Skip to content
Snippets Groups Projects

Master new signalmodel review

Merged Master new signalmodel review
master-new-signalmodel-review into master
1 unresolved thread
Merged Leo Tsukada requested to merge master-new-signalmodel-review into master
1 unresolved thread
  • added more comments to point to Aaron's technical notes.
  • This also fixes the negative
    ∑λ2η2\sum\lambda^2\eta^2
    happening to some IMBH templates.
  • Given the new signal model without KDE, in order to avoid nan values returned, the 2D histogram is not globally normalized to the number of samples, n, but pre-normalized in log scale for each SNR column and let normalize() take care of some numerical error to be normalized in the end. We can ignore n as it won't be KDE-ed anyway.

See this review wiki for some discussion about nan issue.

Edited by Leo Tsukada

Merge request reports

Pipeline #517491 passed with warnings

Stage: level0
Stage: level1
Stage: level2
Stage: docker
Stage: test-gstlal
Stage: test-gstlal-full-build
Stage: test-gstlal-ugly
Stage: test-burst
Stage: test-inspiral
Stage: test-offline
Stage: docs

Pipeline passed with warnings for a6545a06 on master-new-signalmodel-review

Approval is optional

Merged by Leo TsukadaLeo Tsukada 2 years ago (Apr 20, 2023 1:50am UTC)

Merge details

  • Changes merged into master with a6545a06.
  • Deleted the source branch.

Pipeline #517509 canceled

Stage: level0
Stage: level1
Stage: level2
Stage: docker
Stage: test-gstlal
Stage: test-gstlal-full-build
Stage: test-gstlal-ugly
Stage: test-burst
Stage: test-inspiral
Stage: test-offline
Stage: docs

Pipeline canceled for a6545a06 on master

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
gstlal-inspiral/python/stats/inspiral_extrinsics.py
947 947 df = (lambda_sum + lambda_etasq_sum_mismatch)**2 / (lambdasq_sum + 2 * lambdasq_etasq_sum_mismatch)
948 948 bias_correction = (lambda_sum + lambda_etasq_sum_mismatch) / df / norm / snr2
949 949 chi2_grid = numpy.outer(1./bias_correction, rcoss)
950 # arr[snrindices, rcossindices] += gstlalstats.ncx2pdf(chi2_grid, df[:, numpy.newaxis], 1e-20) / bias_correction[:, numpy.newaxis]
951 arr[snrindices, rcossindices] += stats.chi2.pdf(chi2_grid, df[:, numpy.newaxis]) / bias_correction[:, numpy.newaxis]
950 logarr = stats.chi2.logpdf(chi2_grid, df[:, numpy.newaxis]) - numpy.log(bias_correction[:, numpy.newaxis])
951 arr[snrindices, rcossindices] = numpy.exp(numpy.logaddexp(numpy.log(arr[snrindices, rcossindices]), logarr))
  • Leo Tsukada added 1 commit

    added 1 commit

    • f65d8eda - stats/inspiral_extrinsics.py: try to keep log scale as much as possible and...

    Compare with previous version

  • Leo Tsukada added 1 commit

    added 1 commit

    • befac472 - stats/inspiral_extrinsics.py : first attempt to introduce logarr attribute in...

    Compare with previous version

  • Leo Tsukada changed the description

    changed the description

  • Leo Tsukada changed the description

    changed the description

  • Leo Tsukada added 1 commit

    added 1 commit

    • e80c5465 - Revert "stats/inspiral_extrinsics.py : remove boundary of the signal

    Compare with previous version

  • Leo Tsukada added 1 commit

    added 1 commit

    • f0451942 - move the fast cut to NegInf for nan case from __call__ to internal __getitem__ method

    Compare with previous version

  • Leo Tsukada marked the checklist item After talking with Kipp, we decided that it would be straightforward to put here the existing fast cut to -inf. as completed

    marked the checklist item After talking with Kipp, we decided that it would be straightforward to put here the existing fast cut to -inf. as completed

  • Leo Tsukada marked the checklist item After talking with Kipp, we decided that it would be straightforward to put here the existing fast cut to -inf. as incomplete

    marked the checklist item After talking with Kipp, we decided that it would be straightforward to put here the existing fast cut to -inf. as incomplete

  • Leo Tsukada added 1 commit

    added 1 commit

    • 71ea1f93 - inspiral_extrinsics.py inspiral_lr.py : make consistent boundary for sampling...

