Changeset e1fe7e for src/Potentials/CompoundPotential.cpp

Timestamp:

Jun 27, 2014, 9:32:55 PM (11 years ago)

Author:

Frederik Heber <heber@…>

Branches:

Action_Thermostats, Add_AtomRandomPerturbation, Add_FitFragmentPartialChargesAction, Add_RotateAroundBondAction, Add_SelectAtomByNameAction, Added_ParseSaveFragmentResults, AddingActions_SaveParseParticleParameters, Adding_Graph_to_ChangeBondActions, Adding_MD_integration_tests, Adding_ParticleName_to_Atom, Adding_StructOpt_integration_tests, AtomFragments, Automaking_mpqc_open, AutomationFragmentation_failures, Candidate_v1.5.4, Candidate_v1.6.0, Candidate_v1.6.1, ChangeBugEmailaddress, ChangingTestPorts, ChemicalSpaceEvaluator, CombiningParticlePotentialParsing, Combining_Subpackages, Debian_Package_split, Debian_package_split_molecuildergui_only, Disabling_MemDebug, Docu_Python_wait, EmpiricalPotential_contain_HomologyGraph, EmpiricalPotential_contain_HomologyGraph_documentation, Enable_parallel_make_install, Enhance_userguide, Enhanced_StructuralOptimization, Enhanced_StructuralOptimization_continued, Example_ManyWaysToTranslateAtom, Exclude_Hydrogens_annealWithBondGraph, FitPartialCharges_GlobalError, Fix_BoundInBox_CenterInBox_MoleculeActions, Fix_ChargeSampling_PBC, Fix_ChronosMutex, Fix_FitPartialCharges, Fix_FitPotential_needs_atomicnumbers, Fix_ForceAnnealing, Fix_IndependentFragmentGrids, Fix_ParseParticles, Fix_ParseParticles_split_forward_backward_Actions, Fix_PopActions, Fix_QtFragmentList_sorted_selection, Fix_Restrictedkeyset_FragmentMolecule, Fix_StatusMsg, Fix_StepWorldTime_single_argument, Fix_Verbose_Codepatterns, Fix_fitting_potentials, Fixes, ForceAnnealing_goodresults, ForceAnnealing_oldresults, ForceAnnealing_tocheck, ForceAnnealing_with_BondGraph, ForceAnnealing_with_BondGraph_continued, ForceAnnealing_with_BondGraph_continued_betteresults, ForceAnnealing_with_BondGraph_contraction-expansion, FragmentAction_writes_AtomFragments, FragmentMolecule_checks_bonddegrees, GeometryObjects, Gui_Fixes, Gui_displays_atomic_force_velocity, ImplicitCharges, IndependentFragmentGrids, IndependentFragmentGrids_IndividualZeroInstances, IndependentFragmentGrids_IntegrationTest, IndependentFragmentGrids_Sole_NN_Calculation, JobMarket_RobustOnKillsSegFaults, JobMarket_StableWorkerPool, JobMarket_unresolvable_hostname_fix, MoreRobust_FragmentAutomation, ODR_violation_mpqc_open, PartialCharges_OrthogonalSummation, PdbParser_setsAtomName, PythonUI_with_named_parameters, QtGui_reactivate_TimeChanged_changes, Recreated_GuiChecks, Rewrite_FitPartialCharges, RotateToPrincipalAxisSystem_UndoRedo, SaturateAtoms_findBestMatching, SaturateAtoms_singleDegree, StoppableMakroAction, Subpackage_CodePatterns, Subpackage_JobMarket, Subpackage_LinearAlgebra, Subpackage_levmar, Subpackage_mpqc_open, Subpackage_vmg, Switchable_LogView, ThirdParty_MPQC_rebuilt_buildsystem, TrajectoryDependenant_MaxOrder, TremoloParser_IncreasedPrecision, TremoloParser_MultipleTimesteps, TremoloParser_setsAtomName, Ubuntu_1604_changes, stable

Children:

550f2a

Parents:

16227a

git-author:

Frederik Heber <heber@…> (02/27/14 20:15:41)

git-committer:

Frederik Heber <heber@…> (06/27/14 21:32:55)

Message:

FunctionModel now uses list_of_arguments to split sequence of subsets of distances.

