Utility Functions

def build_cox_model(df, duration_col, event_col, crossval=False, n_splits=5, random_state=42):
    """
    Fits a Cox Proportional Hazards model to the data with optional cross-validation.

    Parameters:
    - df: Pandas DataFrame containing features, survival times, and event indicators.
    - duration_col: Column name for survival times.
    - event_col: Column name for event indicator (1 = event occurred, 0 = censored).
    - crossval: If True, performs K-fold cross-validation and returns average C-index.
    - n_splits: Number of folds for cross-validation.
    - random_state: Random seed for reproducibility.

    Returns:
    - cox_model: Fitted CoxPH model on the full data.
    - mean_c_index (optional): Mean C-index from cross-validation if crossval=True.
    """

    def remove_low_variance_survival_features(df, duration_col, event_col, threshold=0.01):
        events = df[event_col].astype(bool)
        low_variance_features = []

        for feature in df.drop(columns=[duration_col, event_col]).columns:
            variance_when_event = df.loc[events, feature].var()
            variance_when_no_event = df.loc[~events, feature].var()
            if variance_when_event < threshold or variance_when_no_event < threshold:
                low_variance_features.append(feature)

        df_filtered = df.drop(columns=low_variance_features)
        if low_variance_features:
            print("Removed low variance features:", low_variance_features)
        return df_filtered

    df = remove_low_variance_survival_features(df, duration_col, event_col)

    if crossval:
        kf = KFold(n_splits=n_splits, shuffle=True, random_state=random_state)
        c_indices = []

        for train_idx, test_idx in kf.split(df):
            train_df = df.iloc[train_idx]
            test_df = df.iloc[test_idx]

            model = CoxPHFitter()
            model.fit(train_df, duration_col=duration_col, event_col=event_col)

            # Compute risk scores on test set and C-index
            risk_scores = model.predict_partial_hazard(test_df)
            ci = concordance_index(test_df[duration_col], -risk_scores, test_df[event_col])
            c_indices.append(ci)

        mean_c_index = np.mean(c_indices)
        print(f"Cross-validated C-index (mean over {n_splits} folds): {mean_c_index:.3f}")
    else:
        mean_c_index = None

    # Fit on full data
    final_model = CoxPHFitter()
    final_model.fit(df, duration_col=duration_col, event_col=event_col)

    return final_model

def compute_transport_cost(embeddings, batch_labels, blur=0.5):
    """
    Compute a transport cost using Sinkhorn loss to align embeddings between batches.

    Parameters:
    - embeddings (torch.Tensor): Tensor of embeddings (shape: [num_samples, num_features]).
    - batch_labels (torch.Tensor): Tensor of batch labels (shape: [num_samples]).
    - blur (float): Regularization parameter for Sinkhorn OT.

    Returns:
    - loss (torch.Tensor): Sinkhorn loss value.
    """
    # Ensure batch labels are integers
    batch_labels = batch_labels.long()

    # Split embeddings by batch
    batch1_embeddings = embeddings[batch_labels == 0]
    batch2_embeddings = embeddings[batch_labels == 1]

    if batch1_embeddings.size(0) == 0 or batch2_embeddings.size(0) == 0:
        raise ValueError("Both batches must have at least one sample for transport cost computation.")

    # Initialize the Sinkhorn loss function
    loss_fn = SamplesLoss("sinkhorn", blur=blur)

    # Compute the Sinkhorn loss between the two batches
    loss = loss_fn(batch1_embeddings, batch2_embeddings)

    return loss

def create_covariate_matrix(covariates, variable_types, ann):
    """
    Convert clinical variables used as covariates into a covariate matrix as a Pandas DataFrame.
    Missing values in numerical variables are imputed using the median.

    Args:
        covariates (list of str): List of variable names that must exist in the "clin.csv".
        variable_types (dict): Dictionary mapping variable names to their types ('categorical' or 'numerical').
        ann (pd.DataFrame): Annotation DataFrame containing batch variable values.

    Returns:
        pd.DataFrame: A covariate matrix DataFrame where categorical variables are one-hot-encoded as 0/1 and numerical variables are imputed,
                      with features as rows and samples as columns.
    """
    covariate_features = []
    feature_names = []

    for var in covariates:
        if variable_types.get(var) == 'categorical':
            # One-hot-encode categorical variables with 0/1 encoding
            one_hot = pd.get_dummies(ann[var], prefix=var).astype(int)
            covariate_features.append(one_hot.T)  # Transpose to make features rows
            feature_names.extend(one_hot.columns.tolist())
        elif variable_types.get(var) == 'numerical':
            # Handle numerical variables with missing values
            numerical_data = ann[[var]].copy()
            # Impute missing values using the median and assign back
            numerical_data[var] = numerical_data[var].fillna(numerical_data[var].median())
            covariate_features.append(numerical_data.T)  # Transpose to make features rows
            feature_names.append(var)
        else:
            raise ValueError(f"Unknown variable type for {var}: {variable_types.get(var)}")

    # Concatenate all covariate features into a single DataFrame
    covariate_matrix = pd.concat(covariate_features, axis=0)

    # Ensure row order matches appended feature names and preserve sample names as columns
    covariate_matrix.index = feature_names
    covariate_matrix.columns = ann.index

    return covariate_matrix

def create_device_from_string(device_str):
    """
    Create a torch.device object from device string, with MPS support.

