce A/B and A/C.
+    - Two-part ratios are assumed to be A:C and produce None for A/B.
+    - Single values produce None for both outputs.
+    """
+    def __init__(self, handle_zeros: bool = False):
+        """
+        Initializes the TriRatioCalculator.
 
-        return pl.select(final_expr.round(4)).to_series()
+        Args:
+            handle_zeros (bool): If True, returns a valid value if either the denominator or numerator is zero; returns zero if both are zero.
+        """
+        self.handle_zeros = handle_zeros
+
+    def _calculate_ratio(self, num: pl.Expr, den: pl.Expr) -> pl.Expr:
+        """Helper to contain the core division logic."""
+        if self.handle_zeros:
+            # Special handling for zeros
+            expr = (
+                pl.when((num == 0) & (den == 0)).then(pl.lit(0.0))
+                .when((num != 0) & (den == 0)).then(num) # Return numerator
+                .when((num == 0) & (den != 0)).then(den) # Return denominator
+                .otherwise(num / den)
+            )
+        else:
+            # Default behavior: return null if denominator is 0
+            expr = pl.when(den != 0).then(num / den).otherwise(None)
+
+        return expr.round(4)
+
+    def __call__(self, column: pl.Series) -> pl.DataFrame:
+        """
+        Applies the robust tri-ratio logic using the lazy API.
+        """
+        column_base_name = column.name
+        # Wrap the input Series in a DataFrame to use the lazy expression API
+        temp_df = column.to_frame()
+
+        # Define all steps as lazy expressions
+        all_numbers_expr = pl.col(column.name).str.extract_all(r"(\d+\.?\d*)")
+        num_parts_expr = all_numbers_expr.list.len()
+
+        expr_A = all_numbers_expr.list.get(0, null_on_oob=True).cast(pl.Float64)
+        expr_B = all_numbers_expr.list.get(1, null_on_oob=True).cast(pl.Float64)
+        expr_C = all_numbers_expr.list.get(2, null_on_oob=True).cast(pl.Float64)
+
+        # Define logic for each output column using expressions
+        ratio_ab_expr = pl.when(num_parts_expr == 3).then(
+            self._calculate_ratio(expr_A, expr_B)
+        ).otherwise(None)
+
+        ratio_ac_expr = pl.when(num_parts_expr == 3).then(
+            self._calculate_ratio(expr_A, expr_C)
+        ).when(num_parts_expr == 2).then(
+            self._calculate_ratio(expr_A, expr_B) # B is actually C in this case
+        ).otherwise(None)
+
+        # Execute the expressions and return the final DataFrame
+        return temp_df.select(
+            ratio_ab_expr.alias(f"{column_base_name}_A_to_B"),
+            ratio_ac_expr.alias(f"{column_base_name}_A_to_C")
+        )
 
 
 class CategoryMapper:
@@ -699,6 +1059,7 @@ class CategoryMapper:
         Returns:
             pl.Series: A new Series with categories mapped to numbers.
         """
+        column_base_name = column.name
         # Ensure the column is treated as a string for matching keys
         str_column = column.cast(pl.Utf8)
 
@@ -715,7 +1076,7 @@ class CategoryMapper:
             pl.lit(self.default_value)
         )
         
-        return pl.select(final_expr).to_series()
+        return pl.select(final_expr).to_series().alias(column_base_name)
 
 
 class RegexMapper:
@@ -779,6 +1140,7 @@ class RegexMapper:
             pl.Series: A new Series with strings mapped to numbers based on
                        the first matching regex pattern.
         """
+        column_base_name = column.name
         # pl.String is the modern alias for pl.Utf8
         str_column = column.cast(pl.String)
 
@@ -793,7 +1155,7 @@ class RegexMapper:
                 .otherwise(mapping_expr)
             )
         
-        return pl.select(mapping_expr).to_series()
+        return pl.select(mapping_expr).to_series().alias(column_base_name)
 
 
 class ValueBinner:
@@ -843,6 +1205,7 @@ class ValueBinner:
             pl.Series: A new Series of integer labels for the bins. Values
                        outside the specified breaks will become null.
         """
+        column_base_name = column.name
         # `cut` creates a new column of type Categorical
         binned_column = column.cut(
             breaks=self.breaks,
@@ -852,7 +1215,7 @@ class ValueBinner:
         
         # to_physical() converts the Categorical type to its underlying
         # integer representation (u32), which is perfect for ML.
-        return binned_column.to_physical()
+        return binned_column.to_physical().alias(column_base_name)
 
 
 class DateFeatureExtractor:
@@ -861,16 +1224,6 @@ class DateFeatureExtractor:
 
