workbench 0.8.213__py3-none-any.whl → 0.8.219__py3-none-any.whl

workbench/model_scripts/custom_models/uq_models/meta_uq.template

@@ -1,377 +0,0 @@
-# Model: NGBoost Regressor with Distribution output
-from ngboost import NGBRegressor
-from ngboost.distns import Cauchy
-from xgboost import XGBRegressor  # Point Estimator
-from sklearn.model_selection import train_test_split
-
-# Model Performance Scores
-from sklearn.metrics import mean_absolute_error, median_absolute_error, r2_score, root_mean_squared_error
-from scipy.stats import spearmanr
-
-from io import StringIO
-import json
-import argparse
-import joblib
-import os
-import numpy as np
-import pandas as pd
-from typing import List, Tuple
-
-# Local Imports
-from proximity import Proximity
-
-
-# Template Placeholders
-TEMPLATE_PARAMS = {
-    "id_column": "{{id_column}}",
-    "target": "{{target_column}}",
-    "features": "{{feature_list}}",
-    "compressed_features": "{{compressed_features}}",
-    "train_all_data": "{{train_all_data}}",
-    "track_columns": "{{track_columns}}",
-}
-
-
-# Function to check if dataframe is empty
-def check_dataframe(df: pd.DataFrame, df_name: str) -> None:
-    """
-    Check if the provided dataframe is empty and raise an exception if it is.
-
-    Args:
-        df (pd.DataFrame): DataFrame to check
-        df_name (str): Name of the DataFrame
-    """
-    if df.empty:
-        msg = f"*** The training data {df_name} has 0 rows! ***STOPPING***"
-        print(msg)
-        raise ValueError(msg)
-
-
-def match_features_case_insensitive(df: pd.DataFrame, model_features: list) -> pd.DataFrame:
-    """
-    Matches and renames DataFrame columns to match model feature names (case-insensitive).
-    Prioritizes exact matches, then case-insensitive matches.
-
-    Raises ValueError if any model features cannot be matched.
-    """
-    df_columns_lower = {col.lower(): col for col in df.columns}
-    rename_dict = {}
-    missing = []
-    for feature in model_features:
-        if feature in df.columns:
-            continue  # Exact match
-        elif feature.lower() in df_columns_lower:
-            rename_dict[df_columns_lower[feature.lower()]] = feature
-        else:
-            missing.append(feature)
-
-    if missing:
-        raise ValueError(f"Features not found: {missing}")
-
-    # Rename the DataFrame columns to match the model features
-    return df.rename(columns=rename_dict)
-
-
-def convert_categorical_types(df: pd.DataFrame, features: list, category_mappings={}) -> tuple:
-    """
-    Converts appropriate columns to categorical type with consistent mappings.
-
-    Args:
-        df (pd.DataFrame): The DataFrame to process.
-        features (list): List of feature names to consider for conversion.
-        category_mappings (dict, optional): Existing category mappings. If empty dict, we're in
-                                            training mode. If populated, we're in inference mode.
-
-    Returns:
-        tuple: (processed DataFrame, category mappings dictionary)
-    """
-    # Training mode
-    if category_mappings == {}:
-        for col in df.select_dtypes(include=["object", "string"]):
-            if col in features and df[col].nunique() < 20:
-                print(f"Training mode: Converting {col} to category")
-                df[col] = df[col].astype("category")
-                category_mappings[col] = df[col].cat.categories.tolist()  # Store category mappings
-
-    # Inference mode
-    else:
-        for col, categories in category_mappings.items():
-            if col in df.columns:
-                print(f"Inference mode: Applying categorical mapping for {col}")
-                df[col] = pd.Categorical(df[col], categories=categories)  # Apply consistent categorical mapping
-
-    return df, category_mappings
-
-
-def decompress_features(
-    df: pd.DataFrame, features: List[str], compressed_features: List[str]
-) -> Tuple[pd.DataFrame, List[str]]:
-    """Prepare features for the model by decompressing bitstring features
-
-    Args:
-        df (pd.DataFrame): The features DataFrame
-        features (List[str]): Full list of feature names
-        compressed_features (List[str]): List of feature names to decompress (bitstrings)
-
-    Returns:
-        pd.DataFrame: DataFrame with the decompressed features
-        List[str]: Updated list of feature names after decompression
-
-    Raises:
-        ValueError: If any missing values are found in the specified features
-    """
-
-    # Check for any missing values in the required features
-    missing_counts = df[features].isna().sum()
-    if missing_counts.any():
-        missing_features = missing_counts[missing_counts > 0]
-        print(
-            f"WARNING: Found missing values in features: {missing_features.to_dict()}. "
-            "WARNING: You might want to remove/replace all NaN values before processing."
-        )
-
-    # Decompress the specified compressed features
-    decompressed_features = features.copy()
-    for feature in compressed_features:
-        if (feature not in df.columns) or (feature not in features):
-            print(f"Feature '{feature}' not in the features list, skipping decompression.")