    Compare with previous version

  • Leo Tsukada added 1 commit

    added 1 commit

    Compare with previous version

  • Leo Tsukada added 4 commits

    added 4 commits

    • 47d9438a - stats/inspiral_lr.py : adjust snr_max and overwrite finish method for the numerator class
    • 7452c0f8 - add print statement for debugging
    • c0e68251 - inspiral_extrinsics.py : introduce logarr when constructing signal model
    • e6b4af4b - inspiral_lr.py : change snr_max for signal model not to give nan

    Compare with previous version

    Toggle commit list
  • Leo Tsukada added 1 commit

    added 1 commit

    • 190dafde - stats/inspiral_lr.py : minor fix

    Compare with previous version

  • Leo Tsukada added 22 commits

    added 22 commits

    • 190dafde...47bb5bf6 - 4 commits from branch master
    • e63c304c - gstlal-inspiral/python/stats/inspiral_extrinsics.py : add some comments for reference
    • 58054722 - svd_bank.py : fix the sliced matrix being off by one element
    • bf35e16f - svd_bank.py : add comments
    • 70205c31 - gstlal-inspiral/python/dags/layers/inspiral.py : make config.prior.df...
    • 758a45d8 - gstlal-inspiral/python/stats/inspiral_lr.py : introduce skip_kde attribute
    • e28f1f34 - minor fix
    • e7f057e0 - stats/inspiral_extrinsics.py : remove the boundary and introduce logsumexp
    • 3a6597fd - stats/inspiral_extrinsics.py: try to keep log scale as much as possible and...
    • 003454bb - stats/inspiral_extrinsics.py : first attempt to introduce logarr attribute in...
    • eff58c0b - Revert "stats/inspiral_extrinsics.py : remove boundary of the signal
    • 9c75b1d7 - move the fast cut to NegInf for nan case from __call__ to internal __getitem__ method
    • 33dc35c0 - inspiral_extrinsics.py inspiral_lr.py : make consistent boundary for sampling...
    • 61ff9b1e - minor fix
    • 55ae07d6 - stats/inspiral_lr.py : adjust snr_max and overwrite finish method for the numerator class
    • 5932dd27 - add print statement for debugging
    • 7782746f - inspiral_extrinsics.py : introduce logarr when constructing signal model
    • a38fe117 - inspiral_lr.py : change snr_max for signal model not to give nan
    • 6aafe980 - stats/inspiral_lr.py : minor fix

