Predict the probability of a sample being outlier. Calling xgboost `predict_proba` function. Parameters ---------- X

Fit the detector, predict on samples, and evaluate the model by predefined metrics, e.g., ROC. Parameters ----------

Utility function to generate synthesized data. Normal data is generated by a multivariate Gaussian distribution and outliers are generated b

Utility function for evaluating and printing the results for examples. Default metrics include ROC and Precision @ n Parameters -----

Predict if a particular sample is an outlier or not, allowing the detector to reject (i.e., output = -2) low confidence predic

Fit detector. y is ignored in unsupervised methods. Parameters ---------- X : numpy array of shape (n_samples, n_features)

Predict the outlyingness rank of a sample by a fitted model. The method is for outlier detector score combination. Parameters

This overrides PyOD base class function in order to use the proper `threshold_` which is quite different in the base class. Internal f

Set the number of classes if `y` is presented, which is not expected. It could be useful for multi-class outlier detection. Parameter

Predict if a particular sample is an outlier or not. Calling xgboost `predict` function. Parameters ---------- X : nu

Predict raw anomaly score of X using the fitted detector. For consistency, outliers are assigned with larger anomaly scores. Paramete

Check if an input is within the defined range. Parameters ---------- param : int, float The input parameter to check. low :

Start an investigation session. Profiles the data and returns an InvestigationState. Parameters ---------- X : array

Plan a detection pipeline. Parameters ---------- profile : dict Output of profile_data(). priority : str

Plan detection: select top-N detectors. Wraps ``plan_detection()`` and extracts primary + alternatives into ``state.plans`` (up to 3

Serialize result to JSON string.

Utility function for visualizing the results in examples. Internal use only. Parameters ---------- clf_name : str The name of

Iterate based on feedback. Structured dicts execute immediately. NL strings are parsed with confidence; ambiguous feedback triggers

Explain why specific samples were flagged as anomalies. Parameters ---------- result : dict Output of run_detecti

Profile the input data. Parameters ---------- X : array-like, list, or dict Input data. data_type : str o

Utility function to generate synthesized categorical data. Parameters ---------- n_train : int, (default=1000) The number of

Load a PyOD detector saved by `save()` or by raw joblib.dump. `load()` understands three input shapes: 1. An envelope dict written by `save(

Execute a detection plan. Parameters ---------- X_train : array-like Training data. plan : dict (Detectio

Generate synthetic time series data with injected anomalies. Creates a sinusoidal base signal with Gaussian noise and injects anomalies at

Get activation function by name Parameters ---------- name : str Activation function name. Available functions: 'el

Analyze detection results with quality assessment. Computes per-detector analysis, consensus analysis, quality metrics (separation, a

Invert the order of a list of values. The smallest value becomes the largest in the inverted list. This is useful while combining multiple de

Analyze detection results. Parameters ---------- result : dict Output of run_detection(). X : array-like

Build and return an unfitted detector from a plan. Parameters ---------- plan : dict (DetectionPlan) Output of pl

Generate synthetic attributed graph data with planted anomalies. Normal nodes have features from N(0, 1). Anomaly nodes have features shif

Return a structured KB snapshot for caller-driven detector selection. This is the agent-facing companion to :meth:`plan_detection`.

List available detectors. Parameters ---------- data_type : str or None Filter by data type (e.g. 'tabular', 'tex

Call the quality function via whichever path is available. During PR 1 the helper module does not yet exist; we use the private metho

Fit detector on raw input data. Encodes X into embeddings, applies preprocessing, then fits the inner detector. Parameters

Fit a detector per modality on the input data. Parameters ---------- X : dict of {str: data} Maps modality name t

Suggest what to try next. Parameters ---------- result : dict Output of run_detection(). analysis : dict

Get algorithm metadata by name. Returns None if not found.

Load the model from the specified path. Parameters ---------- path : str The path to load the model. Ret

Import the generator script as a module by file path.