-            continue
-
-        # Remove the feature from the list of features to avoid duplication
-        decompressed_features.remove(feature)
-
-        # Handle all compressed features as bitstrings
-        bit_matrix = np.array([list(bitstring) for bitstring in df[feature]], dtype=np.uint8)
-        prefix = feature[:3]
-
-        # Create all new columns at once - avoids fragmentation
-        new_col_names = [f"{prefix}_{i}" for i in range(bit_matrix.shape[1])]
-        new_df = pd.DataFrame(bit_matrix, columns=new_col_names, index=df.index)
-
-        # Add to features list
-        decompressed_features.extend(new_col_names)
-
-        # Drop original column and concatenate new ones
-        df = df.drop(columns=[feature])
-        df = pd.concat([df, new_df], axis=1)
-
-    return df, decompressed_features
-
-
-if __name__ == "__main__":
-    # Template Parameters
-    id_column = TEMPLATE_PARAMS["id_column"]
-    target = TEMPLATE_PARAMS["target"]
-    features = TEMPLATE_PARAMS["features"]
-    orig_features = features.copy()
-    compressed_features = TEMPLATE_PARAMS["compressed_features"]
-    train_all_data = TEMPLATE_PARAMS["train_all_data"]
-    track_columns = TEMPLATE_PARAMS["track_columns"]  # Can be None
-    validation_split = 0.2
-
-    # Script arguments for input/output directories
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--model-dir", type=str, default=os.environ.get("SM_MODEL_DIR", "/opt/ml/model"))
-    parser.add_argument("--train", type=str, default=os.environ.get("SM_CHANNEL_TRAIN", "/opt/ml/input/data/train"))
-    parser.add_argument(
-        "--output-data-dir", type=str, default=os.environ.get("SM_OUTPUT_DATA_DIR", "/opt/ml/output/data")
-    )
-    args = parser.parse_args()
-
-    # Read the training data into DataFrames
-    training_files = [os.path.join(args.train, file) for file in os.listdir(args.train) if file.endswith(".csv")]
-    print(f"Training Files: {training_files}")
-
-    # Combine files and read them all into a single pandas dataframe
-    all_df = pd.concat([pd.read_csv(file, engine="python") for file in training_files])
-
-    # Check if the dataframe is empty
-    check_dataframe(all_df, "training_df")
-
-    # Features/Target output
-    print(f"Target: {target}")
-    print(f"Features: {str(features)}")
-
-    # Convert any features that might be categorical to 'category' type
-    all_df, category_mappings = convert_categorical_types(all_df, features)
-
-    # If we have compressed features, decompress them
-    if compressed_features:
-        print(f"Decompressing features {compressed_features}...")
-        all_df, features = decompress_features(all_df, features, compressed_features)
-
-    # Do we want to train on all the data?
-    if train_all_data:
-        print("Training on ALL of the data")
-        df_train = all_df.copy()
-        df_val = all_df.copy()
-
-    # Does the dataframe have a training column?
-    elif "training" in all_df.columns:
-        print("Found training column, splitting data based on training column")
-        df_train = all_df[all_df["training"]]
-        df_val = all_df[~all_df["training"]]
-    else:
-        # Just do a random training Split
-        print("WARNING: No training column found, splitting data with random state=42")
-        df_train, df_val = train_test_split(all_df, test_size=validation_split, random_state=42)
-    print(f"FIT/TRAIN: {df_train.shape}")
-    print(f"VALIDATION: {df_val.shape}")
-
-    # We're using XGBoost for point predictions and NGBoost for uncertainty quantification
-    xgb_model = XGBRegressor()
-    ngb_model = NGBRegressor()  # Dist=Cauchy) Seems to give HUGE prediction intervals
-
-    # Prepare features and targets for training
-    X_train = df_train[features]
-    X_validate = df_val[features]
-    y_train = df_train[target]
-    y_validate = df_val[target]
-
-    # Train both models using the training data
-    xgb_model.fit(X_train, y_train)
-    ngb_model.fit(X_train, y_train, X_val=X_validate, Y_val=y_validate)
-
-    # Make Predictions on the Validation Set
-    print(f"Making Predictions on Validation Set...")
-    preds = xgb_model.predict(X_validate)
-
-    # Calculate various model performance metrics (regression)
-    rmse = root_mean_squared_error(y_validate, preds)
-    mae = mean_absolute_error(y_validate, preds)
-    medae = median_absolute_error(y_validate, preds)
-    r2 = r2_score(y_validate, preds)
-    spearman_corr = spearmanr(y_validate, preds).correlation
-    support = len(df_val)
-    print(f"rmse: {rmse:.3f}")
-    print(f"mae: {mae:.3f}")
-    print(f"medae: {medae:.3f}")
-    print(f"r2: {r2:.3f}")
-    print(f"spearmanr: {spearman_corr:.3f}")
-    print(f"support: {support}")
-
-    # Save the trained XGBoost model
-    xgb_model.save_model(os.path.join(args.model_dir, "xgb_model.json"))
-
-    # Save the trained NGBoost model
-    joblib.dump(ngb_model, os.path.join(args.model_dir, "ngb_model.joblib"))
-
-    # Save the features (this will validate input during predictions)
-    with open(os.path.join(args.model_dir, "feature_columns.json"), "w") as fp:
-        json.dump(orig_features, fp)  # We save the original features, not the decompressed ones
-
-    # Now the Proximity model
-    model = Proximity(df_train, id_column, features, target, track_columns=track_columns)
-
-    # Now serialize the model
-    model.serialize(args.model_dir)
-
-
-#
-# Inference Section
-#
-def model_fn(model_dir) -> dict:
-    """Load and return XGBoost, NGBoost, and Prox Model from model directory."""