    Args:
        device_str (str): Device string ('cuda', 'mps', 'cpu', 'gpu', 'auto')

    Returns:
        torch.device: PyTorch device object
    """
    if device_str in ['gpu', 'auto']:
        optimal_device, _ = get_optimal_device()
        return torch.device(optimal_device)
    elif device_str == 'mps':
        if torch.backends.mps.is_available():
            return torch.device('mps')
        else:
            warnings.warn("MPS not available, falling back to CPU")
            return torch.device('cpu')
    elif device_str == 'cuda':
        if torch.cuda.is_available():
            return torch.device('cuda')
        else:
            warnings.warn("CUDA not available, falling back to CPU")
            return torch.device('cpu')
    else:
        return torch.device('cpu')

def evaluate_baseline_performance(train_dataset, test_dataset, variable_name, methods, n_folds=5, n_jobs=4, use_pca=False, n_components=100):
    """
    Evaluates the performance of machine learning models on a given variable with optional PCA for dimensionality reduction.

    Args:
        train_dataset (Dataset): A MultiOmicDataset object containing training data and metadata.
        test_dataset (Dataset): A MultiOmicDataset object containing testing data.
        variable_name (str): The name of the target variable for prediction.
        methods (list of str): List of machine learning methods to evaluate, e.g., ['RandomForest', 'SVM', 'XGBoost'].
        n_folds (int, optional): Number of folds for K-fold cross-validation. Defaults to 5.
        n_jobs (int, optional): Number of jobs to run in parallel. Defaults to 4.
        use_pca (bool, optional): Whether to apply PCA for dimensionality reduction. Defaults to False.
        n_components (int, optional): Number of principal components to keep if PCA is applied. Defaults to 50.

    Returns:
        pd.DataFrame: A DataFrame containing metrics for each method.
    """
    def prepare_data(data_object, pca_model=None, fit_pca=False):
        # Concatenate Data Matrices
        X = np.concatenate([tensor for tensor in data_object.dat.values()], axis=1)
        y = np.array(data_object.ann[variable_name])

        # Filter out samples without a valid label
        valid_indices = ~np.isnan(y)
        X = X[valid_indices]
        y = y[valid_indices]

        if use_pca:
            if fit_pca:
                pca_model.fit(X)
                print(f"PCA fitted: Reduced to {n_components} components")
            X = pca_model.transform(X)
            print(f"PCA applied: Transformed to {n_components} components")

        return X, y, np.where(valid_indices)[0]

    # Initialize PCA model if PCA is used
    pca_model = PCA(n_components=n_components) if use_pca else None

    # Determine variable type
    variable_type = train_dataset.variable_types[variable_name]

    # Cross-Validation and Training
    kf = KFold(n_splits=n_folds, shuffle=True, random_state=42)
    X_train, y_train, train_indices = prepare_data(train_dataset, pca_model=pca_model, fit_pca=True)
    print("Train:", X_train.shape)
    X_test, y_test, test_indices = prepare_data(test_dataset, pca_model=pca_model, fit_pca=False)
    print("Test:", X_test.shape)

    metrics_list = []
    predictions = []  # Collect all predictions

    for method in methods:
        if variable_type == 'categorical':
            if method == 'RandomForest':
                model = RandomForestClassifier(random_state=42)
                params = {'n_estimators': [100, 200, 300], 'max_depth': [10, 20, None]}
            elif method == 'SVM':
                model = SVC(probability=True, random_state=42)
                params = {'C': [0.1, 1, 10], 'kernel': ['rbf', 'poly']}
            elif method == 'XGBoost':
                model = XGBClassifier(eval_metric='logloss', random_state=42)
                params = {'n_estimators': [100, 200, 300], 'max_depth': [3, 6, 9], 'learning_rate': [0.01, 0.1, 0.2]}
        elif variable_type == 'numerical':
            if method == 'RandomForest':
                model = RandomForestRegressor(random_state=42)
                params = {'n_estimators': [100, 200, 300], 'max_depth': [10, 20, None]}
            elif method == 'SVM':
                model = SVR()
                params = {'C': [0.1, 1, 10], 'kernel': ['rbf', 'poly']}
            elif method == 'XGBoost':
                model = XGBRegressor(random_state=42)
                params = {'n_estimators': [100, 200, 300], 'max_depth': [3, 6, 9], 'learning_rate': [0.01, 0.1, 0.2]}

        print("Training method:", method)
        grid_search = GridSearchCV(model, params, cv=kf, n_jobs=n_jobs)
        grid_search.fit(X_train, y_train)
        best_model = grid_search.best_estimator_

        # Predict on test data
        if variable_type == 'categorical':
            y_probs = best_model.predict_proba(X_test)
            metrics = evaluate_classifier(y_test, y_probs, print_report=True)
            y_pred_dict = {variable_name: y_probs}
        elif variable_type == 'numerical':
            y_pred = best_model.predict(X_test)
            metrics = evaluate_regressor(y_test, y_pred)
            y_pred_dict = {variable_name: y_pred}

        # need to get test indices to only consider samples with labels
        df_preds = get_predicted_labels(y_pred_dict, test_dataset.subset(test_indices), 'test', method)
        predictions.append(df_preds)

        for metric, value in metrics.items():
            metrics_list.append({
                'method': method + ('Classifier' if variable_type == 'categorical' else 'Regressor'),
                'var': variable_name,
                'variable_type': variable_type,
                'metric': metric,
                'value': value
            })

    predictions = pd.concat(predictions, ignore_index=True)

    return pd.DataFrame(metrics_list), predictions

def evaluate_baseline_survival_performance(train_dataset, test_dataset, duration_col, event_col, n_folds=5, n_jobs=4):
    """
    Evaluates the baseline performance of a Random Survival Forest model on survival data using the Concordance Index.

    The function preprocesses both training and testing datasets to prepare appropriate survival data (comprising durations 
    and event occurrences), performs cross-validation to assess model robustness, and then calculates the Concordance Index on 
    the test data. It uses a Random Survival Forest (RSF) as the predictive model.