     It can handle columns that are already in a Polars Date/Datetime format,
     or it can parse string columns if a format is provided.
-
-    Args:
-        features (List[str]):
-            A list of the date/time features to extract. Supported features are:
-            'year', 'month', 'day', 'hour', 'minute', 'second', 'millisecond',
-            'microsecond', 'nanosecond', 'ordinal_day' (day of year),
-            'weekday' (Mon=1, Sun=7), 'week' (week of year), and 'timestamp'.
-        format (str | None):
-            The format code used to parse string dates (e.g., "%Y-%m-%d %H:%M:%S").
-            Use if the input column is not a Date or Datetime type.
     """
     
     ALLOWED_FEATURES = {
@@ -883,6 +1236,17 @@ class DateFeatureExtractor:
         features: List[str],
         format: Optional[str] = None,
     ):
+        """
+        Args:
+            features (List[str]):
+                A list of the date/time features to extract. Supported features are:
+                'year', 'month', 'day', 'hour', 'minute', 'second', 'millisecond',
+                'microsecond', 'nanosecond', 'ordinal_day' (day of year),
+                'weekday' (Mon=1, Sun=7), 'week' (week of year), 'timestamp'.
+            format (str | None):
+                The format code used to parse string dates (e.g., "%Y-%m-%d %H:%M:%S").
+                Use if the input column is not a Date or Datetime type.
+        """
         # --- Validation ---
         if not isinstance(features, list) or not features:
             _LOGGER.error("'features' must be a non-empty list of strings.")
@@ -906,6 +1270,7 @@ class DateFeatureExtractor:
         Returns:
             pl.DataFrame: A DataFrame with columns for each extracted feature.
         """
+        column_base_name = column.name
         date_col = column
         # First, parse strings into a datetime object if a format is given
         if self.format is not None:
@@ -921,10 +1286,81 @@ class DateFeatureExtractor:
                 expr = getattr(date_col.dt, feature)()
             
             # Alias with a generic name for the processor to handle
-            output_expressions.append(expr.alias(f"col_{i}"))
+            output_expressions.append(expr.alias(f"{column_base_name}_{feature}"))
             
         return pl.select(output_expressions)
 
 
+class MolecularFormulaTransformer:
+    """
+    Parses a Polars Series of molecular formula strings into a wide DataFrame.
+
+    This one-to-many transformer takes a column of condensed molecular formulas
+    (e.g., 'Li0.115Mn0.529Ni0.339O2') and converts it into a DataFrame where
+    each chemical element has its own column. The value in each column is the
+    stoichiometric quantity of that element.
+
+    It is designed to be used within the DataProcessor pipeline.
+    """
+
+    def __init__(self):
+        """
+        Initializes the transformer and pre-compiles the regex pattern.
+        """
+        # Sort symbols by length to prevent matching 'C' in 'Co'
+        sorted_symbols = sorted(CHEMICAL_ELEMENT_SYMBOLS, key=len, reverse=True)
+        
+        # Pre-compile regex for efficiency
+        self.pattern = re.compile(rf'({"|".join(sorted_symbols)})(\d*\.?\d*)')
+
+    def __call__(self, column: pl.Series) -> pl.DataFrame:
+        """
+        Executes the formula parsing logic.
+
+        Args:
+            column: A Polars Series containing strings of molecular formulas.
+
+        Returns:
+            A Polars DataFrame with columns for every chemical element.
+        """
+        column_base_name = column.name
+        def parse_formula(formula: str) -> dict:
+            """Helper to parse a single formula string into a dictionary."""
+            if not isinstance(formula, str) or not formula:
+                return {}
+            
+            matches = self.pattern.findall(formula)
+            
+            # This dict comprehension is correct for your use case where
+            # each element appears only once in the formula string.
+            return {
+                element: float(value) if value else 1.0 
+                for element, value in matches
+            }
+
+        # Apply the parsing function to each element
+        parsed_series = column.map_elements(parse_formula, return_dtype=pl.Object)
+        
+        # Convert the Series of dictionaries into a DataFrame
+        df = pl.DataFrame(parsed_series.to_list())
+        
+        # Ensure all possible element columns are created, filling with 0
+        select_expressions = []
+        for symbol in CHEMICAL_ELEMENT_SYMBOLS:
+            col_name = f"{column_base_name}_{symbol}"
+            if symbol in df.columns:
+                expr = pl.col(symbol).fill_null(0).alias(col_name)
+            else:
+                expr = pl.lit(0.0, dtype=pl.Float64).alias(col_name)
+            select_expressions.append(expr)
+            
+        # Handle edge case where input series is not empty but parsing yields no rows
+        base_df = df
+        if df.height == 0 and column.len() > 0:
+            base_df = pl.DataFrame({'dummy': range(column.len())})
+
+        return base_df.select(select_expressions)
+
+
 def info():
     _script_info(__all__)