-
-    # Load XGBoost regressor
-    xgb_path = os.path.join(model_dir, "xgb_model.json")
-    xgb_model = XGBRegressor(enable_categorical=True)
-    xgb_model.load_model(xgb_path)
-
-    # Load NGBoost regressor
-    ngb_model = joblib.load(os.path.join(model_dir, "ngb_model.joblib"))
-
-    # Deserialize the proximity model
-    prox_model = Proximity.deserialize(model_dir)
-
-    return {"xgboost": xgb_model, "ngboost": ngb_model, "proximity": prox_model}
-
-
-def input_fn(input_data, content_type):
-    """Parse input data and return a DataFrame."""
-    if not input_data:
-        raise ValueError("Empty input data is not supported!")
-
-    # Decode bytes to string if necessary
-    if isinstance(input_data, bytes):
-        input_data = input_data.decode("utf-8")
-
-    if "text/csv" in content_type:
-        return pd.read_csv(StringIO(input_data))
-    elif "application/json" in content_type:
-        return pd.DataFrame(json.loads(input_data))  # Assumes JSON array of records
-    else:
-        raise ValueError(f"{content_type} not supported!")
-
-
-def output_fn(output_df, accept_type):
-    """Supports both CSV and JSON output formats."""
-    if "text/csv" in accept_type:
-        csv_output = output_df.fillna("N/A").to_csv(index=False)  # CSV with N/A for missing values
-        return csv_output, "text/csv"
-    elif "application/json" in accept_type:
-        return output_df.to_json(orient="records"), "application/json"  # JSON array of records (NaNs -> null)
-    else:
-        raise RuntimeError(f"{accept_type} accept type is not supported by this script.")
-
-
-def predict_fn(df, models) -> pd.DataFrame:
-    """Make Predictions with our XGB Quantile Regression Model
-
-    Args:
-        df (pd.DataFrame): The input DataFrame
-        models (dict): The dictionary of models to use for predictions
-
-    Returns:
-        pd.DataFrame: The DataFrame with the predictions added
-    """
-
-    # Grab our feature columns (from training)
-    model_dir = os.environ.get("SM_MODEL_DIR", "/opt/ml/model")
-    with open(os.path.join(model_dir, "feature_columns.json")) as fp:
-        model_features = json.load(fp)
-
-    # Match features in a case-insensitive manner
-    matched_df = match_features_case_insensitive(df, model_features)
-
-    # Use XGBoost for point predictions
-    df["prediction"] = models["xgboost"].predict(matched_df[model_features])
-
-    # NGBoost predict returns distribution objects
-    y_dists = models["ngboost"].pred_dist(matched_df[model_features])
-
-    # Extract parameters from distribution
-    dist_params = y_dists.params
-
-    # Extract mean and std from distribution parameters
-    df["prediction_uq"] = dist_params["loc"]  # mean
-    df["prediction_std"] = dist_params["scale"]  # standard deviation
-
-    # Add 95% prediction intervals using ppf (percent point function)
-    # Note: Our hybrid model uses XGB point prediction and NGBoost UQ
-    #  so we need to adjust the bounds to include the point prediction
-    df["q_025"] = np.minimum(y_dists.ppf(0.025), df["prediction"])
-    df["q_975"] = np.maximum(y_dists.ppf(0.975), df["prediction"])
-
-    # Add 90% prediction intervals
-    df["q_05"] = y_dists.ppf(0.05)  # 5th percentile
-    df["q_95"] = y_dists.ppf(0.95)  # 95th percentile
-
-    # Add 80% prediction intervals
-    df["q_10"] = y_dists.ppf(0.10)  # 10th percentile
-    df["q_90"] = y_dists.ppf(0.90)  # 90th percentile
-
-    # Add 50% prediction intervals
-    df["q_25"] = y_dists.ppf(0.25)  # 25th percentile
-    df["q_75"] = y_dists.ppf(0.75)  # 75th percentile
-
-    # Reorder the quantile columns for easier reading
-    quantile_cols = ["q_025", "q_05", "q_10", "q_25", "q_75", "q_90", "q_95", "q_975"]
-    other_cols = [col for col in df.columns if col not in quantile_cols]
-    df = df[other_cols + quantile_cols]
-
-    # Compute Nearest neighbors with Proximity model
-    models["proximity"].neighbors(df)
-
-    # Return the modified DataFrame
-    return df