    Args:
        train_dataset (Dataset): The training dataset (a MultiOmicDataset object) containing features and survival data.
        test_dataset (Dataset): The testing dataset  (a MultiOmicDataset object) containing features and survival data.
        duration_col (str): Column name in the dataset for survival time.
        event_col (str): Column name in the dataset for the event occurrence (1 if event occurred, 0 otherwise).
        n_folds (int, optional): Number of folds for K-fold cross-validation. Defaults to 5.
        n_jobs (int, optional): Number of parallel jobs to run for Random Survival Forest training. Defaults to 4.

    Returns:
        pd.DataFrame: A DataFrame containing the performance metrics of the RSF model, specifically the Concordance Index,
                      listed along with the method name and variable details.

    """
    print(f"[INFO] Evaluating baseline survival prediction performance")
    def prepare_data(data_object, duration_col, event_col):
        # Concatenate Data Matrices
        X = np.concatenate([tensor for tensor in data_object.dat.values()], axis=1)

        # Prepare Survival Data (Durations and Events)
        durations = np.array(data_object.ann[duration_col])
        events = np.array(data_object.ann[event_col])
        y = np.array([(event, duration) for event, duration in zip(events, durations)], 
                     dtype=[('Event', '?'), ('Time', '<f8')])

        # Filter out samples without a valid survival data
        valid_indices = ~np.isnan(durations) & ~np.isnan(events)
        X = X[valid_indices]
        y = y[valid_indices]
        return X, y, np.where(valid_indices)[0]

    # Prepare train and test data
    X_train, y_train, train_indices = prepare_data(train_dataset, duration_col, event_col)
    X_test, y_test, test_indices = prepare_data(test_dataset, duration_col, event_col)

    # Initialize Random Survival Forest
    rsf = RandomSurvivalForest(n_estimators=100, max_depth=5, min_samples_split=10,
                               min_samples_leaf=15, max_features="sqrt", n_jobs=n_jobs, random_state=42)

    # Cross-Validation to determine the best model
    kf = KFold(n_splits=n_folds, shuffle=True, random_state=42)
    c_index_scores = []

    for train_index, test_index in kf.split(X_train):
        X_fold_train, X_fold_test = X_train[train_index], X_train[test_index]
        y_fold_train, y_fold_test = y_train[train_index], y_train[test_index]

        rsf.fit(X_fold_train, y_fold_train)
        prediction = rsf.predict(X_fold_test)
        c_index = concordance_index_censored(y_fold_test['Event'], y_fold_test['Time'], prediction)
        c_index_scores.append(c_index[0])

    # Calculate average C-index across all folds
    avg_c_index = np.mean(c_index_scores)
    print(f"[INFO] Average C-index in cross-validation: {avg_c_index}")

    # Retrain on full training data and evaluate on test data
    rsf.fit(X_train, y_train)
    test_prediction = rsf.predict(X_test)
    test_c_index = concordance_index_censored(y_test['Event'], y_test['Time'], test_prediction)
    print(f"[INFO] C-index on test data: {test_c_index[0]}")


    # need to get test indices to only consider samples with labels
    predicted_labels = get_predicted_labels({event_col: test_prediction}, test_dataset.subset(test_indices), 'test', 'RandomSurvivalForest')
    # Reporting
    metrics_list = [{
        'method': 'RandomSurvivalForest',
        'var': event_col, 
        'variable_type': 'numerical',
        'metric': 'cindex',
        'value': test_c_index[0]
    }]

    return pd.DataFrame(metrics_list), predicted_labels

def evaluate_classifier(y_true, y_probs, print_report=False):
    """
    Evaluate the performance of a classifier using multiple metrics and optionally print a detailed classification report.

    This function computes balanced accuracy, F1 score (weighted), Cohen's Kappa score, average AUROC score, and 
    weighted-average AUC-PR score for the given true labels and predicted probabilities.
    If `print_report` is set to True, it prints a detailed classification report.

    Args:
        y_true (array-like): True labels of the data, must be a 1D list or array of labels.
        y_probs (array-like): Predicted probabilities for each class, must be 2D (n_samples, n_classes).
        print_report (bool, optional): If True, prints a detailed classification report. Defaults to False.

    Returns:
        dict: A dictionary containing:
              - 'balanced_acc': The balanced accuracy of the predictions.
              - 'f1_score': The weighted-average F1 score of the predictions.
              - 'kappa': Cohen's Kappa score indicating the level of agreement between the true and predicted labels.
              - 'average_auroc': The weighted average AUROC score across all classes.
              - 'average_aupr': The weighted average AUC-PR score across all classes.
    """
    # Convert probabilities to predicted labels
    y_pred = np.argmax(y_probs, axis=1)

    # Balanced accuracy
    balanced_acc = balanced_accuracy_score(y_true, y_pred)

    # F1 score (weighted)
    f1 = f1_score(y_true, y_pred, average='weighted', zero_division=0)

    # Cohen's Kappa
    kappa = cohen_kappa_score(y_true, y_pred)

    # Compute AUROC/AUPR 
    try:
        if y_probs.shape[1] == 2:  # Binary classification
            y_probs_binary = y_probs[:, 1]  # Use positive class probabilities
            average_auroc = roc_auc_score(y_true, y_probs_binary)
            average_aupr = average_precision_score(y_true, y_probs_binary)  # AUC-PR for binary case
        else:  # Multiclass classification
            average_auroc = roc_auc_score(y_true, y_probs, multi_class='ovr', average='weighted')
            average_aupr = average_precision_score(y_true, y_probs, average='weighted')  # Weighted AUC-PR for multiclass
    except ValueError:
        average_auroc = None  # Handle cases where AUROC cannot be computed
        average_aupr = None  # Handle cases where AUC-PR cannot be computed

    # Full classification report
    if print_report:
        print("\nClassification Report:")
        report = classification_report(y_true, y_pred, zero_division=0)
        print(report)

    return {
        "balanced_acc": balanced_acc,
        "f1_score": f1,
        "kappa": kappa,
        "average_auroc": average_auroc,
        "average_aupr": average_aupr  # Added AUC-PR
    }

def evaluate_regressor(y_true, y_pred):
    """
    Evaluate the performance of a regression model using mean squared error, R-squared, and Pearson correlation coefficient.

