source: src/molecules.cpp@ cc2ee5

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
Last change on this file since cc2ee5 was cc2ee5, checked in by Frederik Heber <heber@…>, 16 years ago

Just a temporary commit

  • Property mode set to 100755
File size: 212.3 KB
Line 
1/** \file molecules.cpp
2 *
3 * Functions for the class molecule.
4 *
5 */
6
7#include "molecules.hpp"
8
9/************************************* Other Functions *************************************/
10
11/** Determines sum of squared distances of \a X to all \a **vectors.
12 * \param *x reference vector
13 * \param *params
14 * \return sum of square distances
15 */
16double LSQ (const gsl_vector * x, void * params)
17{
18 double sum = 0.;
19 struct LSQ_params *par = (struct LSQ_params *)params;
20 Vector **vectors = par->vectors;
21 int num = par->num;
22
23 for (int i=num;i--;) {
24 for(int j=NDIM;j--;)
25 sum += (gsl_vector_get(x,j) - (vectors[i])->x[j])*(gsl_vector_get(x,j) - (vectors[i])->x[j]);
26 }
27
28 return sum;
29};
30
31/************************************* Functions for class molecule *********************************/
32
33/** Constructor of class molecule.
34 * Initialises molecule list with correctly referenced start and end, and sets molecule::last_atom to zero.
35 */
36molecule::molecule(periodentafel *teil)
37{
38 // init atom chain list
39 start = new atom;
40 end = new atom;
41 start->father = NULL;
42 end->father = NULL;
43 link(start,end);
44 // init bond chain list
45 first = new bond(start, end, 1, -1);
46 last = new bond(start, end, 1, -1);
47 link(first,last);
48 // other stuff
49 MDSteps = 0;
50 last_atom = 0;
51 elemente = teil;
52 AtomCount = 0;
53 BondCount = 0;
54 NoNonBonds = 0;
55 NoNonHydrogen = 0;
56 NoCyclicBonds = 0;
57 ListOfBondsPerAtom = NULL;
58 NumberOfBondsPerAtom = NULL;
59 ElementCount = 0;
60 for(int i=MAX_ELEMENTS;i--;)
61 ElementsInMolecule[i] = 0;
62 cell_size[0] = cell_size[2] = cell_size[5]= 20.;
63 cell_size[1] = cell_size[3] = cell_size[4]= 0.;
64};
65
66/** Destructor of class molecule.
67 * Initialises molecule list with correctly referenced start and end, and sets molecule::last_atom to zero.
68 */
69molecule::~molecule()
70{
71 if (ListOfBondsPerAtom != NULL)
72 for(int i=AtomCount;i--;)
73 Free((void **)&ListOfBondsPerAtom[i], "molecule::~molecule: ListOfBondsPerAtom[i]");
74 Free((void **)&ListOfBondsPerAtom, "molecule::~molecule: ListOfBondsPerAtom");
75 Free((void **)&NumberOfBondsPerAtom, "molecule::~molecule: NumberOfBondsPerAtom");
76 CleanupMolecule();
77 delete(first);
78 delete(last);
79 delete(end);
80 delete(start);
81};
82
83/** Adds given atom \a *pointer from molecule list.
84 * Increases molecule::last_atom and gives last number to added atom and names it according to its element::abbrev and molecule::AtomCount
85 * \param *pointer allocated and set atom
86 * \return true - succeeded, false - atom not found in list
87 */
88bool molecule::AddAtom(atom *pointer)
89{
90 if (pointer != NULL) {
91 pointer->sort = &pointer->nr;
92 pointer->nr = last_atom++; // increase number within molecule
93 AtomCount++;
94 if (pointer->type != NULL) {
95 if (ElementsInMolecule[pointer->type->Z] == 0)
96 ElementCount++;
97 ElementsInMolecule[pointer->type->Z]++; // increase number of elements
98 if (pointer->type->Z != 1)
99 NoNonHydrogen++;
100 if (pointer->Name == NULL) {
101 Free((void **)&pointer->Name, "molecule::AddAtom: *pointer->Name");
102 pointer->Name = (char *) Malloc(sizeof(char)*6, "molecule::AddAtom: *pointer->Name");
103 sprintf(pointer->Name, "%2s%02d", pointer->type->symbol, pointer->nr+1);
104 }
105 }
106 return add(pointer, end);
107 } else
108 return false;
109};
110
111/** Adds a copy of the given atom \a *pointer from molecule list.
112 * Increases molecule::last_atom and gives last number to added atom.
113 * \param *pointer allocated and set atom
114 * \return true - succeeded, false - atom not found in list
115 */
116atom * molecule::AddCopyAtom(atom *pointer)
117{
118 if (pointer != NULL) {
119 atom *walker = new atom();
120 walker->type = pointer->type; // copy element of atom
121 walker->x.CopyVector(&pointer->x); // copy coordination
122 walker->v.CopyVector(&pointer->v); // copy velocity
123 walker->FixedIon = pointer->FixedIon;
124 walker->sort = &walker->nr;
125 walker->nr = last_atom++; // increase number within molecule
126 walker->father = pointer; //->GetTrueFather();
127 walker->Name = (char *) Malloc(sizeof(char)*strlen(pointer->Name)+1, "molecule::AddCopyAtom: *Name");
128 strcpy (walker->Name, pointer->Name);
129 add(walker, end);
130 if ((pointer->type != NULL) && (pointer->type->Z != 1))
131 NoNonHydrogen++;
132 AtomCount++;
133 return walker;
134 } else
135 return NULL;
136};
137
138/** Adds a Hydrogen atom in replacement for the given atom \a *partner in bond with a *origin.
139 * Here, we have to distinguish between single, double or triple bonds as stated by \a BondDegree, that each demand
140 * a different scheme when adding \a *replacement atom for the given one.
141 * -# Single Bond: Simply add new atom with bond distance rescaled to typical hydrogen one
142 * -# Double Bond: Here, we need the **BondList of the \a *origin atom, by scanning for the other bonds instead of
143 * *Bond, we use the through these connected atoms to determine the plane they lie in, vector::MakeNormalvector().
144 * The orthonormal vector to this plane along with the vector in *Bond direction determines the plane the two
145 * replacing hydrogens shall lie in. Now, all remains to do is take the usual hydrogen double bond angle for the
146 * element of *origin and form the sin/cos admixture of both plane vectors for the new coordinates of the two
147 * hydrogens forming this angle with *origin.
148 * -# Triple Bond: The idea is to set up a tetraoid (C1-H1-H2-H3) (however the lengths \f$b\f$ of the sides of the base
149 * triangle formed by the to be added hydrogens are not equal to the typical bond distance \f$l\f$ but have to be
150 * determined from the typical angle \f$\alpha\f$ for a hydrogen triple connected to the element of *origin):
151 * We have the height \f$d\f$ as the vector in *Bond direction (from triangle C1-H1-H2).
152 * \f[ h = l \cdot \cos{\left (\frac{\alpha}{2} \right )} \qquad b = 2l \cdot \sin{\left (\frac{\alpha}{2} \right)} \quad \rightarrow \quad d = l \cdot \sqrt{\cos^2{\left (\frac{\alpha}{2} \right)}-\frac{1}{3}\cdot\sin^2{\left (\frac{\alpha}{2}\right )}}
153 * \f]
154 * vector::GetNormalvector() creates one orthonormal vector from this *Bond vector and vector::MakeNormalvector creates
155 * the third one from the former two vectors. The latter ones form the plane of the base triangle mentioned above.
156 * The lengths for these are \f$f\f$ and \f$g\f$ (from triangle H1-H2-(center of H1-H2-H3)) with knowledge that
157 * the median lines in an isosceles triangle meet in the center point with a ratio 2:1.
158 * \f[ f = \frac{b}{\sqrt{3}} \qquad g = \frac{b}{2}
159 * \f]
160 * as the coordination of all three atoms in the coordinate system of these three vectors:
161 * \f$\pmatrix{d & f & 0}\f$, \f$\pmatrix{d & -0.5 \cdot f & g}\f$ and \f$\pmatrix{d & -0.5 \cdot f & -g}\f$.
162 *
163 * \param *out output stream for debugging
164 * \param *Bond pointer to bond between \a *origin and \a *replacement
165 * \param *TopOrigin son of \a *origin of upper level molecule (the atom added to this molecule as a copy of \a *origin)
166 * \param *origin pointer to atom which acts as the origin for scaling the added hydrogen to correct bond length
167 * \param *replacement pointer to the atom which shall be copied as a hydrogen atom in this molecule
168 * \param **BondList list of bonds \a *replacement has (necessary to determine plane for double and triple bonds)
169 * \param NumBond number of bonds in \a **BondList
170 * \param isAngstroem whether the coordination of the given atoms is in AtomicLength (false) or Angstrom(true)
171 * \return number of atoms added, if < bond::BondDegree then something went wrong
172 * \todo double and triple bonds splitting (always use the tetraeder angle!)
173 */
174bool molecule::AddHydrogenReplacementAtom(ofstream *out, bond *TopBond, atom *BottomOrigin, atom *TopOrigin, atom *TopReplacement, bond **BondList, int NumBond, bool IsAngstroem)
175{
176 double bondlength; // bond length of the bond to be replaced/cut
177 double bondangle; // bond angle of the bond to be replaced/cut
178 double BondRescale; // rescale value for the hydrogen bond length
179 bool AllWentWell = true; // flag gathering the boolean return value of molecule::AddAtom and other functions, as return value on exit
180 bond *FirstBond = NULL, *SecondBond = NULL; // Other bonds in double bond case to determine "other" plane
181 atom *FirstOtherAtom = NULL, *SecondOtherAtom = NULL, *ThirdOtherAtom = NULL; // pointer to hydrogen atoms to be added
182 double b,l,d,f,g, alpha, factors[NDIM]; // hold temporary values in triple bond case for coordination determination
183 Vector Orthovector1, Orthovector2; // temporary vectors in coordination construction
184 Vector InBondvector; // vector in direction of *Bond
185 bond *Binder = NULL;
186 double *matrix;
187
188// *out << Verbose(3) << "Begin of AddHydrogenReplacementAtom." << endl;
189 // create vector in direction of bond
190 InBondvector.CopyVector(&TopReplacement->x);
191 InBondvector.SubtractVector(&TopOrigin->x);
192 bondlength = InBondvector.Norm();
193
194 // is greater than typical bond distance? Then we have to correct periodically
195 // the problem is not the H being out of the box, but InBondvector have the wrong direction
196 // due to TopReplacement or Origin being on the wrong side!
197 if (bondlength > BondDistance) {
198// *out << Verbose(4) << "InBondvector is: ";
199// InBondvector.Output(out);
200// *out << endl;
201 Orthovector1.Zero();
202 for (int i=NDIM;i--;) {
203 l = TopReplacement->x.x[i] - TopOrigin->x.x[i];
204 if (fabs(l) > BondDistance) { // is component greater than bond distance
205 Orthovector1.x[i] = (l < 0) ? -1. : +1.;
206 } // (signs are correct, was tested!)
207 }
208 matrix = ReturnFullMatrixforSymmetric(cell_size);
209 Orthovector1.MatrixMultiplication(matrix);
210 InBondvector.SubtractVector(&Orthovector1); // subtract just the additional translation
211 Free((void **)&matrix, "molecule::AddHydrogenReplacementAtom: *matrix");
212 bondlength = InBondvector.Norm();
213// *out << Verbose(4) << "Corrected InBondvector is now: ";
214// InBondvector.Output(out);
215// *out << endl;
216 } // periodic correction finished
217
218 InBondvector.Normalize();
219 // get typical bond length and store as scale factor for later
220 BondRescale = TopOrigin->type->HBondDistance[TopBond->BondDegree-1];
221 if (BondRescale == -1) {
222 cerr << Verbose(3) << "ERROR: There is no typical hydrogen bond distance in replacing bond (" << TopOrigin->Name << "<->" << TopReplacement->Name << ") of degree " << TopBond->BondDegree << "!" << endl;
223 return false;
224 BondRescale = bondlength;
225 } else {
226 if (!IsAngstroem)
227 BondRescale /= (1.*AtomicLengthToAngstroem);
228 }
229
230 // discern single, double and triple bonds
231 switch(TopBond->BondDegree) {
232 case 1:
233 FirstOtherAtom = new atom(); // new atom
234 FirstOtherAtom->type = elemente->FindElement(1); // element is Hydrogen
235 FirstOtherAtom->v.CopyVector(&TopReplacement->v); // copy velocity
236 FirstOtherAtom->FixedIon = TopReplacement->FixedIon;
237 if (TopReplacement->type->Z == 1) { // neither rescale nor replace if it's already hydrogen
238 FirstOtherAtom->father = TopReplacement;
239 BondRescale = bondlength;
240 } else {
241 FirstOtherAtom->father = NULL; // if we replace hydrogen, we mark it as our father, otherwise we are just an added hydrogen with no father
242 }
243 InBondvector.Scale(&BondRescale); // rescale the distance vector to Hydrogen bond length
244 FirstOtherAtom->x.CopyVector(&TopOrigin->x); // set coordination to origin ...
245 FirstOtherAtom->x.AddVector(&InBondvector); // ... and add distance vector to replacement atom
246 AllWentWell = AllWentWell && AddAtom(FirstOtherAtom);
247// *out << Verbose(4) << "Added " << *FirstOtherAtom << " at: ";
248// FirstOtherAtom->x.Output(out);
249// *out << endl;
250 Binder = AddBond(BottomOrigin, FirstOtherAtom, 1);
251 Binder->Cyclic = false;
252 Binder->Type = TreeEdge;
253 break;
254 case 2:
255 // determine two other bonds (warning if there are more than two other) plus valence sanity check
256 for (int i=0;i<NumBond;i++) {
257 if (BondList[i] != TopBond) {
258 if (FirstBond == NULL) {
259 FirstBond = BondList[i];
260 FirstOtherAtom = BondList[i]->GetOtherAtom(TopOrigin);
261 } else if (SecondBond == NULL) {
262 SecondBond = BondList[i];
263 SecondOtherAtom = BondList[i]->GetOtherAtom(TopOrigin);
264 } else {
265 *out << Verbose(3) << "WARNING: Detected more than four bonds for atom " << TopOrigin->Name;
266 }
267 }
268 }
269 if (SecondOtherAtom == NULL) { // then we have an atom with valence four, but only 3 bonds: one to replace and one which is TopBond (third is FirstBond)
270 SecondBond = TopBond;
271 SecondOtherAtom = TopReplacement;
272 }
273 if (FirstOtherAtom != NULL) { // then we just have this double bond and the plane does not matter at all
274// *out << Verbose(3) << "Regarding the double bond (" << TopOrigin->Name << "<->" << TopReplacement->Name << ") to be constructed: Taking " << FirstOtherAtom->Name << " and " << SecondOtherAtom->Name << " along with " << TopOrigin->Name << " to determine orthogonal plane." << endl;
275
276 // determine the plane of these two with the *origin
277 AllWentWell = AllWentWell && Orthovector1.MakeNormalVector(&TopOrigin->x, &FirstOtherAtom->x, &SecondOtherAtom->x);
278 } else {
279 Orthovector1.GetOneNormalVector(&InBondvector);
280 }
281 //*out << Verbose(3)<< "Orthovector1: ";
282 //Orthovector1.Output(out);
283 //*out << endl;
284 // orthogonal vector and bond vector between origin and replacement form the new plane
285 Orthovector1.MakeNormalVector(&InBondvector);
286 Orthovector1.Normalize();
287 //*out << Verbose(3) << "ReScaleCheck: " << Orthovector1.Norm() << " and " << InBondvector.Norm() << "." << endl;
288
289 // create the two Hydrogens ...
