rfi_flagger

ska_sdp_func_python.preprocessing.flagger.rfi_flagger(data: Visibility, alpha=0.5, threshold_magnitude=3.5, threshold_variation=3.5, threshold_broadband=3.5, sampling=8, window=0, window_median_history=10) → Visibility

This is an interface for the dynamic flagger algorithm (FluctuFlagger) based in SKA Processing Functions Library (PFL) to directely use SKA data model as input and output.

Parameters:

  • data – SKA data model visibility

  • alpha – a value between 0 and 1 used in computation of the rate of fluctuations in the visibility magnitudes; the higher the value of alpha, the higher weight for the more recent ‘rate of changes’ in the visibility magnitudes.

  • threshold_magnitude – threshold for the modified z-score (median-based) of visibility magnitudes to be flagged. The recommended value in statistical texts is 3.5.

  • threshold_variation – threshold for the modified z-score of transit scores (see the description of parameter alpha); recommended value is 3.5

  • threshold_broadband – threshold for the modified z-score of a window of medians of the visibility magnitudes in the recent time slots; recommended value is 3.5

  • sampling – intervals at which samples from visibility magnitudes are taken for computing medians. This approximation reduces the cost of computation

  • window – the number of channels to be flagged on each side of a flagged visibility by the statistical approaches above.

  • window_median_history – window of time slots that the medians of the recent time samples are maintained (see the description of parameter threshold_broadband).

Returns:

flagged SKA data model visibility