    This function computes the mean squared error (MSE) between true and predicted values as a measure of prediction accuracy.
    It also performs a linear regression analysis between the true and predicted values to obtain the R-squared value, which
    explains the variance ratio, and the Pearson correlation coefficient, providing insight into the linear relationship strength.

    Args:
        y_true (array-like): True values of the dependent variable, must be a 1D list or array.
        y_pred (array-like): Predicted values as returned by a regressor, must match the dimensions of y_true.

    Returns:
        dict: A dictionary containing:
              - 'mse': The mean squared error between the true and predicted values.
              - 'r2': The R-squared value indicating the proportion of variance in the dependent variable predictable from the independent variable.
              - 'pearson_corr': The Pearson correlation coefficient indicating the linear relationship strength between the true and predicted values.
    """
    mse = mean_squared_error(y_true, y_pred)
    slope, intercept, r_value, p_value, std_err = linregress(y_true,y_pred)
    r2 = r_value**2 
    return {"mse": mse, "r2": r2, "pearson_corr": r_value}

def evaluate_survival(outputs, durations, events):
    """
    Computes the concordance index (c-index) for survival predictions.

    Parameters:
    - outputs: Predicted risk scores (torch.Tensor or np.array).
    - durations: True survival times or durations (torch.Tensor or np.array).
    - events: Event indicators (1=event, 0=censored) (torch.Tensor or np.array).

    Returns:
    - A dictionary containing the c-index.
    """
    # Ensure all inputs are NumPy arrays
    if isinstance(durations, torch.Tensor):
        durations = durations.numpy()
    if isinstance(events, torch.Tensor):
        events = events.numpy()
    if isinstance(outputs, torch.Tensor):
        outputs = outputs.numpy()

    # Remove entries with NaNs
    valid_mask = ~np.isnan(durations) & ~np.isnan(events)
    if valid_mask.sum() > 0:
        durations = durations[valid_mask]
        events = events[valid_mask]
        outputs = outputs[valid_mask]

    # Compute concordance index (lifelines expects higher risk → lower survival)
    c_index = concordance_index(durations, -outputs, events)
    return {'cindex': c_index}

def evaluate_wrapper(method, y_pred_dict, dataset, surv_event_var = None, surv_time_var = None):
    """
    Evaluates predictions for different variables within a dataset using appropriate metrics based on the variable type. 
    Supports evaluation for numerical, categorical, and survival data.

    This function loops through each variable in the predictions dictionary, determines the type of the variable,
    and evaluates the predictions using the appropriate method: regression, classification, or survival analysis.
    It compiles the metrics into a list of dictionaries, which is then converted into a pandas DataFrame.

    Args:
        method (str): Identifier for the prediction method or model used.
        y_pred_dict (dict): A dictionary where keys are variable names and values are arrays of predicted values.
        dataset (Dataset): A dataset object containing actual values and metadata such as variable types.
        surv_event_var (str, optional): The name of the survival event variable. Required if survival analysis is performed.
        surv_time_var (str, optional): The name of the survival time variable. Required if survival analysis is performed.

    Returns:
        pd.DataFrame: A DataFrame where each row contains the method, variable name, variable type, metric name, and metric value.

    """
    metrics_list = []
    for var in y_pred_dict.keys():
        if dataset.variable_types[var] == 'numerical':
            if var == surv_event_var:
                events = dataset.ann[surv_event_var]
                durations = dataset.ann[surv_time_var]
                metrics = evaluate_survival(y_pred_dict[var], durations, events)
            else:
                ind = ~torch.isnan(dataset.ann[var])
                metrics = evaluate_regressor(dataset.ann[var][ind], y_pred_dict[var][ind].flatten())
        else:
            ind = ~torch.isnan(dataset.ann[var])
            metrics = evaluate_classifier(dataset.ann[var][ind], y_pred_dict[var][ind])

        for metric, value in metrics.items():
            metrics_list.append({
                'method': method,
                'var': var,
                'variable_type': dataset.variable_types[var],
                'metric': metric,
                'value': value
            })
    # Convert the list of metrics to a DataFrame
    return pd.DataFrame(metrics_list)

def find_optimal_cutoff(expression, time, event, min_percent=0.1, max_percent=0.9, step=0.01):
    """
    Find the optimal cutoff in a continuous variable (e.g., gene expression)
    that best separates survival curves based on log-rank test.

    Parameters:
    - expression: pd.Series (continuous values)
    - time: survival time vector (aligned)
    - event: event indicator vector (aligned)
    - min_percent, max_percent: range of quantiles to search within
    - step: fraction of quantile steps to test

    Returns:
    - best_cutoff: value of expression that best separates survival
    - best_p: log-rank test p-value at that split
    """
    quantiles = np.arange(min_percent, max_percent, step)
    cutoffs = expression.quantile(quantiles).unique()

    best_p = 1
    best_cutoff = None

    for cutoff in cutoffs:
        group = expression > cutoff
        if group.nunique() < 2:
            continue
        results = logrank_test(
            time[group], time[~group],
            event[group], event[~group],
            alpha=0.99
        )
        if results.p_value < best_p:
            best_p = results.p_value
            best_cutoff = cutoff

    return best_cutoff, best_p

def get_device_memory_info(device_str):
    """
    Get memory information for the specified device.