290 FirstOtherAtom = new atom();
291 SecondOtherAtom = new atom();
292 FirstOtherAtom->type = elemente->FindElement(1);
293 SecondOtherAtom->type = elemente->FindElement(1);
294 FirstOtherAtom->v.CopyVector(&TopReplacement->v); // copy velocity
295 FirstOtherAtom->FixedIon = TopReplacement->FixedIon;
296 SecondOtherAtom->v.CopyVector(&TopReplacement->v); // copy velocity
297 SecondOtherAtom->FixedIon = TopReplacement->FixedIon;
298 FirstOtherAtom->father = NULL; // we are just an added hydrogen with no father
299 SecondOtherAtom->father = NULL; // we are just an added hydrogen with no father
300 bondangle = TopOrigin->type->HBondAngle[1];
301 if (bondangle == -1) {
302 *out << Verbose(3) << "ERROR: There is no typical hydrogen bond angle in replacing bond (" << TopOrigin->Name << "<->" << TopReplacement->Name << ") of degree " << TopBond->BondDegree << "!" << endl;
303 return false;
304 bondangle = 0;
305 }
306 bondangle *= M_PI/180./2.;
307// *out << Verbose(3) << "ReScaleCheck: InBondvector ";
308// InBondvector.Output(out);
309// *out << endl;
310// *out << Verbose(3) << "ReScaleCheck: Orthovector ";
311// Orthovector1.Output(out);
312// *out << endl;
313// *out << Verbose(3) << "Half the bond angle is " << bondangle << ", sin and cos of it: " << sin(bondangle) << ", " << cos(bondangle) << endl;
314 FirstOtherAtom->x.Zero();
315 SecondOtherAtom->x.Zero();
316 for(int i=NDIM;i--;) { // rotate by half the bond angle in both directions (InBondvector is bondangle = 0 direction)
317 FirstOtherAtom->x.x[i] = InBondvector.x[i] * cos(bondangle) + Orthovector1.x[i] * (sin(bondangle));
318 SecondOtherAtom->x.x[i] = InBondvector.x[i] * cos(bondangle) + Orthovector1.x[i] * (-sin(bondangle));
319 }
320 FirstOtherAtom->x.Scale(&BondRescale); // rescale by correct BondDistance
321 SecondOtherAtom->x.Scale(&BondRescale);
322 //*out << Verbose(3) << "ReScaleCheck: " << FirstOtherAtom->x.Norm() << " and " << SecondOtherAtom->x.Norm() << "." << endl;
323 for(int i=NDIM;i--;) { // and make relative to origin atom
324 FirstOtherAtom->x.x[i] += TopOrigin->x.x[i];
325 SecondOtherAtom->x.x[i] += TopOrigin->x.x[i];
326 }
327 // ... and add to molecule
328 AllWentWell = AllWentWell && AddAtom(FirstOtherAtom);
329 AllWentWell = AllWentWell && AddAtom(SecondOtherAtom);
330// *out << Verbose(4) << "Added " << *FirstOtherAtom << " at: ";
331// FirstOtherAtom->x.Output(out);
332// *out << endl;
333// *out << Verbose(4) << "Added " << *SecondOtherAtom << " at: ";
334// SecondOtherAtom->x.Output(out);
335// *out << endl;
336 Binder = AddBond(BottomOrigin, FirstOtherAtom, 1);
337 Binder->Cyclic = false;
338 Binder->Type = TreeEdge;
339 Binder = AddBond(BottomOrigin, SecondOtherAtom, 1);
340 Binder->Cyclic = false;
341 Binder->Type = TreeEdge;
342 break;
343 case 3:
344 // take the "usual" tetraoidal angle and add the three Hydrogen in direction of the bond (height of the tetraoid)
345 FirstOtherAtom = new atom();
346 SecondOtherAtom = new atom();
347 ThirdOtherAtom = new atom();
348 FirstOtherAtom->type = elemente->FindElement(1);
349 SecondOtherAtom->type = elemente->FindElement(1);
350 ThirdOtherAtom->type = elemente->FindElement(1);
351 FirstOtherAtom->v.CopyVector(&TopReplacement->v); // copy velocity
352 FirstOtherAtom->FixedIon = TopReplacement->FixedIon;
353 SecondOtherAtom->v.CopyVector(&TopReplacement->v); // copy velocity
354 SecondOtherAtom->FixedIon = TopReplacement->FixedIon;
355 ThirdOtherAtom->v.CopyVector(&TopReplacement->v); // copy velocity
356 ThirdOtherAtom->FixedIon = TopReplacement->FixedIon;
357 FirstOtherAtom->father = NULL; // we are just an added hydrogen with no father
358 SecondOtherAtom->father = NULL; // we are just an added hydrogen with no father
359 ThirdOtherAtom->father = NULL; // we are just an added hydrogen with no father
360
361 // we need to vectors orthonormal the InBondvector
362 AllWentWell = AllWentWell && Orthovector1.GetOneNormalVector(&InBondvector);
363// *out << Verbose(3) << "Orthovector1: ";
364// Orthovector1.Output(out);
365// *out << endl;
366 AllWentWell = AllWentWell && Orthovector2.MakeNormalVector(&InBondvector, &Orthovector1);
367// *out << Verbose(3) << "Orthovector2: ";
368// Orthovector2.Output(out);
369// *out << endl;
370
371 // create correct coordination for the three atoms
372 alpha = (TopOrigin->type->HBondAngle[2])/180.*M_PI/2.; // retrieve triple bond angle from database
373 l = BondRescale; // desired bond length
374 b = 2.*l*sin(alpha); // base length of isosceles triangle
375 d = l*sqrt(cos(alpha)*cos(alpha) - sin(alpha)*sin(alpha)/3.); // length for InBondvector
376 f = b/sqrt(3.); // length for Orthvector1
377 g = b/2.; // length for Orthvector2
378// *out << Verbose(3) << "Bond length and half-angle: " << l << ", " << alpha << "\t (b,d,f,g) = " << b << ", " << d << ", " << f << ", " << g << ", " << endl;
379// *out << Verbose(3) << "The three Bond lengths: " << sqrt(d*d+f*f) << ", " << sqrt(d*d+(-0.5*f)*(-0.5*f)+g*g) << ", " << sqrt(d*d+(-0.5*f)*(-0.5*f)+g*g) << endl;
380 factors[0] = d;
381 factors[1] = f;
382 factors[2] = 0.;
383 FirstOtherAtom->x.LinearCombinationOfVectors(&InBondvector, &Orthovector1, &Orthovector2, factors);
384 factors[1] = -0.5*f;
385 factors[2] = g;
386 SecondOtherAtom->x.LinearCombinationOfVectors(&InBondvector, &Orthovector1, &Orthovector2, factors);
387 factors[2] = -g;
388 ThirdOtherAtom->x.LinearCombinationOfVectors(&InBondvector, &Orthovector1, &Orthovector2, factors);
389
390 // rescale each to correct BondDistance
391// FirstOtherAtom->x.Scale(&BondRescale);
392// SecondOtherAtom->x.Scale(&BondRescale);
393// ThirdOtherAtom->x.Scale(&BondRescale);
394
395 // and relative to *origin atom
396 FirstOtherAtom->x.AddVector(&TopOrigin->x);
397 SecondOtherAtom->x.AddVector(&TopOrigin->x);
398 ThirdOtherAtom->x.AddVector(&TopOrigin->x);
399
400 // ... and add to molecule
401 AllWentWell = AllWentWell && AddAtom(FirstOtherAtom);
402 AllWentWell = AllWentWell && AddAtom(SecondOtherAtom);
403 AllWentWell = AllWentWell && AddAtom(ThirdOtherAtom);
404// *out << Verbose(4) << "Added " << *FirstOtherAtom << " at: ";
405// FirstOtherAtom->x.Output(out);
406// *out << endl;
407// *out << Verbose(4) << "Added " << *SecondOtherAtom << " at: ";
408// SecondOtherAtom->x.Output(out);
409// *out << endl;
410// *out << Verbose(4) << "Added " << *ThirdOtherAtom << " at: ";
411// ThirdOtherAtom->x.Output(out);
412// *out << endl;
413 Binder = AddBond(BottomOrigin, FirstOtherAtom, 1);
414 Binder->Cyclic = false;
415 Binder->Type = TreeEdge;
416 Binder = AddBond(BottomOrigin, SecondOtherAtom, 1);
417 Binder->Cyclic = false;
418 Binder->Type = TreeEdge;
419 Binder = AddBond(BottomOrigin, ThirdOtherAtom, 1);
420 Binder->Cyclic = false;
421 Binder->Type = TreeEdge;
422 break;
423 default:
424 cerr << "ERROR: BondDegree does not state single, double or triple bond!" << endl;
425 AllWentWell = false;
426 break;
427 }
428
429// *out << Verbose(3) << "End of AddHydrogenReplacementAtom." << endl;
430 return AllWentWell;
431};
432
433/** Adds given atom \a *pointer from molecule list.
434 * Increases molecule::last_atom and gives last number to added atom.
435 * \param filename name and path of xyz file
436 * \return true - succeeded, false - file not found
437 */
438bool molecule::AddXYZFile(string filename)
439{
440 istringstream *input = NULL;
441 int NumberOfAtoms = 0; // atom number in xyz read
442 int i, j; // loop variables
443 atom *Walker = NULL; // pointer to added atom
444 char shorthand[3]; // shorthand for atom name
445 ifstream xyzfile; // xyz file
446 string line; // currently parsed line
447 double x[3]; // atom coordinates
448
449 xyzfile.open(filename.c_str());
450 if (!xyzfile)
451 return false;
452
453 getline(xyzfile,line,'\n'); // Read numer of atoms in file
454 input = new istringstream(line);
455 *input >> NumberOfAtoms;
456 cout << Verbose(0) << "Parsing " << NumberOfAtoms << " atoms in file." << endl;
457 getline(xyzfile,line,'\n'); // Read comment
458 cout << Verbose(1) << "Comment: " << line << endl;
459
460 if (MDSteps == 0) // no atoms yet present
461 MDSteps++;
462 for(i=0;i<NumberOfAtoms;i++){
463 Walker = new atom;
464 getline(xyzfile,line,'\n');
465 istringstream *item = new istringstream(line);
466 //istringstream input(line);
467 //cout << Verbose(1) << "Reading: " << line << endl;
468 *item >> shorthand;
469 *item >> x[0];
470 *item >> x[1];
471 *item >> x[2];
472 Walker->type = elemente->FindElement(shorthand);
473 if (Walker->type == NULL) {
474 cerr << "Could not parse the element at line: '" << line << "', setting to H.";
475 Walker->type = elemente->FindElement(1);
476 }
477 if (Trajectories[Walker].R.size() <= (unsigned int)MDSteps) {
478 Trajectories[Walker].R.resize(MDSteps+10);
479 Trajectories[Walker].U.resize(MDSteps+10);
480 Trajectories[Walker].F.resize(MDSteps+10);
481 }
482 for(j=NDIM;j--;) {
483 Walker->x.x[j] = x[j];
484 Trajectories[Walker].R.at(MDSteps-1).x[j] = x[j];
485 Trajectories[Walker].U.at(MDSteps-1).x[j] = 0;
486 Trajectories[Walker].F.at(MDSteps-1).x[j] = 0;
487 }
488 AddAtom(Walker); // add to molecule
489 delete(item);
490 }
491 xyzfile.close();
492 delete(input);
493 return true;
494};
495
496/** Creates a copy of this molecule.
497 * \return copy of molecule
498 */
499molecule *molecule::CopyMolecule()
500{
501 molecule *copy = new molecule(elemente);
502 atom *CurrentAtom = NULL;
503 atom *LeftAtom = NULL, *RightAtom = NULL;
504 atom *Walker = NULL;
505
506 // copy all atoms
507 Walker = start;
508 while(Walker->next != end) {
509 Walker = Walker->next;
510 CurrentAtom = copy->AddCopyAtom(Walker);
511 }
512
513 // copy all bonds
514 bond *Binder = first;
515 bond *NewBond = NULL;
516 while(Binder->next != last) {
517 Binder = Binder->next;
518 // get the pendant atoms of current bond in the copy molecule
519 LeftAtom = copy->start;
520 while (LeftAtom->next != copy->end) {
521 LeftAtom = LeftAtom->next;
522 if (LeftAtom->father == Binder->leftatom)
523 break;
524 }
525 RightAtom = copy->start;
526 while (RightAtom->next != copy->end) {
527 RightAtom = RightAtom->next;
528 if (RightAtom->father == Binder->rightatom)
529 break;
530 }
531 NewBond = copy->AddBond(LeftAtom, RightAtom, Binder->BondDegree);
532 NewBond->Cyclic = Binder->Cyclic;
533 if (Binder->Cyclic)
534 copy->NoCyclicBonds++;
535 NewBond->Type = Binder->Type;
536 }
537 // correct fathers
538 Walker = copy->start;
539 while(Walker->next != copy->end) {
540 Walker = Walker->next;
541 if (Walker->father->father == Walker->father) // same atom in copy's father points to itself
542 Walker->father = Walker; // set father to itself (copy of a whole molecule)
543 else
544 Walker->father = Walker->father->father; // set father to original's father
545 }
546 // copy values
547 copy->CountAtoms((ofstream *)&cout);
548 copy->CountElements();
549 if (first->next != last) { // if adjaceny list is present
550 copy->BondDistance = BondDistance;
551 copy->CreateListOfBondsPerAtom((ofstream *)&cout);
552 }
553
554 return copy;
555};
556
557/** Adds a bond to a the molecule specified by two atoms, \a *first and \a *second.
558 * Also updates molecule::BondCount and molecule::NoNonBonds.
559 * \param *first first atom in bond
560 * \param *second atom in bond
561 * \return pointer to bond or NULL on failure
562 */
563bond * molecule::AddBond(atom *atom1, atom *atom2, int degree=1)
564{
565 bond *Binder = NULL;
566 if ((atom1 != NULL) && (FindAtom(atom1->nr) != NULL) && (atom2 != NULL) && (FindAtom(atom2->nr) != NULL)) {
567 Binder = new bond(atom1, atom2, degree, BondCount++);
568 if ((atom1->type != NULL) && (atom1->type->Z != 1) && (atom2->type != NULL) && (atom2->type->Z != 1))
569 NoNonBonds++;
570 add(Binder, last);
571 } else {
572 cerr << Verbose(1) << "ERROR: Could not add bond between " << atom1->Name << " and " << atom2->Name << " as one or both are not present in the molecule." << endl;
573 }
574 return Binder;
575};
576
577/** Remove bond from bond chain list.
578 * \todo Function not implemented yet
579 * \param *pointer bond pointer
580 * \return true - bound found and removed, false - bond not found/removed
581 */
582bool molecule::RemoveBond(bond *pointer)
583{
584 //cerr << Verbose(1) << "molecule::RemoveBond: Function not implemented yet." << endl;
585 removewithoutcheck(pointer);
586 return true;
587};
588
589/** Remove every bond from bond chain list that atom \a *BondPartner is a constituent of.
590 * \todo Function not implemented yet
591 * \param *BondPartner atom to be removed
592 * \return true - bounds found and removed, false - bonds not found/removed
593 */
594bool molecule::RemoveBonds(atom *BondPartner)
595{
596 cerr << Verbose(1) << "molecule::RemoveBond: Function not implemented yet." << endl;
597 return false;
598};
599
600/** Sets the molecule::cell_size to the components of \a *dim (rectangular box)
601 * \param *dim vector class
602 */
603void molecule::SetBoxDimension(Vector *dim)
604{
605 cell_size[0] = dim->x[0];
606 cell_size[1] = 0.;
607 cell_size[2] = dim->x[1];
608 cell_size[3] = 0.;
609 cell_size[4] = 0.;
610 cell_size[5] = dim->x[2];
611};
612
613/** Centers the molecule in the box whose lengths are defined by vector \a *BoxLengths.
614 * \param *out output stream for debugging
615 * \param *BoxLengths box lengths
616 */
617bool molecule::CenterInBox(ofstream *out, Vector *BoxLengths)
618{
619 bool status = true;
620 atom *ptr = NULL;
621 Vector *min = new Vector;
622 Vector *max = new Vector;
623
624 // gather min and max for each axis
625 ptr = start->next; // start at first in list
626 if (ptr != end) { //list not empty?
627 for (int i=NDIM;i--;) {
628 max->x[i] = ptr->x.x[i];
629 min->x[i] = ptr->x.x[i];
630 }
631 while (ptr->next != end) { // continue with second if present
632 ptr = ptr->next;
633 //ptr->Output(1,1,out);
634 for (int i=NDIM;i--;) {
635 max->x[i] = (max->x[i] < ptr->x.x[i]) ? ptr->x.x[i] : max->x[i];
636 min->x[i] = (min->x[i] > ptr->x.x[i]) ? ptr->x.x[i] : min->x[i];
637 }
638 }
639 }
640 // sanity check
641 for(int i=NDIM;i--;) {
642 if (max->x[i] - min->x[i] > BoxLengths->x[i])
643 status = false;
644 }
645 // warn if check failed
646 if (!status)
647 *out << "WARNING: molecule is bigger than defined box!" << endl;
648 else { // else center in box
649 max->AddVector(min);
650 max->Scale(-1.);
651 max->AddVector(BoxLengths);
652 max->Scale(0.5);
653 Translate(max);
654 }
655
656 // free and exit
657 delete(min);
658 delete(max);
659 return status;
660};
661
662/** Centers the edge of the atoms at (0,0,0).
663 * \param *out output stream for debugging
664 * \param *max coordinates of other edge, specifying box dimensions.
665 */
666void molecule::CenterEdge(ofstream *out, Vector *max)
667{
668 Vector *min = new Vector;
669
670// *out << Verbose(3) << "Begin of CenterEdge." << endl;
671 atom *ptr = start->next; // start at first in list
672 if (ptr != end) { //list not empty?
673 for (int i=NDIM;i--;) {
674 max->x[i] = ptr->x.x[i];
675 min->x[i] = ptr->x.x[i];
676 }
677 while (ptr->next != end) { // continue with second if present
678 ptr = ptr->next;
679 //ptr->Output(1,1,out);
680 for (int i=NDIM;i--;) {
681 max->x[i] = (max->x[i] < ptr->x.x[i]) ? ptr->x.x[i] : max->x[i];
682 min->x[i] = (min->x[i] > ptr->x.x[i]) ? ptr->x.x[i] : min->x[i];
683 }
684 }
685// *out << Verbose(4) << "Maximum is ";
686// max->Output(out);
687// *out << ", Minimum is ";
688// min->Output(out);
689// *out << endl;
690 min->Scale(-1.);
691 max->AddVector(min);
692 Translate(min);
693 }
694 delete(min);
695// *out << Verbose(3) << "End of CenterEdge." << endl;
696};
697
698/** Centers the center of the atoms at (0,0,0).
699 * \param *out output stream for debugging
700 * \param *center return vector for translation vector
701 */
702void molecule::CenterOrigin(ofstream *out, Vector *center)
703{
704 int Num = 0;
705 atom *ptr = start->next; // start at first in list
706
707 for(int i=NDIM;i--;) // zero center vector
708 center->x[i] = 0.;
709
710 if (ptr != end) { //list not empty?
711 while (ptr->next != end) { // continue with second if present
712 ptr = ptr->next;
713 Num++;
714 center->AddVector(&ptr->x);
715 }
716 center->Scale(-1./Num); // divide through total number (and sign for direction)
717 Translate(center);
718 }
719};
720
721/** Returns vector pointing to center of gravity.
722 * \param *out output stream for debugging
723 * \return pointer to center of gravity vector
724 */
725Vector * molecule::DetermineCenterOfAll(ofstream *out)
726{
727 atom *ptr = start->next; // start at first in list
728 Vector *a = new Vector();
729 Vector tmp;
730 double Num = 0;
731
732 a->Zero();
733
734 if (ptr != end) { //list not empty?
735 while (ptr->next != end) { // continue with second if present
736 ptr = ptr->next;
737 Num += 1.;
738 tmp.CopyVector(&ptr->x);
739 a->AddVector(&tmp);
740 }
741 a->Scale(-1./Num); // divide through total mass (and sign for direction)
742 }
743 //cout << Verbose(1) << "Resulting center of gravity: ";
744 //a->Output(out);
745 //cout << endl;
746 return a;
747};
748
749/** Returns vector pointing to center of gravity.
750 * \param *out output stream for debugging
751 * \return pointer to center of gravity vector
752 */
753Vector * molecule::DetermineCenterOfGravity(ofstream *out)
754{
755 atom *ptr = start->next; // start at first in list
756 Vector *a = new Vector();
757 Vector tmp;
758 double Num = 0;
759
760 a->Zero();
761
762 if (ptr != end) { //list not empty?
763 while (ptr->next != end) { // continue with second if present
764 ptr = ptr->next;
765 Num += ptr->type->mass;
766 tmp.CopyVector(&ptr->x);
767 tmp.Scale(ptr->type->mass); // scale by mass
768 a->AddVector(&tmp);
769 }
770 a->Scale(-1./Num); // divide through total mass (and sign for direction)
771 }
772// *out << Verbose(1) << "Resulting center of gravity: ";
773// a->Output(out);
774// *out << endl;
775 return a;
776};
777
778/** Centers the center of gravity of the atoms at (0,0,0).
779 * \param *out output stream for debugging
780 * \param *center return vector for translation vector
781 */
782void molecule::CenterGravity(ofstream *out, Vector *center)
783{
784 if (center == NULL) {
785 DetermineCenter(*center);
786 Translate(center);
787 delete(center);
788 } else {
789 Translate(center);
790 }
791};
792
793/** Scales all atoms by \a *factor.
794 * \param *factor pointer to scaling factor
795 */
796void molecule::Scale(double **factor)
797{
798 atom *ptr = start;
799
800 while (ptr->next != end) {
801 ptr = ptr->next;
802 for (int j=0;j<MDSteps;j++)
803 Trajectories[ptr].R.at(j).Scale(factor);
804 ptr->x.Scale(factor);
805 }
806};
807
808/** Translate all atoms by given vector.
809 * \param trans[] translation vector.
810 */
811void molecule::Translate(const Vector *trans)
812{
813 atom *ptr = start;
814
815 while (ptr->next != end) {
816 ptr = ptr->next;
817 for (int j=0;j<MDSteps;j++)
818 Trajectories[ptr].R.at(j).Translate(trans);
819 ptr->x.Translate(trans);
820 }
821};
822
823/** Mirrors all atoms against a given plane.
824 * \param n[] normal vector of mirror plane.
