README.md

Feature Store & Analytic Function Tools

Dependencies

Assumes Teradata >=17.20. Requires the fs optional extra (uv sync --extra fs), which installs teradataml.

---

Feature Store tools (fs_*)

• reconnect_to_database - reestablishes a connection to the Teradata database
• fs_setFeatureStoreConfig - sets the feature store config for the session
• fs_getFeatureStoreConfig - gets the feature store config for the session
• fs_isFeatureStorePresent - confirms that a feature store has been created
• fs_featureStoreContent - returns a summary of the feature store
• fs_getDataDomains - lists the available domains in a feature store
• fs_getFeatures - gets the features within the feature store
• fs_getAvailableDatasets - lists the available datasets in the feature store
• fs_getFeatureDataModel - returns the schema of the feature store
• fs_getAvailableEntities - returns the entities in a domain
• fs_createDataset - creates a data set from the feature store

---

teradataml Analytic Function tools (tdml_*)

These tools are registered dynamically at server startup from the teradataml library. The full list of registered functions is maintained in tools/constants.py. Each entry maps a teradataml function name to a one-line summary used as the MCP tool description.

Tools are only registered when a database connection is available and the function exists in the connected system's teradataml version. If a function is unavailable, it is skipped with a warning.

Tool	Description
tdml_ANOVA	Performs one-way Analysis of Variance (ANOVA) on a dataset with two or more groups.
tdml_Apriori	Finds association patterns and calculates statistical metrics to understand the influence of item sets on each other.
tdml_BincodeFit	Computes bin boundaries for numeric columns to be applied by BincodeTransform().
tdml_BincodeTransform	Converts continuous numeric data to categorical data using bin boundaries from BincodeFit() output.
tdml_CFilter	Calculates statistical measures of how likely each pair of items is to be purchased together.
tdml_CategoricalSummary	Displays distinct values and their counts for each specified input column.
tdml_ChiSq	Performs Pearson's chi-squared test for independence between two categorical variables.
tdml_ClassificationEvaluator	Evaluates a classification model by computing confusion matrix metrics such as accuracy, precision, recall, and F1.
tdml_ColumnSummary	Provides a quick overview of column datatypes and a summary of NULL and non-NULL counts for a given table.
tdml_ColumnTransformer	Applies multiple transformations to input data columns in a single operation using Fit analytic function outputs.
tdml_ConvertTo	Converts specified input columns to specified data types.
tdml_DecisionForest	Trains an ensemble decision forest model for classification and regression predictive modeling.
tdml_FTest	Performs an F-test where the test statistic follows an F-distribution under the null hypothesis.
tdml_FillRowId	Adds a column of unique row identifiers to the input table.
tdml_Fit	Determines whether specified numeric transformations can be applied to target columns and outputs parameters for Transform().
tdml_GLM	Trains a Generalized Linear Model (GLM) for regression and classification.
tdml_GLMPerSegment	Trains a separate GLM model for each segment of the input data.
tdml_GetFutileColumns	Returns names of columns that are futile — all values unique, all identical, or distinct ratio exceeds a threshold.
tdml_GetRowsWithMissingValues	Returns rows that contain NULL values in any of the specified input columns.
tdml_GetRowsWithoutMissingValues	Returns rows that have non-NULL values in all of the specified input columns.
tdml_Histogram	Calculates frequency distribution of a dataset using Sturges, Scott, variable-width, or equal-width binning methods.
tdml_KMeans	Groups observations into k clusters where each point belongs to the cluster with the nearest centroid.
tdml_KMeansPredict	Assigns input data points to cluster centroids produced by KMeans().
tdml_KNN	Classifies data points based on proximity to training data points with known categories.
tdml_MovingAverage	Computes moving average values in a series using a specified moving average type.
tdml_NERExtractor	Performs Named Entity Recognition (NER) on input text using dictionary words or regular expression patterns.
tdml_NGramSplitter	Tokenizes an input stream of text and outputs n-grams based on specified delimiter and reset parameters.
tdml_NPath	Scans a set of rows looking for user-specified sequential patterns and returns rows that match.
tdml_NaiveBayesTextClassifierPredict	Predicts text categories using a model generated by NaiveBayesTextClassifierTrainer().
tdml_NaiveBayesTextClassifierTrainer	Calculates conditional probabilities and prior probabilities for token-category pairs for text classification.
tdml_NonLinearCombineFit	Computes parameters for a non-linear combination of existing features for use by NonLinearCombineTransform().
tdml_NonLinearCombineTransform	Generates a new feature by applying a non-linear combination formula using NonLinearCombineFit() output.
tdml_NumApply	Applies a predefined numeric operation to specified input columns.
tdml_OneClassSVM	Trains a linear one-class SVM model to identify outliers or novelty in a dataset.
tdml_OneClassSVMPredict	Predicts whether input data points are outliers using a model generated by OneClassSVM().
tdml_OneHotEncodingFit	Identifies categorical values to be encoded and outputs parameters for OneHotEncodingTransform().