    Args:
        device_str (str): Device string ('cuda', 'mps', 'cpu')

    Returns:
        dict: Memory information dictionary
    """
    if device_str == 'cuda' and torch.cuda.is_available():
        return {
            'allocated': torch.cuda.memory_allocated() / (1024**2),  # MB
            'reserved': torch.cuda.memory_reserved() / (1024**2),    # MB
            'max_allocated': torch.cuda.max_memory_allocated() / (1024**2),  # MB
            'device_name': torch.cuda.get_device_name(0),
            'device_count': torch.cuda.device_count()
        }
    elif device_str == 'mps' and torch.backends.mps.is_available():
        # MPS doesn't have the same detailed memory tracking as CUDA
        return {
            'allocated': 'N/A (MPS)',
            'reserved': 'N/A (MPS)', 
            'max_allocated': 'N/A (MPS)',
            'device_name': 'Apple Metal Performance Shaders',
            'device_count': 1
        }
    else:
        return {
            'allocated': 'N/A (CPU)',
            'reserved': 'N/A (CPU)',
            'max_allocated': 'N/A (CPU)', 
            'device_name': 'CPU',
            'device_count': 1
        }

def get_optimal_clusters(data, min_k=2, max_k=10):
    """
    Find the optimal number of clusters (k) for k-means clustering on the given data,
    based on the silhouette score, and return the cluster labels for the optimal k.

    Parameters:
    - data: pandas DataFrame or numpy array, dataset for clustering.
    - min_k: int, minimum number of clusters to try.
    - max_k: int, maximum number of clusters to try.

    Returns:
    - int, the optimal number of clusters based on the silhouette score.
    - DataFrame, silhouette scores for each k.
    - array, cluster labels for the optimal number of clusters.
    """
    silhouette_scores = []
    cluster_labels_dict = {}  # To store cluster labels for each k

    for k in range(min_k, max_k + 1):
        kmeans = KMeans(n_clusters=k, n_init = 'auto', random_state=42)
        cluster_labels = kmeans.fit_predict(data)
        silhouette_avg = silhouette_score(data, cluster_labels)
        silhouette_scores.append((k, silhouette_avg))
        cluster_labels_dict[k] = cluster_labels  # Store cluster labels
        #print(f"Number of clusters: {k}, Silhouette Score: {silhouette_avg:.4f}")

    # Convert silhouette scores to DataFrame for easier handling and visualization
    silhouette_scores_df = pd.DataFrame(silhouette_scores, columns=['k', 'silhouette_score'])

    # Find the optimal k (number of clusters) with the highest silhouette score
    optimal_k = silhouette_scores_df.loc[silhouette_scores_df['silhouette_score'].idxmax()]['k']

    # Retrieve the cluster labels for the optimal k
    optimal_cluster_labels = cluster_labels_dict[optimal_k]

    return optimal_cluster_labels, optimal_k, silhouette_scores_df

def get_optimal_device(device_preference=None):
    """
    Automatically detect and return the optimal device for PyTorch operations.

    Args:
        device_preference (str, optional): Preferred device type ('cuda', 'mps', 'cpu', 'auto').
                                         If None or 'auto', automatically selects the best available device.

    Returns:
        tuple: (device_str, device_type) where:
            - device_str: String suitable for torch.device() and PyTorch Lightning accelerator
            - device_type: String indicating the device type for compatibility
    """
    if device_preference is None:
        device_preference = 'auto'

    # If specific device is requested, validate and return it
    if device_preference == 'cuda':
        if torch.cuda.is_available():
            return 'cuda', 'gpu'
        else:
            warnings.warn("CUDA requested but not available. Falling back to auto-detection.")
    elif device_preference == 'mps':
        if torch.backends.mps.is_available():
            return 'mps', 'mps'
        else:
            warnings.warn("MPS requested but not available. Falling back to auto-detection.")
    elif device_preference == 'cpu':
        return 'cpu', 'cpu'

    # Auto-detection logic (priority: CUDA > MPS > CPU)
    if torch.cuda.is_available():
        return 'cuda', 'gpu'
    elif torch.backends.mps.is_available():
        return 'mps', 'mps'
    else:
        return 'cpu', 'cpu'

def get_predicted_labels(y_pred_dict, dataset, split, method_name):
    """
    Generate a DataFrame with class probabilities and associated metadata.

    Args:
        y_pred_dict (dict): Dictionary containing predicted probabilities for each variable.
        dataset: Dataset object containing variable types, label mappings, and sample information.
        split (str): Split identifier (e.g., 'train', 'test', or 'val').
        method_name: Name of the method used for prediction task

    Returns:
        pd.DataFrame: A DataFrame containing:
            - sample_id
            - variable
            - class label
            - probability for that class label
            - known label (y_true)
            - predicted label (argmax of the probabilities)
            - train/test split
            - method_name 
    """
    dfs = []

    for var in y_pred_dict.keys():
        if dataset.variable_types[var] == 'categorical':
            # Predicted probabilities
            probabilities = y_pred_dict[var]

            # Convert class indices to labels if mappings exist
            if var in dataset.label_mappings.keys():
                class_labels = [dataset.label_mappings[var][idx] for idx in range(probabilities.shape[1])]
            else:
                class_labels = [f'class_{i}' for i in range(probabilities.shape[1])]

            # Get true labels (y_true)
            y_true = [dataset.label_mappings[var][int(x.item())] if var in dataset.label_mappings.keys() and not np.isnan(x.item()) else np.nan for x in dataset.ann[var]]

            # Predicted labels (argmax of probabilities)
            y_pred_indices = np.argmax(probabilities, axis=1)
            y_pred = [dataset.label_mappings[var][idx] if var in dataset.label_mappings.keys() else idx for idx in y_pred_indices]