825 */
826void molecule::Mirror(const Vector *n)
827{
828 atom *ptr = start;
829
830 while (ptr->next != end) {
831 ptr = ptr->next;
832 for (int j=0;j<MDSteps;j++)
833 Trajectories[ptr].R.at(j).Mirror(n);
834 ptr->x.Mirror(n);
835 }
836};
837
838/** Determines center of molecule (yet not considering atom masses).
839 * \param Center reference to return vector
840 */
841void molecule::DetermineCenter(Vector &Center)
842{
843 atom *Walker = start;
844 bond *Binder = NULL;
845 double *matrix = ReturnFullMatrixforSymmetric(cell_size);
846 double tmp;
847 bool flag;
848 Vector Testvector, Translationvector;
849
850 do {
851 Center.Zero();
852 flag = true;
853 while (Walker->next != end) {
854 Walker = Walker->next;
855#ifdef ADDHYDROGEN
856 if (Walker->type->Z != 1) {
857#endif
858 Testvector.CopyVector(&Walker->x);
859 Testvector.InverseMatrixMultiplication(matrix);
860 Translationvector.Zero();
861 for (int i=0;i<NumberOfBondsPerAtom[Walker->nr]; i++) {
862 Binder = ListOfBondsPerAtom[Walker->nr][i];
863 if (Walker->nr < Binder->GetOtherAtom(Walker)->nr) // otherwise we shift one to, the other fro and gain nothing
864 for (int j=0;j<NDIM;j++) {
865 tmp = Walker->x.x[j] - Binder->GetOtherAtom(Walker)->x.x[j];
866 if ((fabs(tmp)) > BondDistance) {
867 flag = false;
868 cout << Verbose(0) << "Hit: atom " << Walker->Name << " in bond " << *Binder << " has to be shifted due to " << tmp << "." << endl;
869 if (tmp > 0)
870 Translationvector.x[j] -= 1.;
871 else
872 Translationvector.x[j] += 1.;
873 }
874 }
875 }
876 Testvector.AddVector(&Translationvector);
877 Testvector.MatrixMultiplication(matrix);
878 Center.AddVector(&Testvector);
879 cout << Verbose(1) << "vector is: ";
880 Testvector.Output((ofstream *)&cout);
881 cout << endl;
882#ifdef ADDHYDROGEN
883 // now also change all hydrogens
884 for (int i=0;i<NumberOfBondsPerAtom[Walker->nr]; i++) {
885 Binder = ListOfBondsPerAtom[Walker->nr][i];
886 if (Binder->GetOtherAtom(Walker)->type->Z == 1) {
887 Testvector.CopyVector(&Binder->GetOtherAtom(Walker)->x);
888 Testvector.InverseMatrixMultiplication(matrix);
889 Testvector.AddVector(&Translationvector);
890 Testvector.MatrixMultiplication(matrix);
891 Center.AddVector(&Testvector);
892 cout << Verbose(1) << "Hydrogen vector is: ";
893 Testvector.Output((ofstream *)&cout);
894 cout << endl;
895 }
896 }
897 }
898#endif
899 }
900 } while (!flag);
901 Free((void **)&matrix, "molecule::DetermineCenter: *matrix");
902 Center.Scale(1./(double)AtomCount);
903};
904
905/** Transforms/Rotates the given molecule into its principal axis system.
906 * \param *out output stream for debugging
907 * \param DoRotate whether to rotate (true) or only to determine the PAS.
908 */
909void molecule::PrincipalAxisSystem(ofstream *out, bool DoRotate)
910{
911 atom *ptr = start; // start at first in list
912 double InertiaTensor[NDIM*NDIM];
913 Vector *CenterOfGravity = DetermineCenterOfGravity(out);
914
915 CenterGravity(out, CenterOfGravity);
916
917 // reset inertia tensor
918 for(int i=0;i<NDIM*NDIM;i++)
919 InertiaTensor[i] = 0.;
920
921 // sum up inertia tensor
922 while (ptr->next != end) {
923 ptr = ptr->next;
924 Vector x;
925 x.CopyVector(&ptr->x);
926 //x.SubtractVector(CenterOfGravity);
927 InertiaTensor[0] += ptr->type->mass*(x.x[1]*x.x[1] + x.x[2]*x.x[2]);
928 InertiaTensor[1] += ptr->type->mass*(-x.x[0]*x.x[1]);
929 InertiaTensor[2] += ptr->type->mass*(-x.x[0]*x.x[2]);
930 InertiaTensor[3] += ptr->type->mass*(-x.x[1]*x.x[0]);
931 InertiaTensor[4] += ptr->type->mass*(x.x[0]*x.x[0] + x.x[2]*x.x[2]);
932 InertiaTensor[5] += ptr->type->mass*(-x.x[1]*x.x[2]);
933 InertiaTensor[6] += ptr->type->mass*(-x.x[2]*x.x[0]);
934 InertiaTensor[7] += ptr->type->mass*(-x.x[2]*x.x[1]);
935 InertiaTensor[8] += ptr->type->mass*(x.x[0]*x.x[0] + x.x[1]*x.x[1]);
936 }
937 // print InertiaTensor for debugging
938 *out << "The inertia tensor is:" << endl;
939 for(int i=0;i<NDIM;i++) {
940 for(int j=0;j<NDIM;j++)
941 *out << InertiaTensor[i*NDIM+j] << " ";
942 *out << endl;
943 }
944 *out << endl;
945
946 // diagonalize to determine principal axis system
947 gsl_eigen_symmv_workspace *T = gsl_eigen_symmv_alloc(NDIM);
948 gsl_matrix_view m = gsl_matrix_view_array(InertiaTensor, NDIM, NDIM);
949 gsl_vector *eval = gsl_vector_alloc(NDIM);
950 gsl_matrix *evec = gsl_matrix_alloc(NDIM, NDIM);
951 gsl_eigen_symmv(&m.matrix, eval, evec, T);
952 gsl_eigen_symmv_free(T);
953 gsl_eigen_symmv_sort(eval, evec, GSL_EIGEN_SORT_ABS_DESC);
954
955 for(int i=0;i<NDIM;i++) {
956 *out << Verbose(1) << "eigenvalue = " << gsl_vector_get(eval, i);
957 *out << ", eigenvector = (" << evec->data[i * evec->tda + 0] << "," << evec->data[i * evec->tda + 1] << "," << evec->data[i * evec->tda + 2] << ")" << endl;
958 }
959
960 // check whether we rotate or not
961 if (DoRotate) {
962 *out << Verbose(1) << "Transforming molecule into PAS ... ";
963 // the eigenvectors specify the transformation matrix
964 ptr = start;
965 while (ptr->next != end) {
966 ptr = ptr->next;
967 for (int j=0;j<MDSteps;j++)
968 Trajectories[ptr].R.at(j).MatrixMultiplication(evec->data);
969 ptr->x.MatrixMultiplication(evec->data);
970 }
971 *out << "done." << endl;
972
973 // summing anew for debugging (resulting matrix has to be diagonal!)
974 // reset inertia tensor
975 for(int i=0;i<NDIM*NDIM;i++)
976 InertiaTensor[i] = 0.;
977
978 // sum up inertia tensor
979 ptr = start;
980 while (ptr->next != end) {
981 ptr = ptr->next;
982 Vector x;
983 x.CopyVector(&ptr->x);
984 //x.SubtractVector(CenterOfGravity);
985 InertiaTensor[0] += ptr->type->mass*(x.x[1]*x.x[1] + x.x[2]*x.x[2]);
986 InertiaTensor[1] += ptr->type->mass*(-x.x[0]*x.x[1]);
987 InertiaTensor[2] += ptr->type->mass*(-x.x[0]*x.x[2]);
988 InertiaTensor[3] += ptr->type->mass*(-x.x[1]*x.x[0]);
989 InertiaTensor[4] += ptr->type->mass*(x.x[0]*x.x[0] + x.x[2]*x.x[2]);
990 InertiaTensor[5] += ptr->type->mass*(-x.x[1]*x.x[2]);
991 InertiaTensor[6] += ptr->type->mass*(-x.x[2]*x.x[0]);
992 InertiaTensor[7] += ptr->type->mass*(-x.x[2]*x.x[1]);
993 InertiaTensor[8] += ptr->type->mass*(x.x[0]*x.x[0] + x.x[1]*x.x[1]);
994 }
995 // print InertiaTensor for debugging
996 *out << "The inertia tensor is:" << endl;
997 for(int i=0;i<NDIM;i++) {
998 for(int j=0;j<NDIM;j++)
999 *out << InertiaTensor[i*NDIM+j] << " ";
1000 *out << endl;
1001 }
1002 *out << endl;
1003 }
1004
1005 // free everything
1006 delete(CenterOfGravity);
1007 gsl_vector_free(eval);
1008 gsl_matrix_free(evec);
1009};
1010
1011/** Parses nuclear forces from file and performs Verlet integration.
1012 * Note that we assume the parsed forces to be in atomic units (hence, if coordinates are in angstroem, we
1013 * have to transform them).
1014 * This adds a new MD step to the config file.
1015 * \param *file filename
1016 * \param delta_t time step width in atomic units
1017 * \param IsAngstroem whether coordinates are in angstroem (true) or bohrradius (false)
1018 * \return true - file found and parsed, false - file not found or imparsable
1019 */
1020bool molecule::VerletForceIntegration(char *file, double delta_t, bool IsAngstroem)
1021{
1022 element *runner = elemente->start;
1023 atom *walker = NULL;
1024 int AtomNo;
1025 ifstream input(file);
1026 string token;
1027 stringstream item;
1028 double a, IonMass;
1029 ForceMatrix Force;
1030 Vector tmpvector;
1031
1032 CountElements(); // make sure ElementsInMolecule is up to date
1033
1034 // check file
1035 if (input == NULL) {
1036 return false;
1037 } else {
1038 // parse file into ForceMatrix
1039 if (!Force.ParseMatrix(file, 0,0,0)) {
1040 cerr << "Could not parse Force Matrix file " << file << "." << endl;
1041 return false;
1042 }
1043 if (Force.RowCounter[0] != AtomCount) {
1044 cerr << "Mismatch between number of atoms in file " << Force.RowCounter[0] << " and in molecule " << AtomCount << "." << endl;
1045 return false;
1046 }
1047 // correct Forces
1048// for(int d=0;d<NDIM;d++)
1049// tmpvector.x[d] = 0.;
1050// for(int i=0;i<AtomCount;i++)
1051// for(int d=0;d<NDIM;d++) {
1052// tmpvector.x[d] += Force.Matrix[0][i][d+5];
1053// }
1054// for(int i=0;i<AtomCount;i++)
1055// for(int d=0;d<NDIM;d++) {
1056// Force.Matrix[0][i][d+5] -= tmpvector.x[d]/(double)AtomCount;
1057// }
1058 // and perform Verlet integration for each atom with position, velocity and force vector
1059 runner = elemente->start;
1060 while (runner->next != elemente->end) { // go through every element
1061 runner = runner->next;
1062 IonMass = runner->mass;
1063 a = delta_t*0.5/IonMass; // (F+F_old)/2m = a and thus: v = (F+F_old)/2m * t = (F + F_old) * a
1064 if (ElementsInMolecule[runner->Z]) { // if this element got atoms
1065 AtomNo = 0;
1066 walker = start;
1067 while (walker->next != end) { // go through every atom of this element
1068 walker = walker->next;
1069 if (walker->type == runner) { // if this atom fits to element
1070 // check size of vectors
1071 if (Trajectories[walker].R.size() <= (unsigned int)(MDSteps)) {
1072 //cout << "Increasing size for trajectory array of " << *walker << " to " << (size+10) << "." << endl;
1073 Trajectories[walker].R.resize(MDSteps+10);
1074 Trajectories[walker].U.resize(MDSteps+10);
1075 Trajectories[walker].F.resize(MDSteps+10);
1076 }
1077 // 1. calculate x(t+\delta t)
1078 for (int d=0; d<NDIM; d++) {
1079 Trajectories[walker].F.at(MDSteps).x[d] = -Force.Matrix[0][AtomNo][d+5];
1080 Trajectories[walker].R.at(MDSteps).x[d] = Trajectories[walker].R.at(MDSteps-1).x[d];
1081 Trajectories[walker].R.at(MDSteps).x[d] += delta_t*(Trajectories[walker].U.at(MDSteps-1).x[d]);
1082 Trajectories[walker].R.at(MDSteps).x[d] += 0.5*delta_t*delta_t*(Trajectories[walker].F.at(MDSteps-1).x[d])/IonMass; // F = m * a and s = 0.5 * F/m * t^2 = F * a * t
1083 }
1084 // 2. Calculate v(t+\delta t)
1085 for (int d=0; d<NDIM; d++) {
1086 Trajectories[walker].U.at(MDSteps).x[d] = Trajectories[walker].U.at(MDSteps-1).x[d];
1087 Trajectories[walker].U.at(MDSteps).x[d] += 0.5*delta_t*(Trajectories[walker].F.at(MDSteps-1).x[d]+Trajectories[walker].F.at(MDSteps).x[d])/IonMass;
1088 }
1089// cout << "Integrated position&velocity of step " << (MDSteps) << ": (";
1090// for (int d=0;d<NDIM;d++)
1091// cout << Trajectories[walker].R.at(MDSteps).x[d] << " "; // next step
1092// cout << ")\t(";
1093// for (int d=0;d<NDIM;d++)
1094// cout << Trajectories[walker].U.at(MDSteps).x[d] << " "; // next step
1095// cout << ")" << endl;
1096 // next atom
1097 AtomNo++;
1098 }
1099 }
1100 }
1101 }
1102 }
1103// // correct velocities (rather momenta) so that center of mass remains motionless
1104// tmpvector.zero()
1105// IonMass = 0.;
1106// walker = start;
1107// while (walker->next != end) { // go through every atom
1108// walker = walker->next;
1109// IonMass += walker->type->mass; // sum up total mass
1110// for(int d=0;d<NDIM;d++) {
1111// tmpvector.x[d] += Trajectories[walker].U.at(MDSteps).x[d]*walker->type->mass;
1112// }
1113// }
1114// walker = start;
1115// while (walker->next != end) { // go through every atom of this element
1116// walker = walker->next;
1117// for(int d=0;d<NDIM;d++) {
1118// Trajectories[walker].U.at(MDSteps).x[d] -= tmpvector.x[d]*walker->type->mass/IonMass;
1119// }
1120// }
1121 MDSteps++;
1122
1123
1124 // exit
1125 return true;
1126};
1127
1128/** Align all atoms in such a manner that given vector \a *n is along z axis.
1129 * \param n[] alignment vector.
1130 */
1131void molecule::Align(Vector *n)
1132{
1133 atom *ptr = start;
1134 double alpha, tmp;
1135 Vector z_axis;
1136 z_axis.x[0] = 0.;
1137 z_axis.x[1] = 0.;
1138 z_axis.x[2] = 1.;
1139
1140 // rotate on z-x plane
1141 cout << Verbose(0) << "Begin of Aligning all atoms." << endl;
1142 alpha = atan(-n->x[0]/n->x[2]);
1143 cout << Verbose(1) << "Z-X-angle: " << alpha << " ... ";
1144 while (ptr->next != end) {
1145 ptr = ptr->next;
1146 tmp = ptr->x.x[0];
1147 ptr->x.x[0] = cos(alpha) * tmp + sin(alpha) * ptr->x.x[2];
1148 ptr->x.x[2] = -sin(alpha) * tmp + cos(alpha) * ptr->x.x[2];
1149 for (int j=0;j<MDSteps;j++) {
1150 tmp = Trajectories[ptr].R.at(j).x[0];
1151 Trajectories[ptr].R.at(j).x[0] = cos(alpha) * tmp + sin(alpha) * Trajectories[ptr].R.at(j).x[2];
1152 Trajectories[ptr].R.at(j).x[2] = -sin(alpha) * tmp + cos(alpha) * Trajectories[ptr].R.at(j).x[2];
1153 }
1154 }
1155 // rotate n vector
1156 tmp = n->x[0];
1157 n->x[0] = cos(alpha) * tmp + sin(alpha) * n->x[2];
1158 n->x[2] = -sin(alpha) * tmp + cos(alpha) * n->x[2];
1159 cout << Verbose(1) << "alignment vector after first rotation: ";
1160 n->Output((ofstream *)&cout);
1161 cout << endl;
1162
1163 // rotate on z-y plane
1164 ptr = start;
1165 alpha = atan(-n->x[1]/n->x[2]);
1166 cout << Verbose(1) << "Z-Y-angle: " << alpha << " ... ";
1167 while (ptr->next != end) {
1168 ptr = ptr->next;
1169 tmp = ptr->x.x[1];
1170 ptr->x.x[1] = cos(alpha) * tmp + sin(alpha) * ptr->x.x[2];
1171 ptr->x.x[2] = -sin(alpha) * tmp + cos(alpha) * ptr->x.x[2];
1172 for (int j=0;j<MDSteps;j++) {
1173 tmp = Trajectories[ptr].R.at(j).x[1];
1174 Trajectories[ptr].R.at(j).x[1] = cos(alpha) * tmp + sin(alpha) * Trajectories[ptr].R.at(j).x[2];
1175 Trajectories[ptr].R.at(j).x[2] = -sin(alpha) * tmp + cos(alpha) * Trajectories[ptr].R.at(j).x[2];
1176 }
1177 }
1178 // rotate n vector (for consistency check)
1179 tmp = n->x[1];
1180 n->x[1] = cos(alpha) * tmp + sin(alpha) * n->x[2];
1181 n->x[2] = -sin(alpha) * tmp + cos(alpha) * n->x[2];
1182
1183 cout << Verbose(1) << "alignment vector after second rotation: ";
1184 n->Output((ofstream *)&cout);
1185 cout << Verbose(1) << endl;
1186 cout << Verbose(0) << "End of Aligning all atoms." << endl;
1187};
1188
1189/** Removes atom from molecule list.
1190 * \param *pointer atom to be removed
1191 * \return true - succeeded, false - atom not found in list
1192 */
1193bool molecule::RemoveAtom(atom *pointer)
1194{
1195 if (ElementsInMolecule[pointer->type->Z] != 0) // this would indicate an error
1196 ElementsInMolecule[pointer->type->Z]--; // decrease number of atom of this element
1197 else
1198 cerr << "ERROR: Atom " << pointer->Name << " is of element " << pointer->type->Z << " but the entry in the table of the molecule is 0!" << endl;
1199 if (ElementsInMolecule[pointer->type->Z] == 0) // was last atom of this element?
1200 ElementCount--;
1201 Trajectories.erase(pointer);
1202 return remove(pointer, start, end);
1203};
1204
1205/** Removes every atom from molecule list.
1206 * \return true - succeeded, false - atom not found in list
1207 */
1208bool molecule::CleanupMolecule()
1209{
1210 return (cleanup(start,end) && cleanup(first,last));
1211};
1212
1213/** Finds an atom specified by its continuous number.