Load an artifact whose sklearn Tree node dtype no longer matches. Mirrors `joblib.load` but plugs a dispatch-table override into joblib's unp

Encode multi-modal input and concatenate. Parameters ---------- X : dict of {str: data} Maps modality name to inp

Utility function to generate synthesized data in clusters. Generated data can involve the low density pattern problem and global outli

One-shot investigation: start → plan → run → analyze. Parameters ---------- X : array-like Input data. da

Conduct Z-normalization on data to turn input samples become zero-mean and unit variance. Parameters ---------- X : numpy array of sh

Validate and reshape time series input. Parameters ---------- X : array-like Time series data. 1D (n_timestamps,) or 2D (n_timest

Save ``detector`` to ``path`` after rewiring its trees through ``_OldDtypeTree``. ``transform`` overrides the per-nodes-array transformation;

Compare detectors. When `names` is provided, returns explanations for those detectors in input order. When `names` is omitte

Diagnostic helper for contamination calibration. Reports the contamination value the run actually used, the actual flagged rate from

Resolve an encoder from various input types. Parameters ---------- encoder : str, BaseEncoder, or callable - If BaseEncoder insta

Write a hand-crafted envelope for tests that need to control dependency-version fields. Mirrors ``save()`` but lets the test inject specific v

Utility function to compute the column wise empirical cumulative distribution of a 2D feature matrix, where the rows are samples and the colu

Hindsight validation of consensus and per-detector results. Computes label-based metrics from `y` against the consensus labels and ea

Raise ValueError if feedback dict is malformed. Validates only structure and required fields. Does not validate semantic content (e.g., wheth

Lazy ADEngine singleton for the tool functions.

Z-normalize a vector. Returns zero vector if std is near zero.

Parse an LLM routing response into ``(detector_choices, justifications)``. Parameters ---------- response : str The raw LLM text.

Fit detector on time series data. Validates the input, creates sliding windows, fits the inner detector on the window matrix, and map

Fit the detector on graph data. Parameters ---------- X : Data or array-like PyG Data object, or node featur

Fit detector. y is assumed to be 0 for all training samples. Parameters ---------- X : numpy array of shape (n_samples, n_feat

Commit a caller-driven detector plan and return a DetectionPlan. Companion to :meth:`get_kb_for_routing`. The caller (LLM agent, rule

Run detection with all planned detectors. Wraps ``run_detection()`` per plan. Computes consensus via rank normalization and majority

Create a HistoryEntry dict.

Create an EmbeddingOD configured for text anomaly detection. Configurations are informed by NLP-ADBench (EMNLP 2025). Parameters

Match feedback against the pattern table; return (proposed, confidence). Parameters ---------- state : InvestigationState feedback :

Convert input to PyG Data object. Accepts: - PyG ``Data`` object (returned as-is) - numpy ``X`` (n_nodes, n_features) + numpy ``edge_inde

Return a deep copy of ``detector`` with each ``tree_`` replaced by an ``_OldDtypeTree`` shim. The shim ages the nodes array via ``transform``

Build a minimal state-like object for the parser.

Write mean and std over frames for each row of a feature matrix.

Ensure plan params expose an explicit contamination value (TA2). The MCP `plan_detection` -> `build_detector` chain serializes the pl

Return the index of top n elements in the list if order is set to 'desc', otherwise return the index of n smallest ones. Parameters -----

Combination method to merge the outlier scores from multiple estimators by taking the average. Parameters ---------- scores : numpy a

Explain a detector. Parameters ---------- name : str Detector short name (e.g. 'ECOD'). Returns

Fit the detector on graph data. Parameters ---------- X : Data or array-like y : ignored edge_index : a

Fit the k-Shape anomaly detector on time series data. Parameters ---------- X : array-like of shape (n_timestamps,) or (n_tim

Generate a summary report. Parameters ---------- result : dict Output of run_detection(). analysis : dict

Get criterion by name Parameters ---------- name : str Loss function name. Available functions: 'mse', 'mae', 'bce'.

Initialize weights for a layer Parameters ---------- layer : torch.nn.Module Layer to be initialized. name : str, optio

Map window-level scores back to per-timestamp scores. Parameters ---------- window_scores : np.ndarray of shape (n_windows,) window_s