    Compare with previous version

    Toggle commit list
  • Leo Tsukada changed the description

    changed the description

  • Leo Tsukada added 29 commits

    added 29 commits

    • 6aafe980...7fcc2b6f - 3 commits from branch master
    • 383f17bc - gstlal-inspiral/python/stats/inspiral_extrinsics.py : add some comments for reference
    • 013465c8 - svd_bank.py : fix the sliced matrix being off by one element
    • 615afda1 - svd_bank.py : add comments
    • c941b132 - gstlal-inspiral/python/dags/layers/inspiral.py : make config.prior.df...
    • 7cdf31fe - gstlal-inspiral/python/stats/inspiral_lr.py : introduce skip_kde attribute
    • b174188c - minor fix
    • f57f4c68 - stats/inspiral_extrinsics.py : remove the boundary and introduce logsumexp
    • 575d0d2f - stats/inspiral_extrinsics.py: try to keep log scale as much as possible and...
    • 0527ed7d - stats/inspiral_extrinsics.py : first attempt to introduce logarr attribute in...
    • 7b27d562 - Revert "stats/inspiral_extrinsics.py : remove boundary of the signal
    • 4bde167a - move the fast cut to NegInf for nan case from __call__ to internal __getitem__ method
    • 124519e5 - inspiral_extrinsics.py inspiral_lr.py : make consistent boundary for sampling...
    • e74f6c87 - minor fix
    • 29a27d6d - stats/inspiral_lr.py : adjust snr_max and overwrite finish method for the numerator class
    • f93c1f97 - add print statement for debugging
    • bae607c3 - inspiral_extrinsics.py : introduce logarr when constructing signal model
    • f31bf5b8 - inspiral_lr.py : change snr_max for signal model not to give nan
    • 49ac4510 - stats/inspiral_lr.py : minor fix
    • 6c428003 - introduce pre-normalization in log-scale per SNR column for signal model and a...
    • 291592cc - far.py : minor fix in fast_path_cut
    • 21b239b3 - blah
    • 460b4a2f - stats/inspiral_lr.py : lower snr_max and make snr_max for signal model...
    • f5edaf20 - gstlal-inspiral/bin/gstlal_inspiral_create_prior_diststats : set upper limit...
    • 358ab285 - remove the argument of the number of samples for the new signal model
    • acdfab08 - remove the argument of the number of samples for the new signal model
    • 0a8e86ce - far.py : fix typo

    Compare with previous version

    Toggle commit list
  • Leo Tsukada added 1 commit

    added 1 commit

    • 709fa897 - inspiral_extrinsics.py inspiral_lr.py : fix some issues showing up after reabse

    Compare with previous version

  • Leo Tsukada added 1 commit

    added 1 commit

    Compare with previous version

  • Leo Tsukada changed the description

    changed the description

  • Leo Tsukada added 1 commit

    added 1 commit

    • 50dd0507 - python/far.py: remove debugging print statement

    Compare with previous version

  • Leo Tsukada added 30 commits

    added 30 commits

    • 1e8845d4 - 1 commit from branch master
    • e01bbb40 - gstlal-inspiral/python/stats/inspiral_extrinsics.py : add some comments for reference
    • d3ae8fe8 - svd_bank.py : fix the sliced matrix being off by one element
    • b81774f4 - svd_bank.py : add comments
    • f5e3ff09 - gstlal-inspiral/python/dags/layers/inspiral.py : make config.prior.df...
    • ee857472 - gstlal-inspiral/python/stats/inspiral_lr.py : introduce skip_kde attribute
    • d27ce196 - minor fix
    • 4135b9a6 - stats/inspiral_extrinsics.py : remove the boundary and introduce logsumexp
    • c8abf22a - stats/inspiral_extrinsics.py: try to keep log scale as much as possible and...
    • 07bb20dc - stats/inspiral_extrinsics.py : first attempt to introduce logarr attribute in...
    • 277a08a4 - Revert "stats/inspiral_extrinsics.py : remove boundary of the signal
    • 37252531 - move the fast cut to NegInf for nan case from __call__ to internal __getitem__ method
    • 681bd3d8 - inspiral_extrinsics.py inspiral_lr.py : make consistent boundary for sampling...
    • df5553f7 - minor fix
    • df8b0641 - stats/inspiral_lr.py : adjust snr_max and overwrite finish method for the numerator class
    • f0923527 - add print statement for debugging
    • 0bc1b383 - inspiral_extrinsics.py : introduce logarr when constructing signal model
    • d22df8cf - inspiral_lr.py : change snr_max for signal model not to give nan
    • 2b07fb80 - stats/inspiral_lr.py : minor fix
    • 3ddc4ffc - introduce pre-normalization in log-scale per SNR column for signal model and a...
    • 2d32f57f - far.py : minor fix in fast_path_cut
    • e9b21ff6 - blah
    • 8d5d3372 - stats/inspiral_lr.py : lower snr_max and make snr_max for signal model...
    • 1c3a2e4f - gstlal-inspiral/bin/gstlal_inspiral_create_prior_diststats : set upper limit...
    • 099326ca - remove the argument of the number of samples for the new signal model
    • 9c79d742 - remove the argument of the number of samples for the new signal model
    • a2ecdf30 - far.py : fix typo
    • 3399216d - inspiral_extrinsics.py inspiral_lr.py : fix some issues showing up after reabse
    • 313aab10 - remove trailed white line
    • a0c68388 - python/far.py: remove debugging print statement

    Compare with previous version

    Toggle commit list
  • Leo Tsukada added 3 commits

    added 3 commits

    • b4023ff4 - fix the command line docs
    • dd90cea1 - python/stats/inspiral_lr.py : add extra one bin to the snr slice to make sure...
    • 4135920c - python/stats/inspiral_extrinsic.py : add extra one bin to the snr slice to...