1214 * \param Nr number of atom withim molecule
1215 * \return pointer to atom or NULL
1216 */
1217atom * molecule::FindAtom(int Nr) const{
1218 atom * walker = find(&Nr, start,end);
1219 if (walker != NULL) {
1220 //cout << Verbose(0) << "Found Atom Nr. " << walker->nr << endl;
1221 return walker;
1222 } else {
1223 cout << Verbose(0) << "Atom not found in list." << endl;
1224 return NULL;
1225 }
1226};
1227
1228/** Asks for atom number, and checks whether in list.
1229 * \param *text question before entering
1230 */
1231atom * molecule::AskAtom(string text)
1232{
1233 int No;
1234 atom *ion = NULL;
1235 do {
1236 //cout << Verbose(0) << "============Atom list==========================" << endl;
1237 //mol->Output((ofstream *)&cout);
1238 //cout << Verbose(0) << "===============================================" << endl;
1239 cout << Verbose(0) << text;
1240 cin >> No;
1241 ion = this->FindAtom(No);
1242 } while (ion == NULL);
1243 return ion;
1244};
1245
1246/** Checks if given coordinates are within cell volume.
1247 * \param *x array of coordinates
1248 * \return true - is within, false - out of cell
1249 */
1250bool molecule::CheckBounds(const Vector *x) const
1251{
1252 bool result = true;
1253 int j =-1;
1254 for (int i=0;i<NDIM;i++) {
1255 j += i+1;
1256 result = result && ((x->x[i] >= 0) && (x->x[i] < cell_size[j]));
1257 }
1258 //return result;
1259 return true; /// probably not gonna use the check no more
1260};
1261
1262/** Calculates sum over least square distance to line hidden in \a *x.
1263 * \param *x offset and direction vector
1264 * \param *params pointer to lsq_params structure
1265 * \return \f$ sum_i^N | y_i - (a + t_i b)|^2\f$
1266 */
1267double LeastSquareDistance (const gsl_vector * x, void * params)
1268{
1269 double res = 0, t;
1270 Vector a,b,c,d;
1271 struct lsq_params *par = (struct lsq_params *)params;
1272 atom *ptr = par->mol->start;
1273
1274 // initialize vectors
1275 a.x[0] = gsl_vector_get(x,0);
1276 a.x[1] = gsl_vector_get(x,1);
1277 a.x[2] = gsl_vector_get(x,2);
1278 b.x[0] = gsl_vector_get(x,3);
1279 b.x[1] = gsl_vector_get(x,4);
1280 b.x[2] = gsl_vector_get(x,5);
1281 // go through all atoms
1282 while (ptr != par->mol->end) {
1283 ptr = ptr->next;
1284 if (ptr->type == ((struct lsq_params *)params)->type) { // for specific type
1285 c.CopyVector(&ptr->x); // copy vector to temporary one
1286 c.SubtractVector(&a); // subtract offset vector
1287 t = c.ScalarProduct(&b); // get direction parameter
1288 d.CopyVector(&b); // and create vector
1289 d.Scale(&t);
1290 c.SubtractVector(&d); // ... yielding distance vector
1291 res += d.ScalarProduct((const Vector *)&d); // add squared distance
1292 }
1293 }
1294 return res;
1295};
1296
1297/** By minimizing the least square distance gains alignment vector.
1298 * \bug this is not yet working properly it seems
1299 */
1300void molecule::GetAlignvector(struct lsq_params * par) const
1301{
1302 int np = 6;
1303
1304 const gsl_multimin_fminimizer_type *T =
1305 gsl_multimin_fminimizer_nmsimplex;
1306 gsl_multimin_fminimizer *s = NULL;
1307 gsl_vector *ss;
1308 gsl_multimin_function minex_func;
1309
1310 size_t iter = 0, i;
1311 int status;
1312 double size;
1313
1314 /* Initial vertex size vector */
1315 ss = gsl_vector_alloc (np);
1316
1317 /* Set all step sizes to 1 */
1318 gsl_vector_set_all (ss, 1.0);
1319
1320 /* Starting point */
1321 par->x = gsl_vector_alloc (np);
1322 par->mol = this;
1323
1324 gsl_vector_set (par->x, 0, 0.0); // offset
1325 gsl_vector_set (par->x, 1, 0.0);
1326 gsl_vector_set (par->x, 2, 0.0);
1327 gsl_vector_set (par->x, 3, 0.0); // direction
1328 gsl_vector_set (par->x, 4, 0.0);
1329 gsl_vector_set (par->x, 5, 1.0);
1330
1331 /* Initialize method and iterate */
1332 minex_func.f = &LeastSquareDistance;
1333 minex_func.n = np;
1334 minex_func.params = (void *)par;
1335
1336 s = gsl_multimin_fminimizer_alloc (T, np);
1337 gsl_multimin_fminimizer_set (s, &minex_func, par->x, ss);
1338
1339 do
1340 {
1341 iter++;
1342 status = gsl_multimin_fminimizer_iterate(s);
1343
1344 if (status)
1345 break;
1346
1347 size = gsl_multimin_fminimizer_size (s);
1348 status = gsl_multimin_test_size (size, 1e-2);
1349
1350 if (status == GSL_SUCCESS)
1351 {
1352 printf ("converged to minimum at\n");
1353 }
1354
1355 printf ("%5d ", (int)iter);
1356 for (i = 0; i < (size_t)np; i++)
1357 {
1358 printf ("%10.3e ", gsl_vector_get (s->x, i));
1359 }
1360 printf ("f() = %7.3f size = %.3f\n", s->fval, size);
1361 }
1362 while (status == GSL_CONTINUE && iter < 100);
1363
1364 for (i=0;i<(size_t)np;i++)
1365 gsl_vector_set(par->x, i, gsl_vector_get(s->x, i));
1366 //gsl_vector_free(par->x);
1367 gsl_vector_free(ss);
1368 gsl_multimin_fminimizer_free (s);
1369};
1370
1371/** Prints molecule to *out.
1372 * \param *out output stream
1373 */
1374bool molecule::Output(ofstream *out)
1375{
1376 element *runner;
1377 atom *walker = NULL;
1378 int ElementNo, AtomNo;
1379 CountElements();
1380
1381 if (out == NULL) {
1382 return false;
1383 } else {
1384 *out << "#Ion_TypeNr._Nr.R[0] R[1] R[2] MoveType (0 MoveIon, 1 FixedIon)" << endl;
1385 ElementNo = 0;
1386 runner = elemente->start;
1387 while (runner->next != elemente->end) { // go through every element
1388 runner = runner->next;
1389 if (ElementsInMolecule[runner->Z]) { // if this element got atoms
1390 ElementNo++;
1391 AtomNo = 0;
1392 walker = start;
1393 while (walker->next != end) { // go through every atom of this element
1394 walker = walker->next;
1395 if (walker->type == runner) { // if this atom fits to element
1396 AtomNo++;
1397 walker->Output(ElementNo, AtomNo, out); // removed due to trajectories
1398 }
1399 }
1400 }
1401 }
1402 return true;
1403 }
1404};
1405
1406/** Prints molecule with all atomic trajectory positions to *out.
1407 * \param *out output stream
1408 */
1409bool molecule::OutputTrajectories(ofstream *out)
1410{
1411 element *runner = NULL;
1412 atom *walker = NULL;
1413 int ElementNo, AtomNo;
1414 CountElements();
1415
1416 if (out == NULL) {
1417 return false;
1418 } else {
1419 for (int step = 0; step < MDSteps; step++) {
1420 if (step == 0) {
1421 *out << "#Ion_TypeNr._Nr.R[0] R[1] R[2] MoveType (0 MoveIon, 1 FixedIon)" << endl;
1422 } else {
1423 *out << "# ====== MD step " << step << " =========" << endl;
1424 }
1425 ElementNo = 0;
1426 runner = elemente->start;
1427 while (runner->next != elemente->end) { // go through every element
1428 runner = runner->next;
1429 if (ElementsInMolecule[runner->Z]) { // if this element got atoms
1430 ElementNo++;
1431 AtomNo = 0;
1432 walker = start;
1433 while (walker->next != end) { // go through every atom of this element
1434 walker = walker->next;
1435 if (walker->type == runner) { // if this atom fits to element
1436 AtomNo++;
1437 *out << "Ion_Type" << ElementNo << "_" << AtomNo << "\t" << fixed << setprecision(9) << showpoint;
1438 *out << Trajectories[walker].R.at(step).x[0] << "\t" << Trajectories[walker].R.at(step).x[1] << "\t" << Trajectories[walker].R.at(step).x[2];
1439 *out << "\t" << walker->FixedIon;
1440 if (Trajectories[walker].U.at(step).Norm() > MYEPSILON)
1441 *out << "\t" << scientific << setprecision(6) << Trajectories[walker].U.at(step).x[0] << "\t" << Trajectories[walker].U.at(step).x[1] << "\t" << Trajectories[walker].U.at(step).x[2] << "\t";
1442 if (Trajectories[walker].F.at(step).Norm() > MYEPSILON)
1443 *out << "\t" << scientific << setprecision(6) << Trajectories[walker].F.at(step).x[0] << "\t" << Trajectories[walker].F.at(step).x[1] << "\t" << Trajectories[walker].F.at(step).x[2] << "\t";
1444 *out << "\t# Number in molecule " << walker->nr << endl;
1445 }
1446 }
1447 }
1448 }
1449 }
1450 return true;
1451 }
1452};
1453
1454/** Outputs contents of molecule::ListOfBondsPerAtom.
1455 * \param *out output stream
1456 */
1457void molecule::OutputListOfBonds(ofstream *out) const
1458{
1459 *out << Verbose(2) << endl << "From Contents of ListOfBondsPerAtom, all non-hydrogen atoms:" << endl;
1460 atom *Walker = start;
1461 while (Walker->next != end) {
1462 Walker = Walker->next;
1463#ifdef ADDHYDROGEN
1464 if (Walker->type->Z != 1) { // regard only non-hydrogen
1465#endif
1466 *out << Verbose(2) << "Atom " << Walker->Name << " has Bonds: "<<endl;
1467 for(int j=0;j<NumberOfBondsPerAtom[Walker->nr];j++) {
1468 *out << Verbose(3) << *(ListOfBondsPerAtom)[Walker->nr][j] << endl;
1469 }
1470#ifdef ADDHYDROGEN
1471 }
1472#endif
1473 }
1474 *out << endl;
1475};
1476
1477/** Output of element before the actual coordination list.
1478 * \param *out stream pointer
1479 */
1480bool molecule::Checkout(ofstream *out) const
1481{
1482 return elemente->Checkout(out, ElementsInMolecule);
1483};
1484
1485/** Prints molecule with all its trajectories to *out as xyz file.
1486 * \param *out output stream
1487 */
1488bool molecule::OutputTrajectoriesXYZ(ofstream *out)
1489{
1490 atom *walker = NULL;
1491 int No = 0;
1492 time_t now;
1493
1494 now = time((time_t *)NULL); // Get the system time and put it into 'now' as 'calender time'
1495 walker = start;
1496 while (walker->next != end) { // go through every atom and count
1497 walker = walker->next;
1498 No++;
1499 }
1500 if (out != NULL) {
1501 for (int step=0;step<MDSteps;step++) {
1502 *out << No << "\n\tCreated by molecuilder, step " << step << ", on " << ctime(&now);
1503 walker = start;
1504 while (walker->next != end) { // go through every atom of this element
1505 walker = walker->next;
1506 *out << walker->type->symbol << "\t" << Trajectories[walker].R.at(step).x[0] << "\t" << Trajectories[walker].R.at(step).x[1] << "\t" << Trajectories[walker].R.at(step).x[2] << endl;
1507 }
1508 }
1509 return true;
1510 } else
1511 return false;
1512};
1513
1514/** Prints molecule to *out as xyz file.
1515* \param *out output stream
1516 */
1517bool molecule::OutputXYZ(ofstream *out) const
1518{
1519 atom *walker = NULL;
1520 int AtomNo = 0, ElementNo;
1521 time_t now;
1522 element *runner = NULL;
1523
1524 now = time((time_t *)NULL); // Get the system time and put it into 'now' as 'calender time'
1525 walker = start;
1526 while (walker->next != end) { // go through every atom and count
1527 walker = walker->next;
1528 AtomNo++;
1529 }
1530 if (out != NULL) {
1531 *out << AtomNo << "\n\tCreated by molecuilder on " << ctime(&now);
1532 ElementNo = 0;
1533 runner = elemente->start;
1534 while (runner->next != elemente->end) { // go through every element
1535 runner = runner->next;
1536 if (ElementsInMolecule[runner->Z]) { // if this element got atoms
1537 ElementNo++;
1538 walker = start;
1539 while (walker->next != end) { // go through every atom of this element
1540 walker = walker->next;
1541 if (walker->type == runner) { // if this atom fits to element
1542 walker->OutputXYZLine(out);
1543 }
1544 }
1545 }
1546 }
1547 return true;
1548 } else
1549 return false;
1550};
1551
1552/** Brings molecule::AtomCount and atom::*Name up-to-date.
1553 * \param *out output stream for debugging
1554 */
1555void molecule::CountAtoms(ofstream *out)
1556{
1557 int i = 0;
1558 atom *Walker = start;
1559 while (Walker->next != end) {
1560 Walker = Walker->next;
1561 i++;
1562 }
1563 if ((AtomCount == 0) || (i != AtomCount)) {
1564 *out << Verbose(3) << "Mismatch in AtomCount " << AtomCount << " and recounted number " << i << ", renaming all." << endl;
1565 AtomCount = i;
1566
1567 // count NonHydrogen atoms and give each atom a unique name
1568 if (AtomCount != 0) {
1569 i=0;
1570 NoNonHydrogen = 0;
1571 Walker = start;
1572 while (Walker->next != end) {
1573 Walker = Walker->next;
1574 Walker->nr = i; // update number in molecule (for easier referencing in FragmentMolecule lateron)
1575 if (Walker->type->Z != 1) // count non-hydrogen atoms whilst at it
1576 NoNonHydrogen++;
1577 Free((void **)&Walker->Name, "molecule::CountAtoms: *walker->Name");
1578 Walker->Name = (char *) Malloc(sizeof(char)*6, "molecule::CountAtoms: *walker->Name");
1579 sprintf(Walker->Name, "%2s%02d", Walker->type->symbol, Walker->nr+1);
1580 *out << "Naming atom nr. " << Walker->nr << " " << Walker->Name << "." << endl;
1581 i++;
1582 }
1583 } else
1584 *out << Verbose(3) << "AtomCount is still " << AtomCount << ", thus counting nothing." << endl;
1585 }
1586};
1587
1588/** Brings molecule::ElementCount and molecule::ElementsInMolecule up-to-date.
1589 */
1590void molecule::CountElements()
1591{
1592 int i = 0;
1593 for(i=MAX_ELEMENTS;i--;)
1594 ElementsInMolecule[i] = 0;
1595 ElementCount = 0;
1596
1597 atom *walker = start;
1598 while (walker->next != end) {
1599 walker = walker->next;
1600 ElementsInMolecule[walker->type->Z]++;
1601 i++;
1602 }
1603 for(i=MAX_ELEMENTS;i--;)
1604 ElementCount += (ElementsInMolecule[i] != 0 ? 1 : 0);
1605};
1606
1607/** Counts all cyclic bonds and returns their number.
1608 * \note Hydrogen bonds can never by cyclic, thus no check for that
1609 * \param *out output stream for debugging
1610 * \return number opf cyclic bonds
1611 */
1612int molecule::CountCyclicBonds(ofstream *out)
1613{
1614 int No = 0;
1615 int *MinimumRingSize = NULL;
1616 MoleculeLeafClass *Subgraphs = NULL;
1617 class StackClass<bond *> *BackEdgeStack = NULL;
1618 bond *Binder = first;
1619 if ((Binder->next != last) && (Binder->next->Type == Undetermined)) {
1620 *out << Verbose(0) << "No Depth-First-Search analysis performed so far, calling ..." << endl;
1621 Subgraphs = DepthFirstSearchAnalysis(out, BackEdgeStack);
1622 while (Subgraphs->next != NULL) {
1623 Subgraphs = Subgraphs->next;
1624 delete(Subgraphs->previous);
1625 }
1626 delete(Subgraphs);
1627 delete[](MinimumRingSize);
1628 }
1629 while(Binder->next != last) {
1630 Binder = Binder->next;
1631 if (Binder->Cyclic)
1632 No++;
1633 }
1634 delete(BackEdgeStack);
1635 return No;
1636};
1637/** Returns Shading as a char string.
1638 * \param color the Shading
1639 * \return string of the flag
1640 */
1641string molecule::GetColor(enum Shading color)
1642{
1643 switch(color) {
1644 case white:
1645 return "white";
1646 break;
1647 case lightgray:
1648 return "lightgray";
1649 break;
1650 case darkgray:
1651 return "darkgray";
1652 break;
1653 case black:
1654 return "black";
1655 break;
1656 default:
1657 return "uncolored";
1658 break;
1659 };
1660};
1661
1662
1663/** Counts necessary number of valence electrons and returns number and SpinType.
1664 * \param configuration containing everything
1665 */
1666void molecule::CalculateOrbitals(class config &configuration)
1667{
1668 configuration.MaxPsiDouble = configuration.PsiMaxNoDown = configuration.PsiMaxNoUp = configuration.PsiType = 0;
1669 for(int i=MAX_ELEMENTS;i--;) {
1670 if (ElementsInMolecule[i] != 0) {
1671 //cout << "CalculateOrbitals: " << elemente->FindElement(i)->name << " has a valence of " << (int)elemente->FindElement(i)->Valence << " and there are " << ElementsInMolecule[i] << " of it." << endl;
1672 configuration.MaxPsiDouble += ElementsInMolecule[i]*((int)elemente->FindElement(i)->Valence);
1673 }
1674 }
1675 configuration.PsiMaxNoDown = configuration.MaxPsiDouble/2 + (configuration.MaxPsiDouble % 2);
1676 configuration.PsiMaxNoUp = configuration.MaxPsiDouble/2;
1677 configuration.MaxPsiDouble /= 2;
1678 configuration.PsiType = (configuration.PsiMaxNoDown == configuration.PsiMaxNoUp) ? 0 : 1;
1679 if ((configuration.PsiType == 1) && (configuration.ProcPEPsi < 2)) {
1680 configuration.ProcPEGamma /= 2;
1681 configuration.ProcPEPsi *= 2;
1682 } else {
1683 configuration.ProcPEGamma *= configuration.ProcPEPsi;
1684 configuration.ProcPEPsi = 1;
1685 }
1686 configuration.InitMaxMinStopStep = configuration.MaxMinStopStep = configuration.MaxPsiDouble;
1687};
1688
1689/** Creates an adjacency list of the molecule.
1690 * We obtain an outside file with the indices of atoms which are bondmembers.
1691 */
1692void molecule::CreateAdjacencyList2(ofstream *out, ifstream *input)
1693{
1694
1695 // 1 We will parse bonds out of the dbond file created by tremolo.