            # Create a DataFrame for each sample and its probabilities
            for i, sample_id in enumerate(dataset.samples):
                for j, class_label in enumerate(class_labels):
                    dfs.append({
                        'sample_id': sample_id,
                        'variable': var,
                        'class_label': class_label,
                        'probability': probabilities[i, j],
                        'known_label': y_true[i],
                        'predicted_label': y_pred[i],
                        'split': split,
                        'method': method_name
                    })
        else:
            # For numerical variables, set class_label and probability to NA
            y_true = [x.item() for x in dataset.ann[var]]
            y_pred = [x.item() for x in y_pred_dict[var]]
            for i, sample_id in enumerate(dataset.samples):
                dfs.append({
                    'sample_id': sample_id,
                    'variable': var,
                    'class_label': np.nan,
                    'probability': np.nan,
                    'known_label': y_true[i],
                    'predicted_label': y_pred[i],
                    'split': split,
                    'method': method_name
                })

    # Combine all rows into a DataFrame
    return pd.DataFrame(dfs)

def k_means_clustering(data, k):
    """
    Perform k-means clustering on a given pandas DataFrame.

    Parameters:
    - data: pandas DataFrame, where rows are samples and columns are features.
    - k: int, the number of clusters to form.

    Returns:
    - cluster_labels: A pandas Series indicating the cluster label for each sample.
    - kmeans: The fitted KMeans instance, which can be used to access cluster centers and other attributes.
    """
    # Initialize the KMeans model
    kmeans = KMeans(n_clusters=k, n_init='auto', random_state=42)

    # Fit the model to the data
    kmeans.fit(data)

    # Extract the cluster labels for each data point
    cluster_labels = pd.Series(kmeans.labels_, index=data.index)

    return cluster_labels, kmeans

def louvain_clustering(X, threshold=None, k=None):
    """
    Create a graph from pairwise distances within X. You can define a threshold to connect edges
    or specify k for k-nearest neighbors.

    Parameters:
    - X: numpy array, shape (n_samples, n_features)
    - threshold: float, distance threshold to create an edge between two nodes.
    - k: int, number of nearest neighbors to connect for each node.

    Returns:
    - G: a networkx graph
    """
    distances = euclidean_distances(X)
    G = nx.Graph()
    for i in range(distances.shape[0]):
        for j in range(i + 1, distances.shape[0]):
            # If a threshold is defined, use it to create edges
            if threshold is not None and distances[i, j] < threshold:
                G.add_edge(i, j, weight=1/distances[i, j])
            # If k is defined, add an edge if j is one of i's k-nearest neighbors
            elif k is not None:
                if np.argsort(distances[i])[:k + 1].__contains__(j):
                    G.add_edge(i, j, weight=1/distances[i, j])
    partition = community_louvain.best_partition(G)

    cluster_labels = np.full(len(X), np.nan, dtype=float)
    # Fill the array with the cluster labels from the partition dictionary
    for node_id, cluster_label in partition.items():
        if node_id in range(len(X)):  # Check if the node_id is a valid index in X
            cluster_labels[node_id] = cluster_label
        else:
            # If node_id is not a valid index in X, it's already set to NaN
            continue

    return cluster_labels, G, partition

def optimal_transport_align(embeddings, batch_labels, standardize_by_labels = False):
    """
    Align embeddings from two batches using Optimal Transport, preserving the order of samples.

    Parameters:
    - embeddings (pd.DataFrame): A DataFrame where rows are samples and columns are features.
    - batch_labels (np.ndarray or pd.Series): Batch labels corresponding to the rows of embeddings.

    Returns:
    - aligned_embeddings (pd.DataFrame): A DataFrame containing the aligned embeddings for all samples, with original indices preserved.
    - aligned_batch_labels (pd.Series): A Series containing the corresponding batch labels for the aligned embeddings.
    """
    # Ensure batch labels are a NumPy array
    batch_labels_np = np.array(batch_labels)

    # Identify unique batches
    unique_batches = np.unique(batch_labels_np)
    if len(unique_batches) != 2:
        raise ValueError("Optimal transport supports aligning exactly two batches.")

    # Split embeddings by batch, preserving the original indices
    batch1_indices = np.where(batch_labels_np == unique_batches[0])[0]
    batch2_indices = np.where(batch_labels_np == unique_batches[1])[0]

    batch1_embeddings = embeddings.iloc[batch1_indices].to_numpy()
    batch2_embeddings = embeddings.iloc[batch2_indices].to_numpy()

    # Compute the cost matrix (e.g., Euclidean distances)
    cost_matrix = ot.dist(batch1_embeddings, batch2_embeddings, metric='euclidean')

    # Solve the optimal transport problem
    n_samples_1 = batch1_embeddings.shape[0]
    n_samples_2 = batch2_embeddings.shape[0]
    uniform_dist_1 = np.ones(n_samples_1) / n_samples_1
    uniform_dist_2 = np.ones(n_samples_2) / n_samples_2
    transport_plan = ot.emd(uniform_dist_1, uniform_dist_2, cost_matrix)

    # Align batch 2 embeddings by transporting them to batch 1's distribution
    aligned_batch2 = np.dot(transport_plan.T, batch1_embeddings)

    # Create an array to store the aligned embeddings in the original order
    aligned_embeddings = np.zeros_like(embeddings.to_numpy())
    aligned_embeddings[batch1_indices] = batch1_embeddings
    aligned_embeddings[batch2_indices] = aligned_batch2

    # Standardize the aligned embeddings separately for each batch
    if standardize_by_labels: 
        scaler1 = StandardScaler()
        scaler2 = StandardScaler()

        aligned_embeddings[batch1_indices] = scaler1.fit_transform(aligned_embeddings[batch1_indices])
        aligned_embeddings[batch2_indices] = scaler2.fit_transform(aligned_embeddings[batch2_indices])

    # Convert back to pandas DataFrame and Series, preserving indices
    aligned_embeddings_df = pd.DataFrame(aligned_embeddings, columns=embeddings.columns, index=embeddings.index)
    aligned_batch_labels = pd.Series(batch_labels, index=embeddings.index, name="batch_labels")

    return aligned_embeddings_df, aligned_batch_labels

def plot_dim_reduced(matrix, labels, method='pca', color_type='categorical', title=None):
    """
    Plots the first two dimensions of the transformed input matrix using plotnine,
    with PCA or UMAP, and includes explained variance for PCA.

