:param score: a score, assumed normalised (between 0 and 1) representing anomalousness :param threshold: a user-specified threshold above wh

:param n: number of total samples, including anomalies, defaults to 1000 :param anomalies: number of anomalies in the sample (should be a sma

:param old: Old data :param new: New data :param w: Controls the size of the rolling window :return: The w most recent datapoints fr

Upload dataframe to Elasticsearch

:param a: one summand (e.g., partial sum) :param b: another summand (e.g., the new datapoint) :param weight: the weight of the update (e

Attempt to detect the time dimension in a dataframe

Initialize anomaly detector models

Evaluate modules as Python code and load selecter anomaly detector

Main function

Parse command line arguments

Restream selected sensors & anomaly detector scores from an input pandas dataframe to an existing Elasticsearch instance and/or to a

:param effective_sample_size: :param batch_size: :param forgetting_factor: :return: >>> update_effective_sample_size(1.0,1.0,1.

send 10 datapoints a second

Replay input stream into Elasticsearch

Fits the model on X and scores each datapoint in X. Parameters ---------- X : ndarray, shape (n_samples, n_features)

Returns a bokeh server configured with the dsio dashboard app

Generate a Kibana dashboard given a list of sensor names

Generates the dashboard document

Restream dataframe to bokeh and/or Elasticsearch

Check the queue for updates sent by the restreamer thread and pass them over to the bokeh data soure