Merge pull request #3330 from RosettaCommons/JackMaguire/DebugMSRS
David Thieker found a bug where multistage rosetta scripts does not produce any output if too many trajectories fail their filters. The fix is pretty simple: now I update the number of results a stage expects to produce when the previous stage under-performs.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #2760 from RosettaCommons/fordas/pyrosetta_distributed
Basic support for pyrosetta & rosettascripts via python distributed libraries.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3311 from RosettaCommons/jadolfbr/new_per_residue_metrics
Add a few PerResidueRealMetrics and put native pose into the data map as resource
New PerResidueRealMetrics
====================
This PR adds 4 new `PerResidueRealMetrics`, and refactors other SimpleMetrics to use them as their base code:
- `PerResidueEnergyMetric` - calculate per-residue total_energy or any given score type. Uses delta if input is given
- `PerResidueSasaMetric` - calculate per-residue sasa
- `PerResidueRMSDMetric` - calculates per-residue RMSD in same vein as the RMSDMetric
- `PerResidueDensityFitMetric` - calculates the density fit for each residue. This is now used by the `DensityFitResidueSelector`
Add Integration test for all of these. Their core code is unit tested by others as sums.
Fix up code templates, ref_pose functionality has been moved to core. All metrics that were previously tied to loading the native pose have been moved to core to increase their utility.
Note:
====
This also adds any set `-in:file:native` as a resource in the DataMap. This reduces code duplication and allows the native to be loaded once and accessed as a resource in any RS class without global data. This also works with JD3 - as it uses a local `OptionsCollection`, and is MUCH easier to use than the `SavePoseMover`
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3254 from RosettaCommons/vmullig/bluegene_firstime_issue
Attempting fix to Blue Gene/Q first-time execution.
First-time rotamer binary generation fails on the Blue Gene/Q infrastructure. This is another fconfig issue.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3297 from CyrusBiotechnology/indigo/argobody
Code from @indigogo at Cyrus: Antibody grafting, option to define CDRs with an input JSON in exact same format as the preexisting output JSON
(from Indigo:)
Additions to antibody grafting public app. This adds a non-default option to define CDR loop begin/end positions with an input JSON configuration file. If this file is defined on cmdline input, then the RegEx-based CDR prediction is skipped and the CDR begin/end points are defined by the information in the input JSON. This PR adds the following:
a new class to set the CDR sequence position information in AntibodySequence.sequence.cdr.* based on input json numbers. the input json is parsed using the json parsing tools in utility/json_utitlties.hh. JSON format is checked every time and exit with informative message is called if input JSON is misformatted.
control logic in apps/public/antibody.cc to check if -json flag is defined on input, and if it is, skip the RegEx step and do the JSON reading step instead. I am not touching anything else in the antibody.cc code.
The point of all this is to add functionality for connecting any CDR detection tool's output with antibody grafting protocol input.
Steven Lewis 6 years I don't know what to make of the test failures - particularly as the tests were green in the PR pre-merge (http://benchmark.graylab.jhu.edu/revision?id=10139&branch=) and it's irrelevant code anyway. If someone knows what is up, and/or wants to revert, please do so! If we identify that this code is at fault I won't have time to fix it until the weekend.
Steven Lewis 6 years I don't know what to make of the test failures - particularly as the tests were green in the PR pre-merge (http://benchmark.graylab.jhu.edu/revision?id=10139&branch=) and it's irrelevant code anyway. If someone knows what is up, and/or wants to revert, please do so! If we identify that this code is at fault I won't have time to fix it until the weekend.
reposting with "notify list"[list]
Steven Lewis 6 years I don't know what to make of the test failures - particularly as the tests were green in the PR pre-merge (http://benchmark.graylab.jhu.edu/revision?id=10139&branch=) and it's irrelevant code anyway. If someone knows what is up, and/or wants to revert, please do so! If we identify that this code is at fault I won't have time to fix it until the weekend.
reposting with "notify list"[list]
Sergey Lyskov 6 years Steven, this is not new: http://benchmark.graylab.jhu.edu/revision?branch=master&id=60301 - i think this is related to recent changes (http://benchmark.graylab.jhu.edu/revision?branch=master&id=60294 ?) and manifest itself as random failure of D110_DNA_interface T880_Other tests[list]
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3312 from RosettaCommons/sergey/ui
Updating pose_viewer and hal apps
Vikram K. Mulligan 6 years Note that this broke the serialization tests. Sergey, could you take a look at this?[list]
Sergey Lyskov 6 years Thank you for letting me know Vikram, - i missed that. I will look this up (the functionality itself is not broken, it just test give false positive)
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3020 from RosettaCommons/jadolfbr/glycan_symmetry
Fix GlycanRelax compatibility with Symmetry
Make sure ResidueSelector subset and MoveMap are only for the main asymmetric unit in GlycanRelaxMover.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3299 from RosettaCommons/jadolfbr/per_residue_metrics
Add support for Per-Residue SimpleMetrics
This PR adds two new types of `SimpleMetrics` and associated Code Templates:
- `PerResidueRealMetric`
- `PerResidueStringMetric`
These metrics return values for each residue that is masked by a `ResidueSelector` (which is specified in the base classes for a more consistent interface across all PerResidue metrics).
During their apply, they write a value for each residue using Rosetta (default) or PDB identification, as well as the SUM over all of the residues specified.
These have also been incorporated into the `SimpleMetricFilter` and `SimpleMetricFeaturesReporter` systems. For the filter, these can act as a `CompositeMetric`, or a user can filter on the Sum, acting as a normal SimpleMetric. For the Reporter, the data goes into their own per-residue table, with `resNum` as a foreign Key to aid in robust analysis.
Unit tests have been updated to test this new functionality. Existing SimpleMetrics such as the `SASAMetric`, `RMSDMetric`, and `EnergyMetric` will be converted to have their primary code as a `PerResidueRealMetric`.
notify author
notify list [rosetta-logs@googlegroups.com]
Merge pull request #3260 from RosettaCommons/JWLabonte/sugars/sampling
Carbohydrates: Adding RingPlaneFlipMover
This merge will introduce a new pyranose-specific Mover that will flip an appropriate ring 180 degrees about its anomeric bond with a counter move (similar to how shear moves work) at an equatorial bond at the opposite position on the ring. This move can thus only occur for 4-linked aldopyranose or 5-linked ketopyranose residues, but most natural sugars are aldopyranoses, and the 4-linkage is common. It is expected that this Mover will help sample in cases where a oligo- or poly-saccharide is "stuck" with its ring plane in the wrong orientation, with substituents trapped on the wrong side of the ring.
All tests pass.
Vikram K. Mulligan 6 years Note that there's no reason that this merge should have fixed the unit test that was failing. Andy, when you get the chance, could you take a look at the Stepwise unit tests? (See the linux.srlz unit tests from the previous revision.)[list]
notify author
notify list [rosetta-logs@googlegroups.com]