Imports & setup

!ls

ALGORITHM

Outline:

  1. Take as input bank.xml and a range of parameters + template model + signal model,
  2. Generate test injections for that parameter range using the signal model,
  3. Compute overlaps between each template in bank.xml and each of the test injections,
  4. Maximize the overlap between each injection, and all templates in bank.xml,
  5. Store those maximized overlaps, and templates corresponding to them.

Simplification of algorithm for implementation:

  1. Read in the ini file, and call gwnr_sample_parameter_space to generate the requested injections. Store it as inj.xml.
  2. Split inj.xml into sub-parts of N_injection_per_job test points each. Store them as injection/injection*.xml.
  3. Similarly split bank.xml into sub-parts of N_template_per_job templates each. Store them as bank/bank*.xml.
  4. Compute overlaps between injections injection/injectionA.xml and templates bank/bankB.xml and store them as match-part/matchApartB.dat. This is done using gwnr_banksim.
  5. Use gwnr_banksim_match_combine to read in all overlap data from match-part/matchApart*.dat for injection/injectionA.xml; combine that information by maximizing the match between each injection and all templates! Store the output as match/matchA.dat.
  6. Use gwnr_banksim_collect_results to collect all match values from match/match*.dat into a single file.

Technical detail

  1. What information do we store in match-part/matchApartB.dat?
  2. What information do we store in match/matchA.dat?
  3. How do we uniquely tag injections and templates precisely using the information provided in these data files?

Setup and execute banksim run

Visualize test results