    Args:
        matrix (np.array): Input data matrix (n_samples, n_features).
        labels (list or array): Labels for coloring.
        method (str): 'pca' or 'umap'.
        color_type (str): 'categorical' or 'numerical'.
        title (str or None): Optional title for the plot.
    """
    method = method.lower()

    # Fit transformation
    if method == 'pca':
        transformer = PCA(n_components=2)
        transformed_matrix = transformer.fit_transform(matrix)
        var_exp = transformer.explained_variance_ratio_ * 100
        xlab = f"PC1 ({var_exp[0]:.1f}%)"
        ylab = f"PC2 ({var_exp[1]:.1f}%)"
        colnames = ['PC1', 'PC2']
    elif method == 'umap':
        transformer = UMAP(n_components=2)
        # Convert to numpy array and handle UMAP compatibility
        import numpy as np
        matrix_np = np.array(matrix, dtype=np.float32)
        try:
            # Try with ensure_all_finite parameter (newer versions)
            transformed_matrix = transformer.fit_transform(matrix_np, ensure_all_finite=True)
        except TypeError:
            # Fallback for older UMAP versions that don't support ensure_all_finite
            transformed_matrix = transformer.fit_transform(matrix_np)
        xlab = "UMAP1"
        ylab = "UMAP2"
        colnames = ['UMAP1', 'UMAP2']
    else:
        raise ValueError("Invalid method. Expected 'pca' or 'umap'.")

    # Create DataFrame
    df = pd.DataFrame(transformed_matrix, columns=colnames)
    df["Label"] = list(labels)

    # Title
    plot_title = title if title else f"{method.upper()} Scatter Plot"

    # Plot
    if color_type == 'categorical':
        df["Label"] = df["Label"].astype(str)
        plot = (
            ggplot(df, aes(x=colnames[0], y=colnames[1], color='Label')) +
            geom_point() +
            labs(title=plot_title, x=xlab, y=ylab, color="Labels") +
            theme_minimal()
        )
    elif color_type == 'numerical':
        df["Label"] = pd.to_numeric(df["Label"], errors='coerce')
        plot = (
            ggplot(df, aes(x=colnames[0], y=colnames[1], color='Label')) +
            geom_point() +
            scale_color_gradient(low="blue", high="red") +
            labs(title=plot_title, x=xlab, y=ylab, color="Label") +
            theme_minimal()
        )
    else:
        raise ValueError("Invalid color_type. Choose 'categorical' or 'numerical'.")

    return plot

def plot_hazard_ratios(cox_model):
    """
    Plots the sorted log hazard ratios using plotnine from a fitted Cox Proportional Hazards model,
    with 95% CI and statistical significance annotations. Displays the C-index in the top-right.
    """
    # Extract summary
    coef_summary = cox_model.summary[['coef', 'coef lower 95%', 'coef upper 95%', 'p']].copy()
    coef_summary.columns = ['coef', 'coef_lower_95', 'coef_upper_95', 'p']
    coef_summary['variable'] = coef_summary.index

    # Sort by p-value
    coef_summary_sorted = coef_summary.sort_values('p').reset_index(drop=True)

    # Add significance stars
    def significance(p):
        if p < 0.0001:
            return '***'
        elif p < 0.001:
            return '**'
        elif p < 0.05:
            return '*'
        elif p < 0.1:
            return '.'
        else:
            return ''

    coef_summary_sorted['stars'] = coef_summary_sorted['p'].apply(significance)

    # Reverse the order for top-to-bottom importance
    coef_summary_sorted['variable'] = pd.Categorical(
        coef_summary_sorted['variable'],
        categories=coef_summary_sorted['variable'][::-1],
        ordered=True
    )
    c_index = cox_model.concordance_index_

    # Plot
    p = (
        ggplot(coef_summary_sorted, aes(x='coef', y='variable'))
        + geom_errorbarh(
            aes(xmin='coef_lower_95', xmax='coef_upper_95'),
            height=0.2, color='skyblue'
        )
        + geom_point(color='skyblue', size=3)
        + geom_text(aes(label='stars'), nudge_y=0.1, size=10)
        + annotate('vline', xintercept=0, linetype='dashed', color='gray')
        + labs(x='Log Hazard Ratio', y='', title=f'Log Hazard Ratios Sorted by P-Value with 95% CI\n Model C-index: {c_index:.2f}')
        + theme_bw()
        + theme(
            axis_text_y=element_text(size=10),
            axis_text_x=element_text(size=10),
            plot_title=element_text(weight='bold'),
        )
    )
    return p    

def plot_kaplan_meier_curves(durations, events, categorical_variable):
    """
    Plots Kaplan-Meier survival curves using plotnine and annotates log-rank test p-values.
    """
    # Prepare DataFrame
    data = pd.DataFrame({
        'Duration': durations,
        'Event': events,
        'Group': categorical_variable
    })

    kmf = KaplanMeierFitter()
    survival_curves = []

    # Fit Kaplan-Meier for each group and collect survival data
    for group in data['Group'].unique():
        group_data = data[data['Group'] == group]
        kmf.fit(group_data['Duration'], group_data['Event'], label=str(group))
        surv_df = kmf.survival_function_.reset_index()
        surv_df.columns = ['Time', 'Survival']
        surv_df['Group'] = str(group)
        survival_curves.append(surv_df)