• this fixes ambiguities with the set of distances: Imagine the distances within a water molecule as OH (A) and HH (B). We then may have a sequence of argument_t as AABAAB. And with the current implementation of CompoundPotential::splitUpArgumentsByModels() we would always choose the latter (and more complex) model. Hence, we make two calls to TriplePotential_Angle, instead of calls twice to PairPotential_Harmonic for A, one to PairPotential_Harmonic for B, and once to TriplePotential_Angle for AAB.
• now, we new list looks like A,A,B,AAB where each tuple of distances can be uniquely associated with a specific potential.
• changed signatures of EmpiricalPotential::operator(), ::derivative(), ::parameter_derivative(). This involved changing all of the current specific potentials and CompoundPotential.
• TrainingData must discern between the InputVector_t (just all distances) and the FilteredInputVector_t (tuples of subsets of distances).
• FunctionApproximation now has list_of_arguments_t as parameter to evaluate() and evaluate_derivative().
• DOCU: docu change in TrainingData.

File:

• src/Potentials/CompoundPotential.cpp (modified) (10 diffs)

src/Potentials/CompoundPotential.cpp

@@ -228 +228 @@
 {
   arguments_by_model_t partial_args;
-  // go through each model and have it filter out its arguments
+  // go through each model and have it filter out its arguments, this already
+  // returns a list of tuples associated with the specific model
   for(models_t::const_iterator modeliter = models.begin();
       modeliter != models.end(); ++modeliter) {
     FunctionModel::filter_t filterfunction = (*modeliter)->getSpecificFilter();
-    arguments_t tempargs = filterfunction(arguments);
+    list_of_arguments_t tempargs = filterfunction(arguments);
     // then split up all the bunches, too.
-    arguments_t::const_iterator advanceiter = tempargs.begin();
-    for (arguments_t::const_iterator argiter = tempargs.begin();
-        argiter != tempargs.end(); argiter = advanceiter) {
-      advanceiter += (*modeliter)->getSpecificArgumentCount();
+    for (list_of_arguments_t::const_iterator argiter = tempargs.begin();
+        argiter != tempargs.end(); ++argiter) {
+      const arguments_t &args = *argiter;
       partial_args.push_back(
           std::make_pair(
               *modeliter,
-              arguments_t(argiter, advanceiter)
+              args
             )
           );
@@ -251 +251 @@
 
 CompoundPotential::arguments_by_model_t CompoundPotential::splitUpArgumentsByModels(
-    const arguments_t &arguments) const
+    const list_of_arguments_t &listarguments) const
 {
   arguments_by_model_t partial_args;
-  arguments_t::const_iterator argiter = arguments.begin();
   particletypes_per_model_t::const_iterator typesiter = particletypes_per_model.begin();
   models_t::const_iterator modeliter = models.begin();
 
-  // add constant model (which is always first model) with empty args if present
+  /// add constant model (which is always first model) with empty args if present
   if (typesiter->empty()) {
     partial_args.push_back(
@@ -266 +265 @@
     ++typesiter;
   }
+
   // then check other models
-  while (argiter != arguments.end()) {
+  /// we only have to check whether the current model still matches or whether
+  /// have to use the next model.
+  for (list_of_arguments_t::const_iterator argiter = listarguments.begin();
+      argiter != listarguments.end(); ++argiter) {
+    const arguments_t &arguments = *argiter;
     if (typesiter+1 != particletypes_per_model.end()) {
       // check whether next argument bunch is for same model or different one
@@ -275 +279 @@
 