tdml_OneHotEncodingTransform	Encodes categorical columns as one-hot numeric vectors using OneHotEncodingFit() output.
tdml_OrdinalEncodingFit	Identifies distinct categorical values and generates ordinal mappings for use with OrdinalEncodingTransform().
tdml_OrdinalEncodingTransform	Maps categorical values to ordinal integers using OrdinalEncodingFit() output.
tdml_OutlierFilterFit	Calculates percentile bounds and median for target columns for use by OutlierFilterTransform().
tdml_OutlierFilterTransform	Filters rows containing outlier values using bounds from OutlierFilterFit() output.
tdml_Pack	Packs data from multiple input columns into a single column.
tdml_Pivoting	Pivots data from sparse format to dense format (rows to columns).
tdml_PolynomialFeaturesFit	Computes polynomial combination parameters for existing features for use by PolynomialFeaturesTransform().
tdml_PolynomialFeaturesTransform	Generates polynomial feature combinations from existing features using PolynomialFeaturesFit() output.
tdml_QQNorm	Determines whether values in input columns follow a normal distribution.
tdml_ROC	Computes true positive rate, false positive rate, AUC, and Gini coefficient for a binary classification model.
tdml_RandomProjectionFit	Generates a random projection matrix for use by RandomProjectionTransform().
tdml_RandomProjectionMinComponents	Calculates the minimum number of components required for random projection given an epsilon distortion value.
tdml_RandomProjectionTransform	Reduces high-dimensional input data to a lower-dimensional space using RandomProjectionFit() output.
tdml_RegressionEvaluator	Computes metrics to evaluate regression model predictions including RMSE, MAE, and R-squared.
tdml_RoundColumns	Rounds values in specified input columns to a specified number of decimal places.
tdml_RowNormalizeFit	Computes row-wise normalization parameters for specified columns for use by RowNormalizeTransform().
tdml_RowNormalizeTransform	Normalizes input columns row-wise using RowNormalizeFit() output.
tdml_SMOTE	Generates synthetic minority class samples using SMOTE, ADASYN, Borderline-2, or SMOTE-NC algorithms.
tdml_SVM	Trains a linear Support Vector Machine (SVM) for classification and regression.
tdml_SVMPredict	Predicts target values or class labels on new data using a model generated by SVM().
tdml_ScaleFit	Computes scaling statistics for specified columns for use by ScaleTransform().
tdml_ScaleTransform	Scales specified columns using statistics from ScaleFit() output.
tdml_Sessionize	Maps each click event in a user session to a unique session identifier.
tdml_SentimentExtractor	Extracts the sentiment (positive, negative, or neutral) of each input document or sentence.
tdml_Shap	Computes Shapley values to explain individual predictions (feature contributions) for a machine learning model.
tdml_Silhouette	Measures the consistency of cluster assignments by computing silhouette scores for each data point.
tdml_SimpleImputeFit	Computes imputation values (mean, median, or mode) for missing values in the input data.
tdml_SimpleImputeTransform	Substitutes missing values in the input data using imputation values from SimpleImputeFit() output.
tdml_StrApply	Applies a predefined string operation to specified input columns.
tdml_StringSimilarity	Calculates similarity between two strings using Jaro, Jaro-Winkler, N-Gram, or Levenshtein distance.
tdml_TFIDF	Computes Term Frequency (TF), Inverse Document Frequency (IDF), and TF-IDF scores for each term in a document set.
tdml_TDDecisionForestPredict	Predicts target values or class labels using a DecisionForest() model.
tdml_TDGLMPredict	Predicts target values or class labels for test data using a GLM() model.
tdml_TDNaiveBayesPredict	Predicts classification labels using a model generated by NaiveBayes().
tdml_TargetEncodingFit	Computes target encoding values (expected value per category) for categorical columns.
tdml_TargetEncodingTransform	Encodes categorical columns using encoding values from TargetEncodingFit() output.
tdml_TextMorph	Generates morphological variants (morphs) of words in the input dataset.
tdml_TextParser	Tokenizes text, removes punctuation, converts to lowercase, removes stopwords, and applies stemming or lemmatization.
tdml_TrainTestSplit	Splits input data into training and test sets to simulate model performance on new data.
tdml_Transform	Applies numeric transformations to input columns using parameters from Fit() output.
tdml_UnivariateStatistics	Displays descriptive statistics (mean, min, max, stddev, etc.) for each specified numeric input column.
tdml_Unpack	Unpacks data from a single packed column into multiple separate columns.
tdml_Unpivoting	Unpivots data from dense format to sparse format (columns to rows).
tdml_VectorDistance	Computes distances between target vectors and reference vectors.
tdml_WhichMax	Returns all rows that contain the maximum value in a specified input column.
tdml_WhichMin	Returns all rows that contain the minimum value in a specified input column.
tdml_WordEmbeddings	Produces embedding vectors for text and computes similarity between texts.
tdml_XGBoost	Trains an XGBoost (eXtreme Gradient Boosting) model for classification or regression.
tdml_XGBoostPredict	Predicts target values or class labels using a model generated by XGBoost().
tdml_ZTest	Tests the equality of two means under the assumption that the population variances are known.

To add a new analytic function, see How to Add a New Function.

---

Return to Main README