1696 int atom1, atom2, temp;
1697 atom *Walker, *OtherWalker;
1698
1699 if (!input)
1700 {
1701 cout << Verbose(1) << "Opening silica failed \n";
1702 };
1703
1704 *input >> ws >> atom1;
1705 *input >> ws >> atom2;
1706 cout << Verbose(1) << "Scanning file\n";
1707 while (!input->eof()) // Check whether we read everything already
1708 {
1709 *input >> ws >> atom1;
1710 *input >> ws >> atom2;
1711 if(atom2<atom1) //Sort indices of atoms in order
1712 {
1713 temp=atom1;
1714 atom1=atom2;
1715 atom2=temp;
1716 };
1717
1718 Walker=start;
1719 while(Walker-> nr != atom1) // Find atom corresponding to first index
1720 {
1721 Walker = Walker->next;
1722 };
1723 OtherWalker = Walker->next;
1724 while(OtherWalker->nr != atom2) // Find atom corresponding to second index
1725 {
1726 OtherWalker= OtherWalker->next;
1727 };
1728 AddBond(Walker, OtherWalker); //Add the bond between the two atoms with respective indices.
1729
1730 }
1731
1732 CreateListOfBondsPerAtom(out);
1733
1734};
1735
1736
1737/** Creates an adjacency list of the molecule.
1738 * Generally, we use the CSD approach to bond recognition, that is the the distance
1739 * between two atoms A and B must be within [Rcov(A)+Rcov(B)-t,Rcov(A)+Rcov(B)+t] with
1740 * a threshold t = 0.4 Angstroem.
1741 * To make it O(N log N) the function uses the linked-cell technique as follows:
1742 * The procedure is step-wise:
1743 * -# Remove every bond in list
1744 * -# Count the atoms in the molecule with CountAtoms()
1745 * -# partition cell into smaller linked cells of size \a bonddistance
1746 * -# put each atom into its corresponding cell
1747 * -# go through every cell, check the atoms therein against all possible bond partners in the 27 adjacent cells, add bond if true
1748 * -# create the list of bonds via CreateListOfBondsPerAtom()
1749 * -# correct the bond degree iteratively (single->double->triple bond)
1750 * -# finally print the bond list to \a *out if desired
1751 * \param *out out stream for printing the matrix, NULL if no output
1752 * \param bonddistance length of linked cells (i.e. maximum minimal length checked)
1753 * \param IsAngstroem whether coordinate system is gauged to Angstroem or Bohr radii
1754 */
1755void molecule::CreateAdjacencyList(ofstream *out, double bonddistance, bool IsAngstroem)
1756{
1757
1758 atom *Walker = NULL, *OtherWalker = NULL, *Candidate = NULL;
1759 int No, NoBonds, CandidateBondNo;
1760 int NumberCells, divisor[NDIM], n[NDIM], N[NDIM], index, Index, j;
1761 molecule **CellList;
1762 double distance, MinDistance, MaxDistance;
1763 double *matrix = ReturnFullMatrixforSymmetric(cell_size);
1764 Vector x;
1765 int FalseBondDegree = 0;
1766
1767 BondDistance = bonddistance; // * ((IsAngstroem) ? 1. : 1./AtomicLengthToAngstroem);
1768 *out << Verbose(0) << "Begin of CreateAdjacencyList." << endl;
1769 // remove every bond from the list
1770 if ((first->next != last) && (last->previous != first)) { // there are bonds present
1771 cleanup(first,last);
1772 }
1773
1774 // count atoms in molecule = dimension of matrix (also give each unique name and continuous numbering)
1775 CountAtoms(out);
1776 *out << Verbose(1) << "AtomCount " << AtomCount << "." << endl;
1777
1778 if (AtomCount != 0) {
1779 // 1. find divisor for each axis, such that a sphere with radius of at least bonddistance can be placed into each cell
1780 j=-1;
1781 for (int i=0;i<NDIM;i++) {
1782 j += i+1;
1783 divisor[i] = (int)floor(cell_size[j]/bonddistance); // take smaller value such that size of linked cell is at least bonddistance
1784 //*out << Verbose(1) << "divisor[" << i << "] = " << divisor[i] << "." << endl;
1785 }
1786 // 2a. allocate memory for the cell list
1787 NumberCells = divisor[0]*divisor[1]*divisor[2];
1788 *out << Verbose(1) << "Allocating " << NumberCells << " cells." << endl;
1789 CellList = (molecule **) Malloc(sizeof(molecule *)*NumberCells, "molecule::CreateAdjacencyList - ** CellList");
1790 for (int i=NumberCells;i--;)
1791 CellList[i] = NULL;
1792
1793 // 2b. put all atoms into its corresponding list
1794 Walker = start;
1795 while(Walker->next != end) {
1796 Walker = Walker->next;
1797 //*out << Verbose(1) << "Current atom is " << *Walker << " with coordinates ";
1798 //Walker->x.Output(out);
1799 //*out << "." << endl;
1800 // compute the cell by the atom's coordinates
1801 j=-1;
1802 for (int i=0;i<NDIM;i++) {
1803 j += i+1;
1804 x.CopyVector(&(Walker->x));
1805 x.KeepPeriodic(out, matrix);
1806 n[i] = (int)floor(x.x[i]/cell_size[j]*(double)divisor[i]);
1807 }
1808 index = n[2] + (n[1] + n[0] * divisor[1]) * divisor[2];
1809 //*out << Verbose(1) << "Atom " << *Walker << " goes into cell number [" << n[0] << "," << n[1] << "," << n[2] << "] = " << index << "." << endl;
1810 // add copy atom to this cell
1811 if (CellList[index] == NULL) // allocate molecule if not done
1812 CellList[index] = new molecule(elemente);
1813 OtherWalker = CellList[index]->AddCopyAtom(Walker); // add a copy of walker to this atom, father will be walker for later reference
1814 //*out << Verbose(1) << "Copy Atom is " << *OtherWalker << "." << endl;
1815 }
1816 //for (int i=0;i<NumberCells;i++)
1817 //*out << Verbose(1) << "Cell number " << i << ": " << CellList[i] << "." << endl;
1818
1819
1820 // 3a. go through every cell
1821 for (N[0]=divisor[0];N[0]--;)
1822 for (N[1]=divisor[1];N[1]--;)
1823 for (N[2]=divisor[2];N[2]--;) {
1824 Index = N[2] + (N[1] + N[0] * divisor[1]) * divisor[2];
1825 if (CellList[Index] != NULL) { // if there atoms in this cell
1826 //*out << Verbose(1) << "Current cell is " << Index << "." << endl;
1827 // 3b. for every atom therein
1828 Walker = CellList[Index]->start;
1829 while (Walker->next != CellList[Index]->end) { // go through every atom
1830 Walker = Walker->next;
1831 //*out << Verbose(0) << "Current Atom is " << *Walker << "." << endl;
1832 // 3c. check for possible bond between each atom in this and every one in the 27 cells
1833 for (n[0]=-1;n[0]<=1;n[0]++)
1834 for (n[1]=-1;n[1]<=1;n[1]++)
1835 for (n[2]=-1;n[2]<=1;n[2]++) {
1836 // compute the index of this comparison cell and make it periodic
1837 index = ((N[2]+n[2]+divisor[2])%divisor[2]) + (((N[1]+n[1]+divisor[1])%divisor[1]) + ((N[0]+n[0]+divisor[0])%divisor[0]) * divisor[1]) * divisor[2];
1838 //*out << Verbose(1) << "Number of comparison cell is " << index << "." << endl;
1839 if (CellList[index] != NULL) { // if there are any atoms in this cell
1840 OtherWalker = CellList[index]->start;
1841 while(OtherWalker->next != CellList[index]->end) { // go through every atom in this cell
1842 OtherWalker = OtherWalker->next;
1843 //*out << Verbose(0) << "Current comparison atom is " << *OtherWalker << "." << endl;
1844 /// \todo periodic check is missing here!
1845 //*out << Verbose(1) << "Checking distance " << OtherWalker->x.PeriodicDistance(&(Walker->x), cell_size) << " against typical bond length of " << bonddistance*bonddistance << "." << endl;
1846 MinDistance = OtherWalker->type->CovalentRadius + Walker->type->CovalentRadius;
1847 MinDistance *= (IsAngstroem) ? 1. : 1./AtomicLengthToAngstroem;
1848 MaxDistance = MinDistance + BONDTHRESHOLD;
1849 MinDistance -= BONDTHRESHOLD;
1850 distance = OtherWalker->x.PeriodicDistance(&(Walker->x), cell_size);
1851 if ((OtherWalker->father->nr > Walker->father->nr) && (distance <= MaxDistance*MaxDistance) && (distance >= MinDistance*MinDistance)) { // create bond if distance is smaller
1852 //*out << Verbose(0) << "Adding Bond between " << *Walker << " and " << *OtherWalker << "." << endl;
1853 AddBond(Walker->father, OtherWalker->father, 1); // also increases molecule::BondCount
1854 BondCount++;
1855 } else {
1856 //*out << Verbose(1) << "Not Adding: Wrong label order or distance too great." << endl;
1857 }
1858 }
1859 }
1860 }
1861 }
1862 }
1863 }
1864
1865
1866
1867 // 4. free the cell again
1868 for (int i=NumberCells;i--;)
1869 if (CellList[i] != NULL) {
1870 delete(CellList[i]);
1871 }
1872 Free((void **)&CellList, "molecule::CreateAdjacencyList - ** CellList");
1873
1874 // create the adjacency list per atom
1875 CreateListOfBondsPerAtom(out);
1876
1877 // correct Bond degree of each bond by checking both bond partners for a mismatch between valence and current sum of bond degrees,
1878 // iteratively increase the one first where the other bond partner has the fewest number of bonds (i.e. in general bonds oxygene
1879 // preferred over carbon bonds). Beforehand, we had picked the first mismatching partner, which lead to oxygenes with single instead of
1880 // double bonds as was expected.
1881 if (BondCount != 0) {
1882 NoCyclicBonds = 0;
1883 *out << Verbose(1) << "Correcting Bond degree of each bond ... ";
1884 do {
1885 No = 0; // No acts as breakup flag (if 1 we still continue)
1886 Walker = start;
1887 while (Walker->next != end) { // go through every atom
1888 Walker = Walker->next;
1889 // count valence of first partner
1890 NoBonds = 0;
1891 for(j=0;j<NumberOfBondsPerAtom[Walker->nr];j++)
1892 NoBonds += ListOfBondsPerAtom[Walker->nr][j]->BondDegree;
1893 *out << Verbose(3) << "Walker " << *Walker << ": " << (int)Walker->type->NoValenceOrbitals << " > " << NoBonds << "?" << endl;
1894 if ((int)(Walker->type->NoValenceOrbitals) > NoBonds) { // we have a mismatch, check all bonding partners for mismatch
1895 Candidate = NULL;
1896 CandidateBondNo = -1;
1897 for(int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++) { // go through each of its bond partners
1898 OtherWalker = ListOfBondsPerAtom[Walker->nr][i]->GetOtherAtom(Walker);
1899 // count valence of second partner
1900 NoBonds = 0;
1901 for(j=0;j<NumberOfBondsPerAtom[OtherWalker->nr];j++)
1902 NoBonds += ListOfBondsPerAtom[OtherWalker->nr][j]->BondDegree;
1903 *out << Verbose(3) << "OtherWalker " << *OtherWalker << ": " << (int)OtherWalker->type->NoValenceOrbitals << " > " << NoBonds << "?" << endl;
1904 if ((int)(OtherWalker->type->NoValenceOrbitals) > NoBonds) { // check if possible candidate
1905 if ((Candidate == NULL) || (NumberOfBondsPerAtom[Candidate->nr] > NumberOfBondsPerAtom[OtherWalker->nr])) { // pick the one with fewer number of bonds first
1906 Candidate = OtherWalker;
1907 CandidateBondNo = i;
1908 *out << Verbose(3) << "New candidate is " << *Candidate << "." << endl;
1909 }
1910 }
1911 }
1912 if ((Candidate != NULL) && (CandidateBondNo != -1)) {
1913 ListOfBondsPerAtom[Walker->nr][CandidateBondNo]->BondDegree++;
1914 *out << Verbose(2) << "Increased bond degree for bond " << *ListOfBondsPerAtom[Walker->nr][CandidateBondNo] << "." << endl;
1915 } else
1916 *out << Verbose(2) << "Could not find correct degree for atom " << *Walker << "." << endl;
1917 FalseBondDegree++;
1918 }
1919 }
1920 } while (No);
1921 *out << " done." << endl;
1922 } else
1923 *out << Verbose(1) << "BondCount is " << BondCount << ", no bonds between any of the " << AtomCount << " atoms." << endl;
1924 *out << Verbose(1) << "I detected " << BondCount << " bonds in the molecule with distance " << bonddistance << ", " << FalseBondDegree << " bonds could not be corrected." << endl;
1925
1926 // output bonds for debugging (if bond chain list was correctly installed)
1927 *out << Verbose(1) << endl << "From contents of bond chain list:";
1928 bond *Binder = first;
1929 while(Binder->next != last) {
1930 Binder = Binder->next;
1931 *out << *Binder << "\t" << endl;
1932 }
1933 *out << endl;
1934 } else
1935 *out << Verbose(1) << "AtomCount is " << AtomCount << ", thus no bonds, no connections!." << endl;
1936 *out << Verbose(0) << "End of CreateAdjacencyList." << endl;
1937 Free((void **)&matrix, "molecule::CreateAdjacencyList: *matrix");
1938
1939};
1940
1941
1942
1943/** Performs a Depth-First search on this molecule.
1944 * Marks bonds in molecule as cyclic, bridge, ... and atoms as
1945 * articulations points, ...
1946 * We use the algorithm from [Even, Graph Algorithms, p.62].
1947 * \param *out output stream for debugging
1948 * \param *&BackEdgeStack NULL pointer to StackClass with all the found back edges, allocated and filled on return
1949 * \return list of each disconnected subgraph as an individual molecule class structure
1950 */
1951MoleculeLeafClass * molecule::DepthFirstSearchAnalysis(ofstream *out, class StackClass<bond *> *&BackEdgeStack)
1952{
1953 class StackClass<atom *> *AtomStack = new StackClass<atom *>(AtomCount);
1954 BackEdgeStack = new StackClass<bond *> (BondCount);
1955 MoleculeLeafClass *SubGraphs = new MoleculeLeafClass(NULL);
1956 MoleculeLeafClass *LeafWalker = SubGraphs;
1957 int CurrentGraphNr = 0, OldGraphNr;
1958 int ComponentNumber = 0;
1959 atom *Walker = NULL, *OtherAtom = NULL, *Root = start->next;
1960 bond *Binder = NULL;
1961 bool BackStepping = false;
1962
1963 *out << Verbose(0) << "Begin of DepthFirstSearchAnalysis" << endl;
1964
1965 ResetAllBondsToUnused();
1966 ResetAllAtomNumbers();
1967 InitComponentNumbers();
1968 BackEdgeStack->ClearStack();
1969 while (Root != end) { // if there any atoms at all
1970 // (1) mark all edges unused, empty stack, set atom->GraphNr = 0 for all
1971 AtomStack->ClearStack();
1972
1973 // put into new subgraph molecule and add this to list of subgraphs
1974 LeafWalker = new MoleculeLeafClass(LeafWalker);
1975 LeafWalker->Leaf = new molecule(elemente);
1976 LeafWalker->Leaf->AddCopyAtom(Root);
1977
1978 OldGraphNr = CurrentGraphNr;
1979 Walker = Root;
1980 do { // (10)
1981 do { // (2) set number and Lowpoint of Atom to i, increase i, push current atom
1982 if (!BackStepping) { // if we don't just return from (8)
1983 Walker->GraphNr = CurrentGraphNr;
1984 Walker->LowpointNr = CurrentGraphNr;
1985 *out << Verbose(1) << "Setting Walker[" << Walker->Name << "]'s number to " << Walker->GraphNr << " with Lowpoint " << Walker->LowpointNr << "." << endl;
1986 AtomStack->Push(Walker);
1987 CurrentGraphNr++;
1988 }
1989 do { // (3) if Walker has no unused egdes, go to (5)
1990 BackStepping = false; // reset backstepping flag for (8)
1991 if (Binder == NULL) // if we don't just return from (11), Binder is already set to next unused
1992 Binder = FindNextUnused(Walker);
1993 if (Binder == NULL)
1994 break;
1995 *out << Verbose(2) << "Current Unused Bond is " << *Binder << "." << endl;
1996 // (4) Mark Binder used, ...
1997 Binder->MarkUsed(black);
1998 OtherAtom = Binder->GetOtherAtom(Walker);
1999 *out << Verbose(2) << "(4) OtherAtom is " << OtherAtom->Name << "." << endl;
2000 if (OtherAtom->GraphNr != -1) {
2001 // (4a) ... if "other" atom has been visited (GraphNr != 0), set lowpoint to minimum of both, go to (3)
2002 Binder->Type = BackEdge;
2003 BackEdgeStack->Push(Binder);
2004 Walker->LowpointNr = ( Walker->LowpointNr < OtherAtom->GraphNr ) ? Walker->LowpointNr : OtherAtom->GraphNr;
2005 *out << Verbose(3) << "(4a) Visited: Setting Lowpoint of Walker[" << Walker->Name << "] to " << Walker->LowpointNr << "." << endl;
2006 } else {
2007 // (4b) ... otherwise set OtherAtom as Ancestor of Walker and Walker as OtherAtom, go to (2)
2008 Binder->Type = TreeEdge;
2009 OtherAtom->Ancestor = Walker;
2010 Walker = OtherAtom;
2011 *out << Verbose(3) << "(4b) Not Visited: OtherAtom[" << OtherAtom->Name << "]'s Ancestor is now " << OtherAtom->Ancestor->Name << ", Walker is OtherAtom " << OtherAtom->Name << "." << endl;
2012 break;
2013 }
2014 Binder = NULL;
2015 } while (1); // (3)
2016 if (Binder == NULL) {
2017 *out << Verbose(2) << "No more Unused Bonds." << endl;
2018 break;
2019 } else
2020 Binder = NULL;
2021 } while (1); // (2)
2022
2023 // if we came from backstepping, yet there were no more unused bonds, we end up here with no Ancestor, because Walker is Root! Then we are finished!
2024 if ((Walker == Root) && (Binder == NULL))
2025 break;
2026
2027 // (5) if Ancestor of Walker is ...
