Return an env dict with HOME/USERPROFILE pointing at a clean directory. Used by tests that need to exercise the ``pyod info`` skill-install-state

Normalize a model name to its canonical registry key.

Enforce workflow phase precondition.

Shape-based distance between two z-normalized sequences. SBD(x, y) = 1 - max_w CC_w(x, y) / (||x|| * ||y||) Parameters ---------- x,

Validate and normalize encoder output. Parameters ---------- embeddings : array-like Raw encoder output.

Aggregate per-channel anomaly scores. Z-normalizes each channel before aggregation to prevent high-variance channels from dominating. Pa

Predict raw anomaly scores for X. Parameters ---------- X : list or array-like Raw input data in the same format

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

Fit the detector on graph data.

Fit the Matrix Profile detector on time series data. Parameters ---------- X : array-like of shape (n_timestamps,) or (n_time

Fit the detector on graph data. Parameters ---------- X : Data or array-like PyG Data or node features (n_no

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

Internal method to separate inliers from outliers. Parameters ---------- X : numpy array of shape (n_samples, n_features) Th

Construct a closed-schema DetectionPlan dict. The closed schema means downstream code can rely on the keys ``'detector_name'``, ``'params'``,

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

Utility function to calculate row-wise euclidean distance of two matrix. Different from pair-wise calculation, this function would not broadcast.

Generate investigation report. Text format wraps ``generate_report()`` for best detector, prepending session-level context. JSON form

Save the model to the specified path. Parameters ---------- path : str The path to save the model.

Extract sliding windows from a time series. Parameters ---------- X : np.ndarray of shape (n_timestamps, n_channels) window_size : in

Return sorted list of all *.md files under pyod/skills/.

Load data from file path.

Drive profile_data + plan_detection through the MCP layer.

Average of Maximum - An ensemble method for combining multiple estimators. See :cite:`aggarwal2015theoretical` for details. First dividing es

Checks if fit and decision_function methods exist for given detector Parameters ---------- detector : pyod.models Detector instan

Predict combined anomaly scores for X. Parameters ---------- X : dict of {str: data} Maps modality name to test d

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

Fit detector on time series data. Parameters ---------- X : array-like of shape (n_timestamps,) or (n_timestamps, n_channels)

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

Fit the SAND detector on time series data. Parameters ---------- X : array-like of shape (n_timestamps,) or (n_timestamps, n_

Create an EmbeddingOD configured for audio anomaly detection. Uses a handcrafted 74-dim acoustic feature encoder (20 MFCC, 12 chroma,

Create an EmbeddingOD configured for image anomaly detection. Configurations are informed by AnomalyDINO (WACV 2025). Parameters

Get optimizer by name Parameters ---------- model : torch.nn.Module Model to be optimized. name : str Optimizer

Get parameters for this estimator. See http://scikit-learn.org/stable/modules/generated/sklearn.base.BaseEstimator.html and sklearn/b

Maximization of Average - An ensemble method for combining multiple estimators. See :cite:`aggarwal2015theoretical` for details. First dividi

Utility function to calculate precision @ rank n. Parameters ---------- y : list or numpy array of shape (n_samples,) The ground

Find ROD scores for 3D Data. note that gm, scaler1 and scaler2 will be returned "as they are" and without being changed if the model has

Save a fitted PyOD detector with a versioned envelope. The envelope records every dependency version that can affect pickle/joblib layout, pl

Internal method to add subplot of inliers and outliers. Parameters ---------- X_inliers : numpy array of shape (n_samples, n_

Calculated the variance of weighted cosine of a point. wcos = (<a_curr, b_curr>/((|a_curr|*|b_curr|)^2) Parameters ---------- curr_pt

Create an encoder from a backend name and kwargs.

Load pyod.utils.knowledge.algorithms.

Resolve the running version of an optional dependency.

Render *path* relative to the repo root if possible, else absolute. Used when building error messages. The negative-test fixture points SKILL

Conservative data type detection.

Internal shape to check input data shapes are consistent. Parameters ---------- X_train : numpy array of shape (n_samples, n_features

Create the discriminator of the GAN for a given latent size. Parameters ---------- latent_size : int The size of the latent

Predict raw anomaly scores of X using the fitted detector. The anomaly score of an input sample is computed based on the fitted detec

Predict raw anomaly score of X using the fitted detector. The anomaly score of an input sample is computed based on different detecto

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

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

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

Fit detector on time series data. Parameters ---------- X : array-like of shape (n_timestamps,) or (n_timestamps, n_channels)

Encode training data and store per-modality mean embeddings. Call this during training (EmbeddingOD.fit) so that mean embeddings are

get the elements in li1 but not li2, and vice versa Parameters ---------- li1 : list or numpy array Input list 1. li2 : list

get the overlapping between two lists Parameters ---------- li1 : list or numpy array Input list 1. li2 : list or numpy arra

Function to turn raw outlier scores into binary labels by assign 1 to top n outlier scores. Parameters ---------- y : list or numpy a

get the elements in li1 but not li2. li1-li2 Parameters ---------- li1 : list or numpy array Input list 1. li2 : list or num

Make stochastic detectors repeatable without changing selection.

Predict the model's confidence in making the same prediction under slightly different training sets. See :cite:`perini2020quantifying`

Summarize the current investigation state.

Find ROD overall scores when Data is higher than 3D: # scale dataset using Robust Scaler # decompose the full space into a combinatio

Compute association discrepancy across layers and heads. AssDis = sum over layers of [KL(P||S) + KL(S||P)] / 2 averaged over heads. Para

Compute the perplexity and the A-row for a specific value of the precision of a Gaussian distribution. Parameters ---------- D : arra

Run the k-Shape clustering algorithm. Parameters ---------- subsequences : np.ndarray of shape (n_windows, window_size) Z-normali

Load pyod.utils.knowledge.algorithms once. Planned (not-yet-implemented) detectors are excluded so the skill's detector counts and lists refl

Create a mock OpenAI embeddings response.

Normalize user input to the Python subpackage name (underscore form). Accepts either the Python subpackage name (``od_expert``) or the Cla

Render a markdown bullet list of detectors for a single modality.

Opposite of the anomaly score defined in the original paper. The anomaly score of an input sample is computed as the mean ano

Return a structured array identical to ``nodes_arr`` with one field removed. Used to simulate a pre-1.3 sklearn save.

Decrease contamination by `_CONTAMINATION_DECREASE_FACTOR`, floored at `_CONTAMINATION_MIN`.

Render a routing prompt from a knowledge-base context dict. Parameters ---------- kb_context : dict Output of :meth:`ADEngine.get

Predict clip-level anomaly scores for X. Parameters ---------- X : list Audio clips in the same formats accepted

Predict raw anomaly score of X using the fitted detector. The anomaly score of an input sample is computed based on different detector

One-shot anomaly detection: profile -> plan -> run -> analyze. Parameters ---------- X_train : array-like Trainin

Score (node_embedding, local_context) pairs.