    # Combine all curves
    plot_data = pd.concat(survival_curves)

    # Compute log-rank p-value
    categories = data['Group'].unique()
    p_text = ""
    if len(categories) == 2:
        group1 = data[data['Group'] == categories[0]]
        group2 = data[data['Group'] == categories[1]]
        result = logrank_test(group1['Duration'], group2['Duration'],
                              event_observed_A=group1['Event'],
                              event_observed_B=group2['Event'])
        p_text = f"Log-rank p = {result.p_value:.2e}"
    elif len(categories) > 2:
        result = multivariate_logrank_test(data['Duration'], data['Group'], data['Event'])
        p_text = f"Multivariate log-rank p = {result.p_value:.2e}"
    else:
        p_text = "Only one group — log-rank test not applicable"

    # Create plot
    p = (
        ggplot(plot_data, aes(x='Time', y='Survival', color='Group'))
        + geom_step()
        + labs(x='Time', y='Survival Probability', color='Group')
        + ggtitle('Kaplan-Meier Survival Curves by Group')
        + annotate("text", x=0.1, y=0.1, label=p_text, size=10, ha='left')
        + theme_minimal()
        + theme(legend_title=element_text(size=10, weight='bold'))
        + scale_color_brewer(type='qual', palette='Set1')
    )

    return p

def plot_label_concordance_heatmap(labels1, labels2, figsize=(12, 10)):
    """
    Plot a heatmap reflecting the concordance between two sets of labels using pandas crosstab.

    Parameters:
    - labels1: The first set of labels.
    - labels2: The second set of labels.
    """
    # Compute the cross-tabulation
    ct = pd.crosstab(pd.Series(labels1, name='Labels Set 1'), pd.Series(labels2, name='Labels Set 2'))
    # Normalize the cross-tabulation matrix column-wise
    ct_normalized = ct.div(ct.sum(axis=1), axis=0)

    # Plot the heatmap
    plt.figure(figsize = figsize)
    sns.heatmap(ct_normalized, annot=True,cmap='viridis', linewidths=.5)# col_cluster=False)
    plt.title('Concordance between label groups')
    plt.show()

def plot_pr_curves(y_true, y_probs):
    """
    Plot Precision-Recall (PR) curves using plotnine for binary or multiclass classification.

    Args:
        y_true (array-like): True class labels.
        y_probs (array-like): Predicted probabilities (n_samples, n_classes).
    """
    y_true = np.array(y_true)
    n_classes = y_probs.shape[1]
    plot_data = []

    if n_classes == 2:
        # Binary classification
        precision, recall, _ = precision_recall_curve(y_true, y_probs[:, 1])
        aupr = average_precision_score(y_true, y_probs[:, 1])
        plot_data.append(pd.DataFrame({
            'recall': recall,
            'precision': precision,
            'label': [f'Class 1 (AUPR = {aupr:.2f})'] * len(recall)
        }))
    else:
        # Multiclass classification (one-vs-rest)
        classes = np.arange(n_classes)
        y_true_bin = label_binarize(y_true, classes=classes)

        for i in classes:
            precision, recall, _ = precision_recall_curve(y_true_bin[:, i], y_probs[:, i])
            aupr = average_precision_score(y_true_bin[:, i], y_probs[:, i])
            df = pd.DataFrame({
                'recall': recall,
                'precision': precision,
                'label': [f'Class {i} (AUPR = {aupr:.2f})'] * len(recall)
            })
            plot_data.append(df)

    # Combine all data
    all_data = pd.concat(plot_data, ignore_index=True)

    # Plot using plotnine
    pr_plot = (
        ggplot(all_data, aes(x='recall', y='precision', color='label')) +
        geom_line(size=1.2) +
        labs(
            title='Precision-Recall Curve',
            x='Recall',
            y='Precision'
        ) +
        theme_minimal()
    )

    return pr_plot

def plot_roc_curves(y_true, y_probs):
    """
    Plot ROC curves using plotnine for binary or multiclass classification.

    Args:
        y_true (array-like): True class labels.
        y_probs (array-like): Predicted probabilities (n_samples, n_classes).
    """
    y_true = np.array(y_true)
    n_classes = y_probs.shape[1]
    plot_data = []

    if n_classes == 2:
        # Binary classification
        fpr, tpr, _ = roc_curve(y_true, y_probs[:, 1])
        auc_score = roc_auc_score(y_true, y_probs[:, 1])
        plot_data.append(pd.DataFrame({'fpr': fpr, 'tpr': tpr, 'label': [f'Class 1 (AUC = {auc_score:.2f})'] * len(fpr)}))
    else:
        # Multiclass classification
        classes = np.arange(n_classes)
        y_true_bin = label_binarize(y_true, classes=classes)

        for i in classes:
            fpr, tpr, _ = roc_curve(y_true_bin[:, i], y_probs[:, i])
            auc_score = roc_auc_score(y_true_bin[:, i], y_probs[:, i])
            df = pd.DataFrame({
                'fpr': fpr,
                'tpr': tpr,
                'label': [f'Class {i} (AUC = {auc_score:.2f})'] * len(fpr)
            })
            plot_data.append(df)

    # Combine all data
    all_data = pd.concat(plot_data, ignore_index=True)

    # Plot using plotnine
    roc_plot = (
        ggplot(all_data, aes(x='fpr', y='tpr', color='label')) +
        geom_line(size=1.2) +
        geom_abline(intercept=0, slope=1, linetype='dashed', color='gray') +
        labs(
            title='ROC Curve',
            x='False Positive Rate',
            y='True Positive Rate'
        ) +
        theme_minimal()
    )

    return roc_plot