2028 *out << Verbose(1) << "(5) Number of Walker[" << Walker->Name << "]'s Ancestor[" << Walker->Ancestor->Name << "] is " << Walker->Ancestor->GraphNr << "." << endl;
2029 if (Walker->Ancestor->GraphNr != Root->GraphNr) {
2030 // (6) (Ancestor of Walker is not Root)
2031 if (Walker->LowpointNr < Walker->Ancestor->GraphNr) {
2032 // (6a) set Ancestor's Lowpoint number to minimum of of its Ancestor and itself, go to Step(8)
2033 Walker->Ancestor->LowpointNr = (Walker->Ancestor->LowpointNr < Walker->LowpointNr) ? Walker->Ancestor->LowpointNr : Walker->LowpointNr;
2034 *out << Verbose(2) << "(6) Setting Walker[" << Walker->Name << "]'s Ancestor[" << Walker->Ancestor->Name << "]'s Lowpoint to " << Walker->Ancestor->LowpointNr << "." << endl;
2035 } else {
2036 // (7) (Ancestor of Walker is a separating vertex, remove all from stack till Walker (including), these and Ancestor form a component
2037 Walker->Ancestor->SeparationVertex = true;
2038 *out << Verbose(2) << "(7) Walker[" << Walker->Name << "]'s Ancestor[" << Walker->Ancestor->Name << "]'s is a separating vertex, creating component." << endl;
2039 SetNextComponentNumber(Walker->Ancestor, ComponentNumber);
2040 *out << Verbose(3) << "(7) Walker[" << Walker->Name << "]'s Ancestor's Compont is " << ComponentNumber << "." << endl;
2041 SetNextComponentNumber(Walker, ComponentNumber);
2042 *out << Verbose(3) << "(7) Walker[" << Walker->Name << "]'s Compont is " << ComponentNumber << "." << endl;
2043 do {
2044 OtherAtom = AtomStack->PopLast();
2045 LeafWalker->Leaf->AddCopyAtom(OtherAtom);
2046 SetNextComponentNumber(OtherAtom, ComponentNumber);
2047 *out << Verbose(3) << "(7) Other[" << OtherAtom->Name << "]'s Compont is " << ComponentNumber << "." << endl;
2048 } while (OtherAtom != Walker);
2049 ComponentNumber++;
2050 }
2051 // (8) Walker becomes its Ancestor, go to (3)
2052 *out << Verbose(2) << "(8) Walker[" << Walker->Name << "] is now its Ancestor " << Walker->Ancestor->Name << ", backstepping. " << endl;
2053 Walker = Walker->Ancestor;
2054 BackStepping = true;
2055 }
2056 if (!BackStepping) { // coming from (8) want to go to (3)
2057 // (9) remove all from stack till Walker (including), these and Root form a component
2058 AtomStack->Output(out);
2059 SetNextComponentNumber(Root, ComponentNumber);
2060 *out << Verbose(3) << "(9) Root[" << Root->Name << "]'s Component is " << ComponentNumber << "." << endl;
2061 SetNextComponentNumber(Walker, ComponentNumber);
2062 *out << Verbose(3) << "(9) Walker[" << Walker->Name << "]'s Component is " << ComponentNumber << "." << endl;
2063 do {
2064 OtherAtom = AtomStack->PopLast();
2065 LeafWalker->Leaf->AddCopyAtom(OtherAtom);
2066 SetNextComponentNumber(OtherAtom, ComponentNumber);
2067 *out << Verbose(3) << "(7) Other[" << OtherAtom->Name << "]'s Compont is " << ComponentNumber << "." << endl;
2068 } while (OtherAtom != Walker);
2069 ComponentNumber++;
2070
2071 // (11) Root is separation vertex, set Walker to Root and go to (4)
2072 Walker = Root;
2073 Binder = FindNextUnused(Walker);
2074 *out << Verbose(1) << "(10) Walker is Root[" << Root->Name << "], next Unused Bond is " << Binder << "." << endl;
2075 if (Binder != NULL) { // Root is separation vertex
2076 *out << Verbose(1) << "(11) Root is a separation vertex." << endl;
2077 Walker->SeparationVertex = true;
2078 }
2079 }
2080 } while ((BackStepping) || (Binder != NULL)); // (10) halt only if Root has no unused edges
2081
2082 // From OldGraphNr to CurrentGraphNr ranges an disconnected subgraph
2083 *out << Verbose(0) << "Disconnected subgraph ranges from " << OldGraphNr << " to " << CurrentGraphNr << "." << endl;
2084 LeafWalker->Leaf->Output(out);
2085 *out << endl;
2086
2087 // step on to next root
2088 while ((Root != end) && (Root->GraphNr != -1)) {
2089 //*out << Verbose(1) << "Current next subgraph root candidate is " << Root->Name << "." << endl;
2090 if (Root->GraphNr != -1) // if already discovered, step on
2091 Root = Root->next;
2092 }
2093 }
2094 // set cyclic bond criterium on "same LP" basis
2095 Binder = first;
2096 while(Binder->next != last) {
2097 Binder = Binder->next;
2098 if (Binder->rightatom->LowpointNr == Binder->leftatom->LowpointNr) { // cyclic ??
2099 Binder->Cyclic = true;
2100 NoCyclicBonds++;
2101 }
2102 }
2103
2104
2105 *out << Verbose(1) << "Final graph info for each atom is:" << endl;
2106 Walker = start;
2107 while (Walker->next != end) {
2108 Walker = Walker->next;
2109 *out << Verbose(2) << "Atom " << Walker->Name << " is " << ((Walker->SeparationVertex) ? "a" : "not a") << " separation vertex, components are ";
2110 OutputComponentNumber(out, Walker);
2111 *out << " with Lowpoint " << Walker->LowpointNr << " and Graph Nr. " << Walker->GraphNr << "." << endl;
2112 }
2113
2114 *out << Verbose(1) << "Final graph info for each bond is:" << endl;
2115 Binder = first;
2116 while(Binder->next != last) {
2117 Binder = Binder->next;
2118 *out << Verbose(2) << ((Binder->Type == TreeEdge) ? "TreeEdge " : "BackEdge ") << *Binder << ": <";
2119 *out << ((Binder->leftatom->SeparationVertex) ? "SP," : "") << "L" << Binder->leftatom->LowpointNr << " G" << Binder->leftatom->GraphNr << " Comp.";
2120 OutputComponentNumber(out, Binder->leftatom);
2121 *out << " === ";
2122 *out << ((Binder->rightatom->SeparationVertex) ? "SP," : "") << "L" << Binder->rightatom->LowpointNr << " G" << Binder->rightatom->GraphNr << " Comp.";
2123 OutputComponentNumber(out, Binder->rightatom);
2124 *out << ">." << endl;
2125 if (Binder->Cyclic) // cyclic ??
2126 *out << Verbose(3) << "Lowpoint at each side are equal: CYCLIC!" << endl;
2127 }
2128
2129 // free all and exit
2130 delete(AtomStack);
2131 *out << Verbose(0) << "End of DepthFirstSearchAnalysis" << endl;
2132 return SubGraphs;
2133};
2134
2135/** Analyses the cycles found and returns minimum of all cycle lengths.
2136 * We begin with a list of Back edges found during DepthFirstSearchAnalysis(). We go through this list - one end is the Root,
2137 * the other our initial Walker - and do a Breadth First Search for the Root. We mark down each Predecessor and as soon as
2138 * we have found the Root via BFS, we may climb back the closed cycle via the Predecessors. Thereby we mark atoms and bonds
2139 * as cyclic and print out the cycles.
2140 * \param *out output stream for debugging
2141 * \param *BackEdgeStack stack with all back edges found during DFS scan. Beware: This stack contains the bonds from the total molecule, not from the subgraph!
2142 * \param *&MinimumRingSize contains smallest ring size in molecular structure on return or -1 if no rings were found, if set is maximum search distance
2143 * \todo BFS from the not-same-LP to find back to starting point of tributary cycle over more than one bond
2144 */
2145void molecule::CyclicStructureAnalysis(ofstream *out, class StackClass<bond *> * BackEdgeStack, int *&MinimumRingSize)
2146{
2147 atom **PredecessorList = (atom **) Malloc(sizeof(atom *)*AtomCount, "molecule::CyclicStructureAnalysis: **PredecessorList");
2148 int *ShortestPathList = (int *) Malloc(sizeof(int)*AtomCount, "molecule::CyclicStructureAnalysis: *ShortestPathList");
2149 enum Shading *ColorList = (enum Shading *) Malloc(sizeof(enum Shading)*AtomCount, "molecule::CyclicStructureAnalysis: *ColorList");
2150 class StackClass<atom *> *BFSStack = new StackClass<atom *> (AtomCount); // will hold the current ring
2151 class StackClass<atom *> *TouchedStack = new StackClass<atom *> (AtomCount); // contains all "touched" atoms (that need to be reset after BFS loop)
2152 atom *Walker = NULL, *OtherAtom = NULL, *Root = NULL;
2153 bond *Binder = NULL, *BackEdge = NULL;
2154 int RingSize, NumCycles, MinRingSize = -1;
2155
2156 // initialise each vertex as white with no predecessor, empty queue, color Root lightgray
2157 for (int i=AtomCount;i--;) {
2158 PredecessorList[i] = NULL;
2159 ShortestPathList[i] = -1;
2160 ColorList[i] = white;
2161 }
2162
2163 *out << Verbose(1) << "Back edge list - ";
2164 BackEdgeStack->Output(out);
2165
2166 *out << Verbose(1) << "Analysing cycles ... " << endl;
2167 NumCycles = 0;
2168 while (!BackEdgeStack->IsEmpty()) {
2169 BackEdge = BackEdgeStack->PopFirst();
2170 // this is the target
2171 Root = BackEdge->leftatom;
2172 // this is the source point
2173 Walker = BackEdge->rightatom;
2174 ShortestPathList[Walker->nr] = 0;
2175 BFSStack->ClearStack(); // start with empty BFS stack
2176 BFSStack->Push(Walker);
2177 TouchedStack->Push(Walker);
2178 *out << Verbose(1) << "---------------------------------------------------------------------------------------------------------" << endl;
2179 OtherAtom = NULL;
2180 do { // look for Root
2181 Walker = BFSStack->PopFirst();
2182 *out << Verbose(2) << "Current Walker is " << *Walker << ", we look for SP to Root " << *Root << "." << endl;
2183 for(int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++) {
2184 Binder = ListOfBondsPerAtom[Walker->nr][i];
2185 if (Binder != BackEdge) { // only walk along DFS spanning tree (otherwise we always find SP of one being backedge Binder)
2186 OtherAtom = Binder->GetOtherAtom(Walker);
2187#ifdef ADDHYDROGEN
2188 if (OtherAtom->type->Z != 1) {
2189#endif
2190 *out << Verbose(2) << "Current OtherAtom is: " << OtherAtom->Name << " for bond " << *Binder << "." << endl;
2191 if (ColorList[OtherAtom->nr] == white) {
2192 TouchedStack->Push(OtherAtom);
2193 ColorList[OtherAtom->nr] = lightgray;
2194 PredecessorList[OtherAtom->nr] = Walker; // Walker is the predecessor
2195 ShortestPathList[OtherAtom->nr] = ShortestPathList[Walker->nr]+1;
2196 *out << Verbose(2) << "Coloring OtherAtom " << OtherAtom->Name << " lightgray, its predecessor is " << Walker->Name << " and its Shortest Path is " << ShortestPathList[OtherAtom->nr] << " egde(s) long." << endl;
2197 //if (ShortestPathList[OtherAtom->nr] < MinimumRingSize[Walker->GetTrueFather()->nr]) { // Check for maximum distance
2198 *out << Verbose(3) << "Putting OtherAtom into queue." << endl;
2199 BFSStack->Push(OtherAtom);
2200 //}
2201 } else {
2202 *out << Verbose(3) << "Not Adding, has already been visited." << endl;
2203 }
2204 if (OtherAtom == Root)
2205 break;
2206#ifdef ADDHYDROGEN
2207 } else {
2208 *out << Verbose(2) << "Skipping hydrogen atom " << *OtherAtom << "." << endl;
2209 ColorList[OtherAtom->nr] = black;
2210 }
2211#endif
2212 } else {
2213 *out << Verbose(2) << "Bond " << *Binder << " not Visiting, is the back edge." << endl;
2214 }
2215 }
2216 ColorList[Walker->nr] = black;
2217 *out << Verbose(1) << "Coloring Walker " << Walker->Name << " black." << endl;
2218 if (OtherAtom == Root) { // if we have found the root, check whether this cycle wasn't already found beforehand
2219 // step through predecessor list
2220 while (OtherAtom != BackEdge->rightatom) {
2221 if (!OtherAtom->GetTrueFather()->IsCyclic) // if one bond in the loop is not marked as cyclic, we haven't found this cycle yet
2222 break;
2223 else
2224 OtherAtom = PredecessorList[OtherAtom->nr];
2225 }
2226 if (OtherAtom == BackEdge->rightatom) { // if each atom in found cycle is cyclic, loop's been found before already
2227 *out << Verbose(3) << "This cycle was already found before, skipping and removing seeker from search." << endl;\
2228 do {
2229 OtherAtom = TouchedStack->PopLast();
2230 if (PredecessorList[OtherAtom->nr] == Walker) {
2231 *out << Verbose(4) << "Removing " << *OtherAtom << " from lists and stacks." << endl;
2232 PredecessorList[OtherAtom->nr] = NULL;
2233 ShortestPathList[OtherAtom->nr] = -1;
2234 ColorList[OtherAtom->nr] = white;
2235 BFSStack->RemoveItem(OtherAtom);
2236 }
2237 } while ((!TouchedStack->IsEmpty()) && (PredecessorList[OtherAtom->nr] == NULL));
2238 TouchedStack->Push(OtherAtom); // last was wrongly popped
2239 OtherAtom = BackEdge->rightatom; // set to not Root
2240 } else
2241 OtherAtom = Root;
2242 }
2243 } while ((!BFSStack->IsEmpty()) && (OtherAtom != Root) && (OtherAtom != NULL)); // || (ShortestPathList[OtherAtom->nr] < MinimumRingSize[Walker->GetTrueFather()->nr])));
2244
2245 if (OtherAtom == Root) {
2246 // now climb back the predecessor list and thus find the cycle members
2247 NumCycles++;
2248 RingSize = 1;
2249 Root->GetTrueFather()->IsCyclic = true;
2250 *out << Verbose(1) << "Found ring contains: ";
2251 Walker = Root;
2252 while (Walker != BackEdge->rightatom) {
2253 *out << Walker->Name << " <-> ";
2254 Walker = PredecessorList[Walker->nr];
2255 Walker->GetTrueFather()->IsCyclic = true;
2256 RingSize++;
2257 }
2258 *out << Walker->Name << " with a length of " << RingSize << "." << endl << endl;
2259 // walk through all and set MinimumRingSize
2260 Walker = Root;
2261 MinimumRingSize[Walker->GetTrueFather()->nr] = RingSize;
2262 while (Walker != BackEdge->rightatom) {
2263 Walker = PredecessorList[Walker->nr];
2264 if (RingSize < MinimumRingSize[Walker->GetTrueFather()->nr])
2265 MinimumRingSize[Walker->GetTrueFather()->nr] = RingSize;
2266 }
2267 if ((RingSize < MinRingSize) || (MinRingSize == -1))
2268 MinRingSize = RingSize;
2269 } else {
2270 *out << Verbose(1) << "No ring containing " << *Root << " with length equal to or smaller than " << MinimumRingSize[Walker->GetTrueFather()->nr] << " found." << endl;
2271 }
2272
2273 // now clean the lists
2274 while (!TouchedStack->IsEmpty()){
2275 Walker = TouchedStack->PopFirst();
2276 PredecessorList[Walker->nr] = NULL;
2277 ShortestPathList[Walker->nr] = -1;
2278 ColorList[Walker->nr] = white;
2279 }
2280 }
2281 if (MinRingSize != -1) {
2282 // go over all atoms
2283 Root = start;
2284 while(Root->next != end) {
2285 Root = Root->next;
2286
2287 if (MinimumRingSize[Root->GetTrueFather()->nr] == AtomCount) { // check whether MinimumRingSize is set, if not BFS to next where it is
2288 Walker = Root;
2289 ShortestPathList[Walker->nr] = 0;
2290 BFSStack->ClearStack(); // start with empty BFS stack
2291 BFSStack->Push(Walker);
2292 TouchedStack->Push(Walker);
2293 //*out << Verbose(1) << "---------------------------------------------------------------------------------------------------------" << endl;
2294 OtherAtom = Walker;
2295 while (OtherAtom != NULL) { // look for Root
2296 Walker = BFSStack->PopFirst();
2297 //*out << Verbose(2) << "Current Walker is " << *Walker << ", we look for SP to Root " << *Root << "." << endl;
2298 for(int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++) {
2299 Binder = ListOfBondsPerAtom[Walker->nr][i];
2300 if ((Binder != BackEdge) || (NumberOfBondsPerAtom[Walker->nr] == 1)) { // only walk along DFS spanning tree (otherwise we always find SP of 1 being backedge Binder), but terminal hydrogens may be connected via backedge, hence extra check
2301 OtherAtom = Binder->GetOtherAtom(Walker);
2302 //*out << Verbose(2) << "Current OtherAtom is: " << OtherAtom->Name << " for bond " << *Binder << "." << endl;
2303 if (ColorList[OtherAtom->nr] == white) {
2304 TouchedStack->Push(OtherAtom);
2305 ColorList[OtherAtom->nr] = lightgray;
2306 PredecessorList[OtherAtom->nr] = Walker; // Walker is the predecessor
2307 ShortestPathList[OtherAtom->nr] = ShortestPathList[Walker->nr]+1;
2308 //*out << Verbose(2) << "Coloring OtherAtom " << OtherAtom->Name << " lightgray, its predecessor is " << Walker->Name << " and its Shortest Path is " << ShortestPathList[OtherAtom->nr] << " egde(s) long." << endl;
2309 if (OtherAtom->GetTrueFather()->IsCyclic) { // if the other atom is connected to a ring
2310 MinimumRingSize[Root->GetTrueFather()->nr] = ShortestPathList[OtherAtom->nr]+MinimumRingSize[OtherAtom->GetTrueFather()->nr];
2311 OtherAtom = NULL; //break;
2312 break;
2313 } else
2314 BFSStack->Push(OtherAtom);
2315 } else {
2316 //*out << Verbose(3) << "Not Adding, has already been visited." << endl;
2317 }
2318 } else {
2319 //*out << Verbose(3) << "Not Visiting, is a back edge." << endl;
2320 }
2321 }
2322 ColorList[Walker->nr] = black;
2323 //*out << Verbose(1) << "Coloring Walker " << Walker->Name << " black." << endl;
2324 }
2325
2326 // now clean the lists
2327 while (!TouchedStack->IsEmpty()){
2328 Walker = TouchedStack->PopFirst();
2329 PredecessorList[Walker->nr] = NULL;
2330 ShortestPathList[Walker->nr] = -1;
2331 ColorList[Walker->nr] = white;
2332 }
2333 }
2334 *out << Verbose(1) << "Minimum ring size of " << *Root << " is " << MinimumRingSize[Root->GetTrueFather()->nr] << "." << endl;
2335 }
2336 *out << Verbose(1) << "Minimum ring size is " << MinRingSize << ", over " << NumCycles << " cycles total." << endl;
2337 } else
2338 *out << Verbose(1) << "No rings were detected in the molecular structure." << endl;
2339
2340 Free((void **)&PredecessorList, "molecule::CyclicStructureAnalysis: **PredecessorList");
2341 Free((void **)&ShortestPathList, "molecule::CyclicStructureAnalysis: **ShortestPathList");
2342 Free((void **)&ColorList, "molecule::CyclicStructureAnalysis: **ColorList");
2343 delete(BFSStack);
2344};
2345
2346/** Sets the next component number.
2347 * This is O(N) as the number of bonds per atom is bound.