    Compare with previous version

  • Leo Tsukada added 1 commit

    added 1 commit

    • e2b41938 - gstlal-inspiral/python/stats/inspiral_lr.py : propagate skip_kde attribute to...

    Compare with previous version

  • Leo Tsukada added 48 commits

    added 48 commits

    • e2b41938...8d5670b7 - 15 commits from branch master
    • 8f123848 - gstlal-inspiral/python/stats/inspiral_extrinsics.py : add some comments for reference
    • 451172b5 - svd_bank.py : fix the sliced matrix being off by one element
    • 892c7133 - svd_bank.py : add comments
    • 95a528e0 - gstlal-inspiral/python/dags/layers/inspiral.py : make config.prior.df...
    • f247a3c6 - gstlal-inspiral/python/stats/inspiral_lr.py : introduce skip_kde attribute
    • 184a76ad - minor fix
    • aa503ca6 - stats/inspiral_extrinsics.py : remove the boundary and introduce logsumexp
    • 4f3500b8 - stats/inspiral_extrinsics.py: try to keep log scale as much as possible and...
    • d479fa3a - stats/inspiral_extrinsics.py : first attempt to introduce logarr attribute in...
    • b061ab63 - Revert "stats/inspiral_extrinsics.py : remove boundary of the signal
    • 27a1eca8 - move the fast cut to NegInf for nan case from __call__ to internal __getitem__ method
    • cd011979 - inspiral_extrinsics.py inspiral_lr.py : make consistent boundary for sampling...
    • 02253893 - minor fix
    • 0f23e5fe - stats/inspiral_lr.py : adjust snr_max and overwrite finish method for the numerator class
    • b0d18f7c - add print statement for debugging
    • ed1b538e - inspiral_extrinsics.py : introduce logarr when constructing signal model
    • a76a61e8 - inspiral_lr.py : change snr_max for signal model not to give nan
    • 7faded44 - stats/inspiral_lr.py : minor fix
    • a7048c0e - introduce pre-normalization in log-scale per SNR column for signal model and a...
    • 470916ba - far.py : minor fix in fast_path_cut
    • 8b346163 - blah
    • 74f0c658 - stats/inspiral_lr.py : lower snr_max and make snr_max for signal model...
    • aa0e6a0b - gstlal-inspiral/bin/gstlal_inspiral_create_prior_diststats : set upper limit...
    • de552e1a - remove the argument of the number of samples for the new signal model
    • fd6536d9 - remove the argument of the number of samples for the new signal model
    • 76fb6ffe - far.py : fix typo
    • 26547488 - inspiral_extrinsics.py inspiral_lr.py : fix some issues showing up after reabse
    • 475b735d - remove trailed white line
    • b0030723 - python/far.py: remove debugging print statement
    • 5b6694ec - fix the command line docs
    • 6d5467cc - python/stats/inspiral_lr.py : add extra one bin to the snr slice to make sure...
    • 921d078d - python/stats/inspiral_extrinsic.py : add extra one bin to the snr slice to...
    • 9166b97d - gstlal-inspiral/python/stats/inspiral_lr.py : propagate skip_kde attribute to...

