template_bank_generator_HM.TemplateBank.test_effectualness#

Back to Class page

Summary#

Computes the mismatches between original astrophysical waveforms and the bank’s representation of them. Used to measure losses in sensitivity

Signature#

def test_effectualness(self, wfs_fd, delta_calpha, fs_in = None, HM = True, wfs_22_fd = None, n_sinc_interp = 2, highpass = True, truncate = True, coeff_grid = None, do_optimization = True, fudge = params.TEMPLATE_SAFETY, force_zero = True, ret_overlaps_guess = False, search_full_bank = False, calpha_search_radius = 5)

Input variables#
Name	Type	Default	Description
`wfs_fd`			n_wf x fs_in array with LOG of frequency domain waveforms on fs_in, If using 2,2 modes only, ensure that the phases are unwrapped so that we can linearly interpolate If using 2,2 + higher modes, interpolation won’t be accurate so provide them on self.fs_fft
`delta_calpha`			Resolution of calphas used for coeff_grid (the finer grid will have 0.5 \* delta_calpha)
`fs_in`		None	Frequency grid for wfs_fd (None = self.fs_fft)
`HM`		True	Flag indicating whether we are using higher mode template bank or 22 only
`wfs_22_fd`		None	If HM is true, provide a separate 22-only version of the full waveform. These 22 wfs are only used to find the best-fit calphas (ensure that the phases are unwrapped so that we can linearly interpolate) Format: n_wf x fs_in array with LOG of frequency domain waveforms on fs_in
`n_sinc_interp`		2	Number of times to sinc-interpolate the overlaps
`highpass`		True	Flag indicating whether to high-pass filter the waveforms
`truncate`		True
`coeff_grid`		None	n_templates x n_dims array with the coefficients of the templates that are in the bank
`do_optimization`		True
`fudge`		params.TEMPLATE_SAFETY
`force_zero`		True
`ret_overlaps_guess`		False	Flag indicating whether to include the guess in the return (guess = sum_i \\|delta_calpha_i\\2\\|). Only used for debugging.
`search_full_bank`		False	Flag indicating whether to search the full bank. If false (default), we first find the best-fit calpha template and only take the overlap with that. Otherwise, we take overlaps with all templates in bank (more expensive). Internally, we only implement this case when overlap_threshold<0.96. Suggestion: Better to not use this flag with multiprocessing as it consumes a lot of memory.
`calpha_search_radius`		5	This will be updated later by Mark. #TODO_Mark Applicable if search_full_bank is true. We only search for templates within this radius of the best-fit calpha. If None, we search the full bank.

Output variables#

Return annotation	Docstring type	Description
`None`

Docstring#

Computes the mismatches between original astrophysical waveforms and
the bank's representation of them. Used to measure losses in sensitivity
:param wfs_fd:
    n_wf x fs_in array with LOG of frequency domain waveforms on fs_in,
    If using 2,2 modes only, ensure that the phases are unwrapped so
    that we can linearly interpolate
    If using 2,2 + higher modes, interpolation won't be accurate so
    provide them on self.fs_fft
:param fs_in: Frequency grid for wfs_fd (None = self.fs_fft)
:param HM: Flag indicating whether we are using higher mode template
           bank or 22 only
:param wfs_22_fd:
    If HM is true, provide a separate 22-only version of the full waveform.
    These 22 wfs are only used to find the best-fit calphas (ensure that the
     phases are unwrapped so that we can linearly interpolate)
    Format: n_wf x fs_in array with LOG of frequency domain waveforms on fs_in
:param delta_calpha:
    Resolution of calphas used for coeff_grid
    (the finer grid will have 0.5 * delta_calpha)
:param coeff_grid:
    n_templates x n_dims array with the coefficients of the templates
    that are in the bank
:param n_sinc_interp: Number of times to sinc-interpolate the overlaps
:param highpass:
    Flag indicating whether to high-pass filter the waveforms
:param ret_overlaps_guess: Flag indicating whether to include the guess
         in the return (guess = sum_i |delta_calpha_i**2|).
         Only used for debugging.
:param search_full_bank: Flag indicating whether to search the full bank.
    If false (default), we first find the best-fit calpha template and
    only take the overlap with that. Otherwise, we take overlaps with
    all templates in bank (more expensive). Internally, we only
    implement this case when overlap_threshold<0.96.
    Suggestion: Better to not use this flag with multiprocessing
    as it consumes a lot of memory.
:param calpha_search_radius:
    This will be updated later by Mark. #TODO_Mark
    Applicable if search_full_bank is true. We only search for templates
      within this radius of the best-fit calpha. If None, we search the full bank.
:return if n_sinc_interp=0,
        Array of size n_wf with maximum overlaps at current resolution
        else:
        Array of size n_wf x 2
        Maximum overlaps at current resolution
        Maximum sinc interpolated overlaps, without angle information (use this as default)