MF Class

This is the MF (match filter) class. Use this class to apply filters to an incoherent noise spectrum.

class SSINS.MF(freq_array, sig_thresh, shape_dict={}, tb_aggro=0, N_samp_thresh=None, narrow=True, streak=True, broadcast_dict={}, broadcast_streak=False)

Defines the Match Filter (MF) class.

apply_match_test(INS, event_record=True, apply_samp_thresh=None, freq_broadcast=False, time_broadcast=False)

A method that uses the match_test() method to flag RFI. The champion from match_test() is flagged and the mean-subtracted spectrum is recalculated. This repeats until there are no more outliers greater than sig_thresh. Also can apply the samp_thresh_test in each iteration, which flags highly occupied channels between match test iterations.

Parameters

• INS – The INS to flag

• event_record (bool) – If True, append events to INS.match_events

• apply_samp_thresh (bool) – Deprecated in favor of the time_broadcast keyword.

• freq_broadcast (bool) – If True, broadcast flags between iterations using the broadcast_dict

• time_broadcast (bool) – If True, broadcasts flags in time if significant flagging in channels. Set tb_aggro parameter for aggression.

broadcast_dict

A dictionary of subbands. Keys are a subband name, values are the lower and upper frequency bounds in Hz.

broadcast_slc_dict

A dictionary whose keys are the same as broadcast_dict and whose values are corresponding slices into the freq_array attribute

freq_array

A 1-d array of frequencies (in hz) for the filter to operate with

freq_broadcast(INS, event, event_record=False)

Broadcast flags in frequency, regardless of flagging fraction. Determined, by the broadcast_dict. An event will be recorded in the match_filter saying which integration/band was flagged.

Parameters

• INS – The incoherent noise spectrum being worked on.

• event – The event to broadcast flags for.

• event_record (bool) – Whether to record a new event for this flagging entry.

match_test(INS)

The primary test that the filter is used with. The greatest outlier in each shape is put forth and a champion among them is chosen. The time, frequencies, and outlier statistic and shape of this champion is returned to the stack.

Parameters

INS – An INS to test

Returns

The time index of the strongest outlier (None if no significant outliers) f_max: The slice in the freq_array for the strongest outlier (None if no significant outliers) R_max: The ratio of the z-score of the outlier to the sig_thresh (-np.inf if no significant outliers) shape_max: The shape of the strongest outlier

Return type

t_max

shape_dict

A dictionary of shapes. Keys are a shape name, values are the lower and upper frequency bounds in Hz.

sig_thresh

A dictionary of significance thresholds to flag per shape. Keys are shapes and values are thresholds.

slice_dict

A dictionary whose keys are the same as shape_dict and whose values are corresponding slices into the freq_array attribute

tb_aggro

The threshold for flagging an entire channel when some flags exist and apply_samp_thresh() is called. Must be between 0 and 1. Represents a fraction of unflagged data remaining.

time_broadcast(INS, event, event_record=False)

Broadcasts flags in time for a subband (determined by the passed event) if the fraction of unflagged samples in the subband is less than the tb_aggro parameter. If so, the entire subband is flagged for the whole object. A ValueError is raised if the aggro parameter is greater than or equal to 1, which will always flag everything.

Parameters

• INS – An INS to test

• event – The event to check.

• event_record (bool) – If true, append events to INS.match_events.

Returns

Possible new event if more flagging happened. Returns old

event if no additional flagging happened.

Return type

new_event

write(prefix, sep='_', clobber=False)

Writes out a yaml file with the important information about the filter.

Parameters

• prefix – The filepath prefix for the output file. Output file will be named f’{prefix}{sep}matchfilter.yaml’

• sep – The separator character between the prefix and the rest of the output filepath.

• clobber – Whether to overwrite an identically named file. True overwrites.