lloidparts module

lloidparts.mkLLOIDSnrChisqToTriggers(pipeline, snr, chisq, bank, verbose=False, nxydump_segment=None, logname='')

! Build pipeline fragment that converts single detector SNR and Chisq into triggers.

lloidparts.mkLLOIDSnrSlicesToTimeSliceChisq(pipeline, branch_heads, bank, block_duration)

! Build pipeline fragment that computes the TimeSliceChisq from SnrSlices.

lloidparts.mkLLOIDbranch(pipeline, src, bank, bank_fragment, control_input_tpl, gate_attack_length, gate_hold_length, block_duration, nxydump_segment=None, fir_stride=None, control_peak_time=None, reconstruction_segment_list=None)

! Make a single slice of one branch of the lloid graph, e.g. one instrument and one template bank fragment. For details see: http://arxiv.org/abs/1107.2665

Specifically this implements the filtering of multirate svd basis and (conditional) resampling and reconstruction of the physical SNR

@param pipeline The gstreamer pipeline in which to place this graph @param src The source of data for this graph provided by a gstreamer element @param bank The template bank class @param bank_fragment The specific fragment (time slice) of the template bank in question @param (control_snk, control_src) An optional tuple of the sink and source elements for a graph that will construct a control time series for the gate which aggregates the orthogonal snrs from each template slice. This is used to conditionally reconstruct the physical SNR of interesting times @param gate_attack_length The attack length in samples for the lal_gate element that controls the reconstruction of physical SNRs @param gate_hold_length The hold length in samples for the lal_gate element that controls the reconstruction of physical SNRs @param block_duration The characteristic buffer size that is passed around, which is useful for constructing queues. @param nxydump_segment Not used @param fir_stride The target length of output buffers from lal_firbank in seconds. Directly effects latency. Making this short will force time-domain convolution. Otherwise FFT convolution will be done to save CPU cycles, but at higher latency. @param control_peak_time The window over which to find peaks in the control signal. Shorter windows increase computational cost but probably also detection efficiency. @param reconstruction_segment_list A segment list object that describes when the control signal should be on. This can be useful in e.g., only reconstructing physical SNRS around the time of injections, which can save an enormous amount of CPU time.

lloidparts.mkLLOIDhoftToSnrSlices(pipeline, hoftdict, bank, control_snksrc=(None, None), block_duration=1000000000, verbose=False, logname='', nxydump_segment=None, fir_stride=None, control_peak_time=None, snrslices=None, reconstruction_segment_list=None)

! Build the pipeline fragment that creates the SnrSlices associated with different sample rates from hoft.

@param reconstruction_segment_list A segment list object that describes when the control signal should be on. This can be useful in e.g., only reconstructing physical SNRS around the time of injections, which can save an enormous amount of CPU time.

lloidparts.mkcontrolsnksrc(pipeline, rate, verbose=False, suffix=None, control_peak_samples=None)

! This function implements a portion of a gstreamer graph to provide a control signal for deciding when to reconstruct physical SNRS

@param pipeline A reference to the gstreamer pipeline in which to add this graph @param rate An integer representing the target sample rate of the resulting src @param verbose Make verbose @param suffix Log name for verbosity @param control_peak_samples If nonzero, this would do peakfinding on the control signal with the window specified by this parameter. The peak finding would give a single sample of “on” state at the peak. This will cause far less CPU to be used if you only want to reconstruct SNR around the peak of the control signal.