    Compare with previous version

    Toggle commit list
  • Leo Tsukada added 34 commits

    added 34 commits

    • a0afba41 - 1 commit from branch master
    • 798e4ad0 - gstlal-inspiral/python/stats/inspiral_extrinsics.py : add some comments for reference
    • ff669aa5 - svd_bank.py : fix the sliced matrix being off by one element
    • 2e53af2f - svd_bank.py : add comments
    • e5dd62e9 - gstlal-inspiral/python/dags/layers/inspiral.py : make config.prior.df...
    • 0c639fb5 - gstlal-inspiral/python/stats/inspiral_lr.py : introduce skip_kde attribute
    • 98027159 - minor fix
    • 3b424467 - stats/inspiral_extrinsics.py : remove the boundary and introduce logsumexp
    • 535a21be - stats/inspiral_extrinsics.py: try to keep log scale as much as possible and...
    • 6611c6cc - stats/inspiral_extrinsics.py : first attempt to introduce logarr attribute in...
    • 643baa70 - Revert "stats/inspiral_extrinsics.py : remove boundary of the signal
    • ddd248ff - move the fast cut to NegInf for nan case from __call__ to internal __getitem__ method
    • 3e60b4db - inspiral_extrinsics.py inspiral_lr.py : make consistent boundary for sampling...
    • 18bc3283 - minor fix
    • 028cc8db - stats/inspiral_lr.py : adjust snr_max and overwrite finish method for the numerator class
    • 652cb9cc - add print statement for debugging
    • 62cf8717 - inspiral_extrinsics.py : introduce logarr when constructing signal model
    • bc3b5a24 - inspiral_lr.py : change snr_max for signal model not to give nan
    • 1073ddac - stats/inspiral_lr.py : minor fix
    • 7b4d1b60 - introduce pre-normalization in log-scale per SNR column for signal model and a...
    • 96e64cd8 - far.py : minor fix in fast_path_cut
    • a6a766af - blah
    • c1bef301 - stats/inspiral_lr.py : lower snr_max and make snr_max for signal model...
    • 248eab00 - gstlal-inspiral/bin/gstlal_inspiral_create_prior_diststats : set upper limit...
    • 81e14166 - remove the argument of the number of samples for the new signal model
    • cbdc6985 - remove the argument of the number of samples for the new signal model
    • 151c6063 - far.py : fix typo
    • 60f1df3a - inspiral_extrinsics.py inspiral_lr.py : fix some issues showing up after reabse
    • fa62d08f - remove trailed white line
    • 7198811d - python/far.py: remove debugging print statement
    • bfa04398 - fix the command line docs
    • c43d12f1 - python/stats/inspiral_lr.py : add extra one bin to the snr slice to make sure...
    • a5fbac29 - python/stats/inspiral_extrinsic.py : add extra one bin to the snr slice to...
    • 6775e428 - gstlal-inspiral/python/stats/inspiral_lr.py : propagate skip_kde attribute to...

    Compare with previous version

    Toggle commit list
  • Leo Tsukada added 5 commits

    added 5 commits

    • 17aa795f - inspiral_extrinsics.py, svd_bank.py : fix the bug caused by one sample
    • d17dcff8 - gstlal-inspiral/python/dags/layers/inspiral.py : make config.prior.df...
    • 983725db - gstlal-inspiral/python/stats/inspiral_lr.py : introduce skip_kde attribute so...
    • 0472e97a - introduce pre-normalization in log-scale per SNR column for signal model and a...
    • 561742c7 - gstlal-inspiral/bin/gstlal_inspiral_create_prior_diststats : set upper limit...

    Compare with previous version

    Toggle commit list
  • Leo Tsukada added 6 commits

    added 6 commits

    • 14e5b093 - 1 commit from branch master
    • fd88d52a - inspiral_extrinsics.py, svd_bank.py : fix the bug caused by one sample
    • d395325a - gstlal-inspiral/python/dags/layers/inspiral.py : make config.prior.df...
    • 5f523822 - gstlal-inspiral/python/stats/inspiral_lr.py : introduce skip_kde attribute so...
    • d34e5866 - introduce pre-normalization in log-scale per SNR column for signal model and a...
    • a6545a06 - gstlal-inspiral/bin/gstlal_inspiral_create_prior_diststats : set upper limit...

    Compare with previous version

    Toggle commit list
  • Leo Tsukada merged

    merged

    • Please register or sign in to reply