You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fixes and improvements related to random number generation. Improvements for data set filtering. (#887)
* Avoid modulo bias by using bit shifting
* Fix Makefile; source changed to cpp
* Change function name, results can include max
* Override Number_UpTo for Marsaglia since it natively returns between 0 and 1
* Use Number_UpTo() for interval functions
* Use rn_num_interval instead of rn_gen
* Set RNG to Marsaglia to recover earlier test results.
* Try generating numbers between 0 and 1 using method of Doornik
* Use the shuffle from Random_Number. Set Marsaglia for Kmeans random test.
* Use rn_num_interval for cluster sieve. Set Marsaglia gen for test.
* Try to avoid range and modulo bias
* Change number of bits reported to dieharder based on RNG. Does not
change results.
* use rng_gen() times two PI instead of integer values for angle
* Try a simpler random algorithm.
* Use new simpler algorithm unless back-tracking is specified
* Allow other names besides RunTest.sh (more flexible)
* Add test for new method, update test for old method
* Increase the residue cutoff for the new method.
* Add code docs
* Only set up CheckResidue() stuff if needed (not needed for new routine). Make max_factor_ a double, and make the default 10.
* Get rid of duplicate check
* Only need to convert to degrees for debug
* Change max factor default back to 2. In new random rotate routine, have
max rotations be on a per-dihedral basis. Make rotation counts unsigned
in old routine. Fix printf formatting.
* Hide some debug info. Fix cleanup
* Set default RNG to Marsaglia for tests
* Set marsaglia for test
* Add some new versions of the randomizeions algorithm, one that does no
distance checking, one that only checks for around.
* Enable algorithm 3
* Split around select around function
* Version that respects around and overlap
* Start adding separate DataFilter class
* Start adding init
* Add support for 'filterset' arg
* Start implementing filterindex
* Add set to be filtered stuff
* Add support for filtering 2d sets
* Add support for 3d sets and a check
* Enable version 4
* Rename functions to better describe how they work
* Move regular index selection into its own function
* Make separate placement with no overlap function
* use swapIons_NoOverlap()
* Add function for only respecting overlap
* Move towards using the new algorithms automatically
* Add 'originalalgorithm' keyword.
* Update help
* Add test for new algorithms.
* Fix 0 to 1 RNG for stdlib generator
* Fix code docs
* Add 'gauss' mode to random.
* Update manual with gauss keyword for random command.
* Add function determining the minimum number of elements among all input
data sets.
* Save pointer to master dsl
* Add Finalize() routine.
* Add functions for returning npassed and nfiltered
* Add printout
* Use DataFilter in FilterByData action
* Use DataFilter for exec filter command. Fix infinite loop in DataFilter.
* Remove old code
* Remove old code.
* Clean up keyword help functions
* Fix up help
* Add count sets and countout keyword
* Add 2d matrix filter and count test.
* Hide debug info.
* Move final printout to individual functions; ensure it only happens when
!multi
* Fix up filter action and datafilter command entries
* Add link to Emmett Leddin's cpptraj vim syntax file
* DRR - Go back to Marsaglia for now. New generators still need more
testing.
* Minor version bump for RNG improvements, data filter improvements, randomizeions improvements, and permutedihedrals improvements.
* Fix up and improve randomizeions manual entry.
* Fix up help for permutedihedrals
* maxfactor option was listed twice
* Hail mary to fix github actions...
* Try to fix GitHub Actions by putting update and install in the same
command instance
* Well that was a bust. Try removing the update command...
0 commit comments