2348 * \param *vertex atom whose next atom::*ComponentNr is to be set
2349 * \param nr number to use
2350 */
2351void molecule::SetNextComponentNumber(atom *vertex, int nr)
2352{
2353 int i=0;
2354 if (vertex != NULL) {
2355 for(;i<NumberOfBondsPerAtom[vertex->nr];i++) {
2356 if (vertex->ComponentNr[i] == -1) { // check if not yet used
2357 vertex->ComponentNr[i] = nr;
2358 break;
2359 }
2360 else if (vertex->ComponentNr[i] == nr) // if number is already present, don't add another time
2361 break; // breaking here will not cause error!
2362 }
2363 if (i == NumberOfBondsPerAtom[vertex->nr])
2364 cerr << "Error: All Component entries are already occupied!" << endl;
2365 } else
2366 cerr << "Error: Given vertex is NULL!" << endl;
2367};
2368
2369/** Output a list of flags, stating whether the bond was visited or not.
2370 * \param *out output stream for debugging
2371 */
2372void molecule::OutputComponentNumber(ofstream *out, atom *vertex)
2373{
2374 for(int i=0;i<NumberOfBondsPerAtom[vertex->nr];i++)
2375 *out << vertex->ComponentNr[i] << " ";
2376};
2377
2378/** Allocates memory for all atom::*ComponentNr in this molecule and sets each entry to -1.
2379 */
2380void molecule::InitComponentNumbers()
2381{
2382 atom *Walker = start;
2383 while(Walker->next != end) {
2384 Walker = Walker->next;
2385 if (Walker->ComponentNr != NULL)
2386 Free((void **)&Walker->ComponentNr, "molecule::InitComponentNumbers: **Walker->ComponentNr");
2387 Walker->ComponentNr = (int *) Malloc(sizeof(int)*NumberOfBondsPerAtom[Walker->nr], "molecule::InitComponentNumbers: *Walker->ComponentNr");
2388 for (int i=NumberOfBondsPerAtom[Walker->nr];i--;)
2389 Walker->ComponentNr[i] = -1;
2390 }
2391};
2392
2393/** Returns next unused bond for this atom \a *vertex or NULL of none exists.
2394 * \param *vertex atom to regard
2395 * \return bond class or NULL
2396 */
2397bond * molecule::FindNextUnused(atom *vertex)
2398{
2399 for(int i=0;i<NumberOfBondsPerAtom[vertex->nr];i++)
2400 if (ListOfBondsPerAtom[vertex->nr][i]->IsUsed() == white)
2401 return(ListOfBondsPerAtom[vertex->nr][i]);
2402 return NULL;
2403};
2404
2405/** Resets bond::Used flag of all bonds in this molecule.
2406 * \return true - success, false - -failure
2407 */
2408void molecule::ResetAllBondsToUnused()
2409{
2410 bond *Binder = first;
2411 while (Binder->next != last) {
2412 Binder = Binder->next;
2413 Binder->ResetUsed();
2414 }
2415};
2416
2417/** Resets atom::nr to -1 of all atoms in this molecule.
2418 */
2419void molecule::ResetAllAtomNumbers()
2420{
2421 atom *Walker = start;
2422 while (Walker->next != end) {
2423 Walker = Walker->next;
2424 Walker->GraphNr = -1;
2425 }
2426};
2427
2428/** Output a list of flags, stating whether the bond was visited or not.
2429 * \param *out output stream for debugging
2430 * \param *list
2431 */
2432void OutputAlreadyVisited(ofstream *out, int *list)
2433{
2434 *out << Verbose(4) << "Already Visited Bonds:\t";
2435 for(int i=1;i<=list[0];i++) *out << Verbose(0) << list[i] << " ";
2436 *out << endl;
2437};
2438
2439/** Estimates by educated guessing (using upper limit) the expected number of fragments.
2440 * The upper limit is
2441 * \f[
2442 * n = N \cdot C^k
2443 * \f]
2444 * where \f$C=2^c\f$ and c is the maximum bond degree over N number of atoms.
2445 * \param *out output stream for debugging
2446 * \param order bond order k
2447 * \return number n of fragments
2448 */
2449int molecule::GuesstimateFragmentCount(ofstream *out, int order)
2450{
2451 int c = 0;
2452 int FragmentCount;
2453 // get maximum bond degree
2454 atom *Walker = start;
2455 while (Walker->next != end) {
2456 Walker = Walker->next;
2457 c = (NumberOfBondsPerAtom[Walker->nr] > c) ? NumberOfBondsPerAtom[Walker->nr] : c;
2458 }
2459 FragmentCount = NoNonHydrogen*(1 << (c*order));
2460 *out << Verbose(1) << "Upper limit for this subgraph is " << FragmentCount << " for " << NoNonHydrogen << " non-H atoms with maximum bond degree of " << c << "." << endl;
2461 return FragmentCount;
2462};
2463
2464/** Scans a single line for number and puts them into \a KeySet.
2465 * \param *out output stream for debugging
2466 * \param *buffer buffer to scan
2467 * \param &CurrentSet filled KeySet on return
2468 * \return true - at least one valid atom id parsed, false - CurrentSet is empty
2469 */
2470bool molecule::ScanBufferIntoKeySet(ofstream *out, char *buffer, KeySet &CurrentSet)
2471{
2472 stringstream line;
2473 int AtomNr;
2474 int status = 0;
2475
2476 line.str(buffer);
2477 while (!line.eof()) {
2478 line >> AtomNr;
2479 if ((AtomNr >= 0) && (AtomNr < AtomCount)) {
2480 CurrentSet.insert(AtomNr); // insert at end, hence in same order as in file!
2481 status++;
2482 } // else it's "-1" or else and thus must not be added
2483 }
2484 *out << Verbose(1) << "The scanned KeySet is ";
2485 for(KeySet::iterator runner = CurrentSet.begin(); runner != CurrentSet.end(); runner++) {
2486 *out << (*runner) << "\t";
2487 }
2488 *out << endl;
2489 return (status != 0);
2490};
2491
2492/** Parses the KeySet file and fills \a *FragmentList from the known molecule structure.
2493 * Does two-pass scanning:
2494 * -# Scans the keyset file and initialises a temporary graph
2495 * -# Scans TEFactors file and sets the TEFactor of each key set in the temporary graph accordingly
2496 * Finally, the temporary graph is inserted into the given \a FragmentList for return.
2497 * \param *out output stream for debugging
2498 * \param *path path to file
2499 * \param *FragmentList empty, filled on return
2500 * \return true - parsing successfully, false - failure on parsing (FragmentList will be NULL)
2501 */
2502bool molecule::ParseKeySetFile(ofstream *out, char *path, Graph *&FragmentList)
2503{
2504 bool status = true;
2505 ifstream InputFile;
2506 stringstream line;
2507 GraphTestPair testGraphInsert;
2508 int NumberOfFragments = 0;
2509 double TEFactor;
2510 char *filename = (char *) Malloc(sizeof(char)*MAXSTRINGSIZE, "molecule::ParseKeySetFile - filename");
2511
2512 if (FragmentList == NULL) { // check list pointer
2513 FragmentList = new Graph;
2514 }
2515
2516 // 1st pass: open file and read
2517 *out << Verbose(1) << "Parsing the KeySet file ... " << endl;
2518 sprintf(filename, "%s/%s%s", path, FRAGMENTPREFIX, KEYSETFILE);
2519 InputFile.open(filename);
2520 if (InputFile != NULL) {
2521 // each line represents a new fragment
2522 char *buffer = (char *) Malloc(sizeof(char)*MAXSTRINGSIZE, "molecule::ParseKeySetFile - *buffer");
2523 // 1. parse keysets and insert into temp. graph
2524 while (!InputFile.eof()) {
2525 InputFile.getline(buffer, MAXSTRINGSIZE);
2526 KeySet CurrentSet;
2527 if ((strlen(buffer) > 0) && (ScanBufferIntoKeySet(out, buffer, CurrentSet))) { // if at least one valid atom was added, write config
2528 testGraphInsert = FragmentList->insert(GraphPair (CurrentSet,pair<int,double>(NumberOfFragments++,1))); // store fragment number and current factor
2529 if (!testGraphInsert.second) {
2530 cerr << "KeySet file must be corrupt as there are two equal key sets therein!" << endl;
2531 }
2532 //FragmentList->ListOfMolecules[NumberOfFragments++] = StoreFragmentFromKeySet(out, CurrentSet, IsAngstroem);
2533 }
2534 }
2535 // 2. Free and done
2536 InputFile.close();
2537 InputFile.clear();
2538 Free((void **)&buffer, "molecule::ParseKeySetFile - *buffer");
2539 *out << Verbose(1) << "done." << endl;
2540 } else {
2541 *out << Verbose(1) << "File " << filename << " not found." << endl;
2542 status = false;
2543 }
2544
2545 // 2nd pass: open TEFactors file and read
2546 *out << Verbose(1) << "Parsing the TEFactors file ... " << endl;
2547 sprintf(filename, "%s/%s%s", path, FRAGMENTPREFIX, TEFACTORSFILE);
2548 InputFile.open(filename);
2549 if (InputFile != NULL) {
2550 // 3. add found TEFactors to each keyset
2551 NumberOfFragments = 0;
2552 for(Graph::iterator runner = FragmentList->begin();runner != FragmentList->end(); runner++) {
2553 if (!InputFile.eof()) {
2554 InputFile >> TEFactor;
2555 (*runner).second.second = TEFactor;
2556 *out << Verbose(2) << "Setting " << ++NumberOfFragments << " fragment's TEFactor to " << (*runner).second.second << "." << endl;
2557 } else {
2558 status = false;
2559 break;
2560 }
2561 }
2562 // 4. Free and done
2563 InputFile.close();
2564 *out << Verbose(1) << "done." << endl;
2565 } else {
2566 *out << Verbose(1) << "File " << filename << " not found." << endl;
2567 status = false;
2568 }
2569
2570 // free memory
2571 Free((void **)&filename, "molecule::ParseKeySetFile - filename");
2572
2573 return status;
2574};
2575
2576/** Stores keysets and TEFactors to file.
2577 * \param *out output stream for debugging
2578 * \param KeySetList Graph with Keysets and factors
2579 * \param *path path to file
2580 * \return true - file written successfully, false - writing failed
2581 */
2582bool molecule::StoreKeySetFile(ofstream *out, Graph &KeySetList, char *path)
2583{
2584 ofstream output;
2585 bool status = true;
2586 string line;
2587
2588 // open KeySet file
2589 line = path;
2590 line.append("/");
2591 line += FRAGMENTPREFIX;
2592 line += KEYSETFILE;
2593 output.open(line.c_str(), ios::out);
2594 *out << Verbose(1) << "Saving key sets of the total graph ... ";
2595 if(output != NULL) {
2596 for(Graph::iterator runner = KeySetList.begin(); runner != KeySetList.end(); runner++) {
2597 for (KeySet::iterator sprinter = (*runner).first.begin();sprinter != (*runner).first.end(); sprinter++) {
2598 if (sprinter != (*runner).first.begin())
2599 output << "\t";
2600 output << *sprinter;
2601 }
2602 output << endl;
2603 }
2604 *out << "done." << endl;
2605 } else {
2606 cerr << "Unable to open " << line << " for writing keysets!" << endl;
2607 status = false;
2608 }
2609 output.close();
2610 output.clear();
2611
2612 // open TEFactors file
2613 line = path;
2614 line.append("/");
2615 line += FRAGMENTPREFIX;
2616 line += TEFACTORSFILE;
2617 output.open(line.c_str(), ios::out);
2618 *out << Verbose(1) << "Saving TEFactors of the total graph ... ";
2619 if(output != NULL) {
2620 for(Graph::iterator runner = KeySetList.begin(); runner != KeySetList.end(); runner++)
2621 output << (*runner).second.second << endl;
2622 *out << Verbose(1) << "done." << endl;
2623 } else {
2624 *out << Verbose(1) << "failed to open " << line << "." << endl;
2625 status = false;
2626 }
2627 output.close();
2628
2629 return status;
2630};
2631
2632/** Storing the bond structure of a molecule to file.
2633 * Simply stores Atom::nr and then the Atom::nr of all bond partners per line.
2634 * \param *out output stream for debugging
2635 * \param *path path to file
2636 * \return true - file written successfully, false - writing failed
2637 */
2638bool molecule::StoreAdjacencyToFile(ofstream *out, char *path)
2639{
2640 ofstream AdjacencyFile;
2641 atom *Walker = NULL;
2642 stringstream line;
2643 bool status = true;
2644
2645 line << path << "/" << FRAGMENTPREFIX << ADJACENCYFILE;
2646 AdjacencyFile.open(line.str().c_str(), ios::out);
2647 *out << Verbose(1) << "Saving adjacency list ... ";
2648 if (AdjacencyFile != NULL) {
2649 Walker = start;
2650 while(Walker->next != end) {
2651 Walker = Walker->next;
2652 AdjacencyFile << Walker->nr << "\t";
2653 for (int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++)
2654 AdjacencyFile << ListOfBondsPerAtom[Walker->nr][i]->GetOtherAtom(Walker)->nr << "\t";
2655 AdjacencyFile << endl;
2656 }
2657 AdjacencyFile.close();
2658 *out << Verbose(1) << "done." << endl;
2659 } else {
2660 *out << Verbose(1) << "failed to open file " << line.str() << "." << endl;
2661 status = false;
2662 }
2663
2664 return status;
2665};
2666
2667/** Checks contents of adjacency file against bond structure in structure molecule.
2668 * \param *out output stream for debugging
2669 * \param *path path to file
2670 * \param **ListOfAtoms allocated (molecule::AtomCount) and filled lookup table for ids (Atom::nr) to *Atom
2671 * \return true - structure is equal, false - not equivalence
2672 */
2673bool molecule::CheckAdjacencyFileAgainstMolecule(ofstream *out, char *path, atom **ListOfAtoms)
2674{
2675 ifstream File;
2676 stringstream filename;
2677 bool status = true;
2678 char *buffer = (char *) Malloc(sizeof(char)*MAXSTRINGSIZE, "molecule::CheckAdjacencyFileAgainstMolecule: *buffer");
2679
2680 filename << path << "/" << FRAGMENTPREFIX << ADJACENCYFILE;
2681 File.open(filename.str().c_str(), ios::out);
2682 *out << Verbose(1) << "Looking at bond structure stored in adjacency file and comparing to present one ... ";
2683 if (File != NULL) {
2684 // allocate storage structure
2685 int NonMatchNumber = 0; // will number of atoms with differing bond structure
2686 int *CurrentBonds = (int *) Malloc(sizeof(int)*8, "molecule::CheckAdjacencyFileAgainstMolecule - CurrentBonds"); // contains parsed bonds of current atom
2687 int CurrentBondsOfAtom;
2688
2689 // Parse the file line by line and count the bonds
2690 while (!File.eof()) {
2691 File.getline(buffer, MAXSTRINGSIZE);
2692 stringstream line;
2693 line.str(buffer);
2694 int AtomNr = -1;
2695 line >> AtomNr;
2696 CurrentBondsOfAtom = -1; // we count one too far due to line end
2697 // parse into structure
2698 if ((AtomNr >= 0) && (AtomNr < AtomCount)) {
2699 while (!line.eof())
2700 line >> CurrentBonds[ ++CurrentBondsOfAtom ];
2701 // compare against present bonds
2702 //cout << Verbose(2) << "Walker is " << *Walker << ", bond partners: ";
2703 if (CurrentBondsOfAtom == NumberOfBondsPerAtom[AtomNr]) {
2704 for(int i=0;i<NumberOfBondsPerAtom[AtomNr];i++) {
2705 int id = ListOfBondsPerAtom[AtomNr][i]->GetOtherAtom(ListOfAtoms[AtomNr])->nr;
2706 int j = 0;
2707 for (;(j<CurrentBondsOfAtom) && (CurrentBonds[j++] != id);); // check against all parsed bonds
2708 if (CurrentBonds[j-1] != id) { // no match ? Then mark in ListOfAtoms
2709 ListOfAtoms[AtomNr] = NULL;
2710 NonMatchNumber++;
2711 status = false;
2712 //out << "[" << id << "]\t";
2713 } else {
2714 //out << id << "\t";
2715 }
2716 }
2717 //out << endl;
2718 } else {
2719 *out << "Number of bonds for Atom " << *ListOfAtoms[AtomNr] << " does not match, parsed " << CurrentBondsOfAtom << " against " << NumberOfBondsPerAtom[AtomNr] << "." << endl;
2720 status = false;
2721 }
2722 }
2723 }
2724 File.close();
2725 File.clear();
2726 if (status) { // if equal we parse the KeySetFile
2727 *out << Verbose(1) << "done: Equal." << endl;
2728 status = true;
2729 } else
2730 *out << Verbose(1) << "done: Not equal by " << NonMatchNumber << " atoms." << endl;
2731 Free((void **)&CurrentBonds, "molecule::CheckAdjacencyFileAgainstMolecule - **CurrentBonds");
2732 } else {
2733 *out << Verbose(1) << "Adjacency file not found." << endl;
2734 status = false;
2735 }
2736 *out << endl;
2737 Free((void **)&buffer, "molecule::CheckAdjacencyFileAgainstMolecule: *buffer");
2738
2739 return status;
2740};
2741
2742/** Checks whether the OrderAtSite is still below \a Order at some site.
2743 * \param *out output stream for debugging
2744 * \param *AtomMask defines true/false per global Atom::nr to mask in/out each nuclear site, used to activate given number of site to increment order adaptively
2745 * \param *GlobalKeySetList list of keysets with global ids (valid in "this" molecule) needed for adaptive increase
2746 * \param Order desired Order if positive, desired exponent in threshold criteria if negative (0 is single-step)
2747 * \param *MinimumRingSize array of max. possible order to avoid loops
2748 * \param *path path to ENERGYPERFRAGMENT file (may be NULL if Order is non-negative)
2749 * \return true - needs further fragmentation, false - does not need fragmentation
2750 */
2751bool molecule::CheckOrderAtSite(ofstream *out, bool *AtomMask, Graph *GlobalKeySetList, int Order, int *MinimumRingSize, char *path)
2752{
2753 atom *Walker = start;
2754 bool status = false;
2755 ifstream InputFile;
2756
2757 // initialize mask list
2758 for(int i=AtomCount;i--;)
2759 AtomMask[i] = false;
2760
2761 if (Order < 0) { // adaptive increase of BondOrder per site
2762 if (AtomMask[AtomCount] == true) // break after one step
2763 return false;
2764 // parse the EnergyPerFragment file
2765 char *buffer = (char *) Malloc(sizeof(char)*MAXSTRINGSIZE, "molecule::CheckOrderAtSite: *buffer");
2766 sprintf(buffer, "%s/%s%s.dat", path, FRAGMENTPREFIX, ENERGYPERFRAGMENT);
2767 InputFile.open(buffer, ios::in);
2768 if ((InputFile != NULL) && (GlobalKeySetList != NULL)) {
2769 // transmorph graph keyset list into indexed KeySetList
2770 map<int,KeySet> IndexKeySetList;
2771 for(Graph::iterator runner = GlobalKeySetList->begin(); runner != GlobalKeySetList->end(); runner++) {
2772 IndexKeySetList.insert( pair<int,KeySet>(runner->second.first,runner->first) );
2773 }
2774 int lines = 0;
2775 // count the number of lines, i.e. the number of fragments
2776 InputFile.getline(buffer, MAXSTRINGSIZE); // skip comment lines
2777 InputFile.getline(buffer, MAXSTRINGSIZE);
2778 while(!InputFile.eof()) {
2779 InputFile.getline(buffer, MAXSTRINGSIZE);
2780 lines++;
2781 }
2782 //*out << Verbose(2) << "Scanned " << lines-1 << " lines." << endl; // one endline too much
2783 InputFile.clear();
2784 InputFile.seekg(ios::beg);
2785 map<int, pair<double,int> > AdaptiveCriteriaList; // (Root No., (Value, Order)) !