       // we always expect N(N-1)/2 distances for N particle types
-      arguments_t::const_iterator enditer = argiter+(types.size()*(types.size()-1)/2);
-      arguments_t::const_iterator nextenditer = argiter+(nexttypes.size()*(nexttypes.size()-1)/2);
-      arguments_t args(argiter, enditer);
-      arguments_t nextargs(argiter, nextenditer);
-      if (types.size() < nexttypes.size()) {
-        if (areValidArguments(nexttypes, nextargs)) {
+      // check first from sizes alone
+      const size_t tuplesize = types.size()*(types.size()-1)/2;
+      const size_t nexttuplesize = nexttypes.size()*(nexttypes.size()-1)/2;
+      if ((tuplesize != nexttuplesize)) {
+        if ((arguments.size() == tuplesize) &&  areValidArguments(types, arguments)) {
+          // only former still matches, don't increment
+          partial_args.push_back(
+              std::make_pair(*modeliter, arguments)
+              );
+        } else if ((arguments.size() == nexttuplesize) &&  areValidArguments(nexttypes, arguments)) {
           // latter matches, increment
           ++typesiter;
           partial_args.push_back(
-              std::make_pair(*(++modeliter), arguments_t(argiter, nextenditer))
+              std::make_pair(*(++modeliter), arguments)
               );
-          argiter = nextenditer;
-        } else if (areValidArguments(types, args)) {
-          // only former matches, don't increment
-          partial_args.push_back(
-              std::make_pair(*modeliter, arguments_t(argiter, enditer))
-              );
-          argiter = enditer;
-        } else
+        } else {
           ASSERT(0,
-              "CompoundPotential::splitUpArgumentsByModels() - neither type matches its argument bunch.");
-      } else {
-        if (areValidArguments(types, args)) {
-          // only former matches, don't increment
-          partial_args.push_back(
-              std::make_pair(*modeliter, arguments_t(argiter, enditer))
-              );
-          argiter = enditer;
-        } else if (areValidArguments(nexttypes, nextargs)) {
-          // latter matches, increment
-          ++typesiter;
-          partial_args.push_back(
-              std::make_pair(*(++modeliter), arguments_t(argiter, nextenditer))
-              );
-          argiter = nextenditer;
+              "CompoundPotential::splitUpArgumentsByModels() - neither this model nor next model match (size) with current tuple.");
+        }
+      } else { // same size, now we have to check the types individually
+        size_t encodeValidity = 0;
+        encodeValidity += 1*areValidArguments(types, arguments);
+        encodeValidity += 2*areValidArguments(nexttypes, arguments);
+
+        switch (encodeValidity) {
+          case 1:
+            // only former still matches, don't increment
+            partial_args.push_back(
+                std::make_pair(*modeliter, arguments)
+                );
+            break;
+          case 2:
+            ++typesiter;
+            partial_args.push_back(
+                std::make_pair(*(++modeliter), arguments)
+                );
+            break;
+          case 0:
+          case 3:
+          default:
+            ASSERT(0,
+                "CompoundPotential::splitUpArgumentsByModels() - neither this model nor next model match (type) with current tuple.");
+            break;
         }
       }
     } else {
       const SerializablePotential::ParticleTypes_t &types = *typesiter;
-      // we always expect N(N-1)/2 distances for N particle types
-      arguments_t::const_iterator enditer = argiter+(types.size()*(types.size()-1)/2);
-      partial_args.push_back(
-          std::make_pair(*modeliter, arguments_t(argiter, enditer))
-          );
-      argiter = enditer;
+      if (areValidArguments(types, arguments)) {
+        // only former matches, don't increment
+        partial_args.push_back(
+            std::make_pair(*modeliter, arguments)
+            );
+      } else {
+        ASSERT(0,
+            "CompoundPotential::splitUpArgumentsByModels() - last model does not match with current tuple.");
+      }
     }
   }
@@ -326 +341 @@
 }
 
-CompoundPotential::results_t CompoundPotential::operator()(const arguments_t &arguments) const
+CompoundPotential::results_t CompoundPotential::operator()(
+    const list_of_arguments_t &listarguments) const
 {
   /// first, we have to split up the given arguments
   arguments_by_model_t partial_args =
-      splitUpArgumentsByModels(arguments);
+      splitUpArgumentsByModels(listarguments);
   // print split up argument list for debugging
   if (DoLog(4)) {
@@ -348 +364 @@
       iter != partial_args.end(); ++iter) {
     partial_results.push_back(
-        (*iter->first)(iter->second)
+        (*iter->first)(
+            FunctionModel::list_of_arguments_t(1, iter->second))
     );
   }
@@ -366 +383 @@
 }
 
-CompoundPotential::results_t CompoundPotential::parameter_derivative(const arguments_t &arguments, const size_t index) const
+CompoundPotential::results_t CompoundPotential::parameter_derivative(
+    const list_of_arguments_t &listarguments,
+    const size_t index) const
 {
   // first, we have to split up the given arguments
   arguments_by_model_t partial_args =
-      splitUpArgumentsByModels(arguments);
+      splitUpArgumentsByModels(listarguments);
   // then, with each bunch of arguments, we call the specific model
   // get parameter dimensions per model
@@ -399 +418 @@
       const size_t indexbase = (iter == dimensions.begin()) ? 0 : *(iter-1);
       CompoundPotential::results_t results =
-          model->parameter_derivative(argiter->second, index-indexbase);
+          model->parameter_derivative(
+              FunctionModel::list_of_arguments_t(1, argiter->second), index-indexbase);
 
       // either set results or add
@@ -462 +482 @@
   // create initial returnfunction
   FunctionModel::filter_t returnfunction =
-      boost::bind(&Helpers::returnEmptyArguments);
+      boost::bind(&Helpers::returnEmptyListArguments);
 
   // every following fragments combines its arguments with the initial function
@@ -468 +488 @@
       modeliter != models.end(); ++modeliter) {
     returnfunction =
-          boost::bind(&Extractors::concatenateArguments,
+          boost::bind(&Extractors::concatenateListOfArguments,
               boost::bind(returnfunction, _1),
               boost::bind((*modeliter)->getSpecificFilter(), _1)