Merge pull request #613 from RosettaCommons/roccomoretti/rand_perfbench
Randomize the order in which the benchmark tests are run.
This will hopefully address any sort of issues where similar tests are correlated due to test grouping.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #611 from RosettaCommons/vmullig/fix_hbond_counter
Trying to fix several problems with the HbondsToResidue filter.
Tasks:
- Allowing apply-time parse of residue number (with support for reference poses).
- Adding proper setters and getters for internal data.
- Adding proper constructors, with full initialization of all data.
- Adding explicit option to consider bonds to other chains, and another to consider bonds in the current chain. Default behaviour is to consider ALL hydrogen bonds to selected residue, now, but the user can turn off out-of-chain or in-chain hydrogen bonds. (Previously, the filter didn't work with single-chain poses, or work properly with poses with more than two chains.)
- Update documentation.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #614 from RosettaCommons/vmullig/update_bettercluster
Tweaking bettercluster.cc -- updating for new owning pointers.
Trivial change to pilot app. Merging without testing.
notify author
notify list [rosetta-logs@googlegroups.com]
Merging fixes to >1000 bugs across >600 files discovered using the static analysis tool cppcheck and fixed manually.
The most common:
1. Postincremented iterators. Postincrementing an iterator requires copying the iterator. It can be slow. Unless you use the original value, this is simply not correct; preincrementing is superior.
2. Variables declared and/or initialized but not used.
3. Variables initialized and then re-initialized without using the original value
4. Variables given too great a scope
5. Mis-used standard library functions (most commonly, the return value of unique was not used, so all the redundant values were still in the container)
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #610 from RosettaCommons/vmullig/fix_energy_per_residue_filter
Small fix to EnergyPerResidueFilter.cc.
The filter previously assumed that all poses have PDB info, which is not necessarily true for poses that Rosetta generates itself. This adds a check for a PDBinfo object before trying to access that object.
This is a very minor bugfix, so I'm merging without putting it through the test server to save test server clock cycles.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #608 from RosettaCommons/dougrenfrew/fix_app_exception_handling_test
Fixing app_exception_handling itest
Apps need a try/catch block in their main function in order to catch exceptions thrown by things like the options system.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #601 from RosettaCommons/dougrenfrew/fix_ncaa_fixbb
This updates the two-component reference energy energy methods, params files, unit tests, integration tests, and score function finger print tests. Changes to the ncaa_fixbb itest are expected (it should stop complaining) and changes to the mm_std_fa_elec_dslf_fa13_split_unfolded score test are expected as well as a result of the modified params.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #592 from RosettaCommons/rhiju/placeholder_residue_type_for_reduced_footprint
This pull request introduces 'placeholder' ResidueTypes. It supports a more general refactoring of ResidueTypeSet that only instantiates full-fledged ResidueTypes when they are needed, for speed and reduced memory footprint. The more general effort is happening at https://github.com/RosettaCommons/main/pull/591 and the biggest speed increase will happen with changes in how ResidueTypeSet is accessed through the code (particularly in pose_from_pdb and pose_from_sequence).
This commit set already produces a modest speed increase in applications like extract_pdbs which don't need to instantiate all residue types with all patches to do their jobs.
All integration test changes are OK -- a few are due to change in ordering of when ResidueTypes are finalized, and their number.
Note that the option of creating these placeholders is on ("-chemical:on_the_fly" is default true), since all tests are OK at the moment, and I want to ensure that others find any issues with ASAP.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #599 from RosettaCommons/mlnance-N-patch
Carbohydrates: corrected the hydrogens in the N-linked patch file
Carbohydrate test changes were expected, other test changes were *hopefully* not related to this change
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #565 from RosettaCommons/vmullig/multirepeat_bundle
Permit helical bundles to have repeating units that are more than one residue.
Currently, our parametric design code assumes that the repeating unit in a helix is a single residue. This changes that to permit the repeating unit to be more than one residue. This allows interesting secondary structures like alpha-sheets (made from alternating D- and L- amino acids in the left- and right-handed helix regions of Ramachandran space, respectively) to be generated parametrically. It also allows things like the collagen triple helix to be created parametrically.
Tasks:
-- Update the write function for Crick params files.
-- Update the minor helix fitting app.
-- Update the minimizer called by the minor helix fitting app.
-- Add an integration test for fitting a helix with a multi-residue repeating unit.
-- Update and test the MakeBundle mover.
-- Update the read function for Crick params files.
-- Update and test the MakeBundleHelix mover called by the MakeBundle mover.
-- Add an integration test for the MakeBundle mover + more than one residue in the repeating unit.
-- Update and test the PerturbBundle mover.
-- Add an integration test for the PerturbBundle mover + more than one residue in the repeating unit.
-- Update and test the BundleGridSampler mover.
-- Add an integration test for the BundleGridSampler mover + more than one residue in the repeating unit.
-- UPDATE DOCUMENTATION.
Also:
- Adding a symmetrically split TBMB to the database, for symmetric linker design.
Integration test changes observed for all parametric generation/design integration tests; these are all cosmetic and expected. Slight numerical changes to the fitter are possible since the fit algorithm has been tweaked. Two new integration tests have been added, and will fail because they're new.
notify author
notify list [rosetta-logs@googlegroups.com]