2786 int No, FragOrder;
2787 double Value;
2788 // each line represents a fragment root (Atom::nr) id and its energy contribution
2789 InputFile.getline(buffer, MAXSTRINGSIZE); // skip comment lines
2790 InputFile.getline(buffer, MAXSTRINGSIZE);
2791 while(!InputFile.eof()) {
2792 InputFile.getline(buffer, MAXSTRINGSIZE);
2793 if (strlen(buffer) > 2) {
2794 //*out << Verbose(2) << "Scanning: " << buffer << endl;
2795 stringstream line(buffer);
2796 line >> FragOrder;
2797 line >> ws >> No;
2798 line >> ws >> Value; // skip time entry
2799 line >> ws >> Value;
2800 No -= 1; // indices start at 1 in file, not 0
2801 //*out << Verbose(2) << " - yields (" << No << "," << Value << ", " << FragOrder << ")" << endl;
2802
2803 // clean the list of those entries that have been superceded by higher order terms already
2804 map<int,KeySet>::iterator marker = IndexKeySetList.find(No); // find keyset to Frag No.
2805 if (marker != IndexKeySetList.end()) { // if found
2806 Value *= 1 + MYEPSILON*(*((*marker).second.begin())); // in case of equal energies this makes em not equal without changing anything actually
2807 // as the smallest number in each set has always been the root (we use global id to keep the doubles away), seek smallest and insert into AtomMask
2808 pair <map<int, pair<double,int> >::iterator, bool> InsertedElement = AdaptiveCriteriaList.insert( make_pair(*((*marker).second.begin()), pair<double,int>( fabs(Value), FragOrder) ));
2809 map<int, pair<double,int> >::iterator PresentItem = InsertedElement.first;
2810 if (!InsertedElement.second) { // this root is already present
2811 if ((*PresentItem).second.second < FragOrder) // if order there is lower, update entry with higher-order term
2812 //if ((*PresentItem).second.first < (*runner).first) // as higher-order terms are not always better, we skip this part (which would always include this site into adaptive increase)
2813 { // if value is smaller, update value and order
2814 (*PresentItem).second.first = fabs(Value);
2815 (*PresentItem).second.second = FragOrder;
2816 *out << Verbose(2) << "Updated element (" << (*PresentItem).first << ",[" << (*PresentItem).second.first << "," << (*PresentItem).second.second << "])." << endl;
2817 } else {
2818 *out << Verbose(2) << "Did not update element " << (*PresentItem).first << " as " << FragOrder << " is less than or equal to " << (*PresentItem).second.second << "." << endl;
2819 }
2820 } else {
2821 *out << Verbose(2) << "Inserted element (" << (*PresentItem).first << ",[" << (*PresentItem).second.first << "," << (*PresentItem).second.second << "])." << endl;
2822 }
2823 } else {
2824 *out << Verbose(1) << "No Fragment under No. " << No << "found." << endl;
2825 }
2826 }
2827 }
2828 // then map back onto (Value, (Root Nr., Order)) (i.e. sorted by value to pick the highest ones)
2829 map<double, pair<int,int> > FinalRootCandidates;
2830 *out << Verbose(1) << "Root candidate list is: " << endl;
2831 for(map<int, pair<double,int> >::iterator runner = AdaptiveCriteriaList.begin(); runner != AdaptiveCriteriaList.end(); runner++) {
2832 Walker = FindAtom((*runner).first);
2833 if (Walker != NULL) {
2834 //if ((*runner).second.second >= Walker->AdaptiveOrder) { // only insert if this is an "active" root site for the current order
2835 if (!Walker->MaxOrder) {
2836 *out << Verbose(2) << "(" << (*runner).first << ",[" << (*runner).second.first << "," << (*runner).second.second << "])" << endl;
2837 FinalRootCandidates.insert( make_pair( (*runner).second.first, pair<int,int>((*runner).first, (*runner).second.second) ) );
2838 } else {
2839 *out << Verbose(2) << "Excluding (" << *Walker << ", " << (*runner).first << ",[" << (*runner).second.first << "," << (*runner).second.second << "]), as it has reached its maximum order." << endl;
2840 }
2841 } else {
2842 cerr << "Atom No. " << (*runner).second.first << " was not found in this molecule." << endl;
2843 }
2844 }
2845 // pick the ones still below threshold and mark as to be adaptively updated
2846 for(map<double, pair<int,int> >::iterator runner = FinalRootCandidates.upper_bound(pow(10.,Order)); runner != FinalRootCandidates.end(); runner++) {
2847 No = (*runner).second.first;
2848 Walker = FindAtom(No);
2849 //if (Walker->AdaptiveOrder < MinimumRingSize[Walker->nr]) {
2850 *out << Verbose(2) << "Root " << No << " is still above threshold (10^{" << Order <<"}: " << runner->first << ", setting entry " << No << " of Atom mask to true." << endl;
2851 AtomMask[No] = true;
2852 status = true;
2853 //} else
2854 //*out << Verbose(2) << "Root " << No << " is still above threshold (10^{" << Order <<"}: " << runner->first << ", however MinimumRingSize of " << MinimumRingSize[Walker->nr] << " does not allow further adaptive increase." << endl;
2855 }
2856 // close and done
2857 InputFile.close();
2858 InputFile.clear();
2859 } else {
2860 cerr << "Unable to parse " << buffer << " file, incrementing all." << endl;
2861 while (Walker->next != end) {
2862 Walker = Walker->next;
2863 #ifdef ADDHYDROGEN
2864 if (Walker->type->Z != 1) // skip hydrogen
2865 #endif
2866 {
2867 AtomMask[Walker->nr] = true; // include all (non-hydrogen) atoms
2868 status = true;
2869 }
2870 }
2871 }
2872 Free((void **)&buffer, "molecule::CheckOrderAtSite: *buffer");
2873 // pick a given number of highest values and set AtomMask
2874 } else { // global increase of Bond Order
2875 while (Walker->next != end) {
2876 Walker = Walker->next;
2877 #ifdef ADDHYDROGEN
2878 if (Walker->type->Z != 1) // skip hydrogen
2879 #endif
2880 {
2881 AtomMask[Walker->nr] = true; // include all (non-hydrogen) atoms
2882 if ((Order != 0) && (Walker->AdaptiveOrder < Order)) // && (Walker->AdaptiveOrder < MinimumRingSize[Walker->nr]))
2883 status = true;
2884 }
2885 }
2886 if ((Order == 0) && (AtomMask[AtomCount] == false)) // single stepping, just check
2887 status = true;
2888
2889 if (!status) {
2890 if (Order == 0)
2891 *out << Verbose(1) << "Single stepping done." << endl;
2892 else
2893 *out << Verbose(1) << "Order at every site is already equal or above desired order " << Order << "." << endl;
2894 }
2895 }
2896
2897 // print atom mask for debugging
2898 *out << " ";
2899 for(int i=0;i<AtomCount;i++)
2900 *out << (i % 10);
2901 *out << endl << "Atom mask is: ";
2902 for(int i=0;i<AtomCount;i++)
2903 *out << (AtomMask[i] ? "t" : "f");
2904 *out << endl;
2905
2906 return status;
2907};
2908
2909/** Create a SortIndex to map from atomic labels to the sequence in which the atoms are given in the config file.
2910 * \param *out output stream for debugging
2911 * \param *&SortIndex Mapping array of size molecule::AtomCount
2912 * \return true - success, false - failure of SortIndex alloc
2913 */
2914bool molecule::CreateMappingLabelsToConfigSequence(ofstream *out, int *&SortIndex)
2915{
2916 element *runner = elemente->start;
2917 int AtomNo = 0;
2918 atom *Walker = NULL;
2919
2920 if (SortIndex != NULL) {
2921 *out << Verbose(1) << "SortIndex is " << SortIndex << " and not NULL as expected." << endl;
2922 return false;
2923 }
2924 SortIndex = (int *) Malloc(sizeof(int)*AtomCount, "molecule::FragmentMolecule: *SortIndex");
2925 for(int i=AtomCount;i--;)
2926 SortIndex[i] = -1;
2927 while (runner->next != elemente->end) { // go through every element
2928 runner = runner->next;
2929 if (ElementsInMolecule[runner->Z]) { // if this element got atoms
2930 Walker = start;
2931 while (Walker->next != end) { // go through every atom of this element
2932 Walker = Walker->next;
2933 if (Walker->type->Z == runner->Z) // if this atom fits to element
2934 SortIndex[Walker->nr] = AtomNo++;
2935 }
2936 }
2937 }
2938 return true;
2939};
2940
2941/** Performs a many-body bond order analysis for a given bond order.
2942 * -# parses adjacency, keysets and orderatsite files
2943 * -# performs DFS to find connected subgraphs (to leave this in was a design decision: might be useful later)
2944 * -# RootStack is created for every subgraph (here, later we implement the "update 10 sites with highest energ
2945y contribution", and that's why this consciously not done in the following loop)
2946 * -# in a loop over all subgraphs
2947 * -# calls FragmentBOSSANOVA with this RootStack and within the subgraph molecule structure
2948 * -# creates molecule (fragment)s from the returned keysets (StoreFragmentFromKeySet)
2949 * -# combines the generated molecule lists from all subgraphs
2950 * -# saves to disk: fragment configs, adjacency, orderatsite, keyset files
2951 * Note that as we split "this" molecule up into a list of subgraphs, i.e. a MoleculeListClass, we have two sets
2952 * of vertex indices: Global always means the index in "this" molecule, whereas local refers to the molecule or
2953 * subgraph in the MoleculeListClass.
2954 * \param *out output stream for debugging
2955 * \param Order up to how many neighbouring bonds a fragment contains in BondOrderScheme::BottumUp scheme
2956 * \param *configuration configuration for writing config files for each fragment
2957 * \return 1 - continue, 2 - stop (no fragmentation occured)
2958 */
2959int molecule::FragmentMolecule(ofstream *out, int Order, config *configuration)
2960{
2961 MoleculeListClass *BondFragments = NULL;
2962 int *SortIndex = NULL;
2963 int *MinimumRingSize = new int[AtomCount];
2964 int FragmentCounter;
2965 MoleculeLeafClass *MolecularWalker = NULL;
2966 MoleculeLeafClass *Subgraphs = NULL; // list of subgraphs from DFS analysis
2967 fstream File;
2968 bool FragmentationToDo = true;
2969 class StackClass<bond *> *BackEdgeStack = NULL, *LocalBackEdgeStack = NULL;
2970 bool CheckOrder = false;
2971 Graph **FragmentList = NULL;
2972 Graph *ParsedFragmentList = NULL;
2973 Graph TotalGraph; // graph with all keysets however local numbers
2974 int TotalNumberOfKeySets = 0;
2975 atom **ListOfAtoms = NULL;
2976 atom ***ListOfLocalAtoms = NULL;
2977 bool *AtomMask = NULL;
2978
2979 *out << endl;
2980#ifdef ADDHYDROGEN
2981 *out << Verbose(0) << "I will treat hydrogen special and saturate dangling bonds with it." << endl;
2982#else
2983 *out << Verbose(0) << "Hydrogen is treated just like the rest of the lot." << endl;
2984#endif
2985
2986 // ++++++++++++++++++++++++++++ INITIAL STUFF: Bond structure analysis, file parsing, ... ++++++++++++++++++++++++++++++++++++++++++
2987
2988 // ===== 1. Check whether bond structure is same as stored in files ====
2989
2990 // fill the adjacency list
2991 CreateListOfBondsPerAtom(out);
2992
2993 // create lookup table for Atom::nr
2994 FragmentationToDo = FragmentationToDo && CreateFatherLookupTable(out, start, end, ListOfAtoms, AtomCount);
2995
2996 // === compare it with adjacency file ===
2997 FragmentationToDo = FragmentationToDo && CheckAdjacencyFileAgainstMolecule(out, configuration->configpath, ListOfAtoms);
2998 Free((void **)&ListOfAtoms, "molecule::FragmentMolecule - **ListOfAtoms");
2999
3000 // ===== 2. perform a DFS analysis to gather info on cyclic structure and a list of disconnected subgraphs =====
3001 Subgraphs = DepthFirstSearchAnalysis(out, BackEdgeStack);
3002 // fill the bond structure of the individually stored subgraphs
3003 Subgraphs->next->FillBondStructureFromReference(out, this, (FragmentCounter = 0), ListOfLocalAtoms, false); // we want to keep the created ListOfLocalAtoms
3004 // analysis of the cycles (print rings, get minimum cycle length) for each subgraph
3005 for(int i=AtomCount;i--;)
3006 MinimumRingSize[i] = AtomCount;
3007 MolecularWalker = Subgraphs;
3008 FragmentCounter = 0;
3009 while (MolecularWalker->next != NULL) {
3010 MolecularWalker = MolecularWalker->next;
3011 LocalBackEdgeStack = new StackClass<bond *> (MolecularWalker->Leaf->BondCount);
3012// // check the list of local atoms for debugging
3013// *out << Verbose(0) << "ListOfLocalAtoms for this subgraph is:" << endl;
3014// for (int i=0;i<AtomCount;i++)
3015// if (ListOfLocalAtoms[FragmentCounter][i] == NULL)
3016// *out << "\tNULL";
3017// else
3018// *out << "\t" << ListOfLocalAtoms[FragmentCounter][i]->Name;
3019 *out << Verbose(0) << "Gathering local back edges for subgraph " << MolecularWalker->Leaf << " with nr. " << FragmentCounter << "." << endl;
3020 MolecularWalker->Leaf->PickLocalBackEdges(out, ListOfLocalAtoms[FragmentCounter++], BackEdgeStack, LocalBackEdgeStack);
3021 *out << Verbose(0) << "Analysing the cycles of subgraph " << MolecularWalker->Leaf << " with nr. " << FragmentCounter << "." << endl;
3022 MolecularWalker->Leaf->CyclicStructureAnalysis(out, LocalBackEdgeStack, MinimumRingSize);
3023 *out << Verbose(0) << "Done with Analysing the cycles of subgraph " << MolecularWalker->Leaf << " with nr. " << FragmentCounter << "." << endl;
3024 delete(LocalBackEdgeStack);
3025 }
3026
3027 // ===== 3. if structure still valid, parse key set file and others =====
3028 FragmentationToDo = FragmentationToDo && ParseKeySetFile(out, configuration->configpath, ParsedFragmentList);
3029
3030 // ===== 4. check globally whether there's something to do actually (first adaptivity check)
3031 FragmentationToDo = FragmentationToDo && ParseOrderAtSiteFromFile(out, configuration->configpath);
3032
3033 // =================================== Begin of FRAGMENTATION ===============================
3034 // ===== 6a. assign each keyset to its respective subgraph =====
3035 Subgraphs->next->AssignKeySetsToFragment(out, this, ParsedFragmentList, ListOfLocalAtoms, FragmentList, (FragmentCounter = 0), true);
3036
3037 // ===== 6b. prepare and go into the adaptive (Order<0), single-step (Order==0) or incremental (Order>0) cycle
3038 KeyStack *RootStack = new KeyStack[Subgraphs->next->Count()];
3039 AtomMask = new bool[AtomCount+1];
3040 AtomMask[AtomCount] = false;
3041 FragmentationToDo = false; // if CheckOrderAtSite just ones recommends fragmentation, we will save fragments afterwards
3042 while ((CheckOrder = CheckOrderAtSite(out, AtomMask, ParsedFragmentList, Order, MinimumRingSize, configuration->configpath))) {
3043 FragmentationToDo = FragmentationToDo || CheckOrder;
3044 AtomMask[AtomCount] = true; // last plus one entry is used as marker that we have been through this loop once already in CheckOrderAtSite()
3045 // ===== 6b. fill RootStack for each subgraph (second adaptivity check) =====
3046 Subgraphs->next->FillRootStackForSubgraphs(out, RootStack, AtomMask, (FragmentCounter = 0));
3047
3048 // ===== 7. fill the bond fragment list =====
3049 FragmentCounter = 0;
3050 MolecularWalker = Subgraphs;
3051 while (MolecularWalker->next != NULL) {
3052 MolecularWalker = MolecularWalker->next;
3053 *out << Verbose(1) << "Fragmenting subgraph " << MolecularWalker << "." << endl;
3054 //MolecularWalker->Leaf->OutputListOfBonds(out); // output ListOfBondsPerAtom for debugging
3055 if (MolecularWalker->Leaf->first->next != MolecularWalker->Leaf->last) {
3056 // call BOSSANOVA method
3057 *out << Verbose(0) << endl << " ========== BOND ENERGY of subgraph " << FragmentCounter << " ========================= " << endl;
3058 MolecularWalker->Leaf->FragmentBOSSANOVA(out, FragmentList[FragmentCounter], RootStack[FragmentCounter], MinimumRingSize);
3059 } else {
3060 cerr << "Subgraph " << MolecularWalker << " has no atoms!" << endl;
3061 }
3062 FragmentCounter++; // next fragment list
3063 }
3064 }
3065 delete[](RootStack);
3066 delete[](AtomMask);
3067 delete(ParsedFragmentList);
3068 delete[](MinimumRingSize);
3069
3070
3071 // ==================================== End of FRAGMENTATION ============================================
3072
3073 // ===== 8a. translate list into global numbers (i.e. ones that are valid in "this" molecule, not in MolecularWalker->Leaf)
3074 Subgraphs->next->TranslateIndicesToGlobalIDs(out, FragmentList, (FragmentCounter = 0), TotalNumberOfKeySets, TotalGraph);
3075
3076 // free subgraph memory again
3077 FragmentCounter = 0;
3078 if (Subgraphs != NULL) {
3079 while (Subgraphs->next != NULL) {
3080 Subgraphs = Subgraphs->next;
3081 delete(FragmentList[FragmentCounter++]);
3082 delete(Subgraphs->previous);
3083 }
3084 delete(Subgraphs);
3085 }
3086 Free((void **)&FragmentList, "molecule::FragmentMolecule - **FragmentList");
3087
3088 // ===== 8b. gather keyset lists (graphs) from all subgraphs and transform into MoleculeListClass =====
3089 //if (FragmentationToDo) { // we should always store the fragments again as coordination might have changed slightly without changing bond structure
3090 // allocate memory for the pointer array and transmorph graphs into full molecular fragments
3091 BondFragments = new MoleculeListClass(TotalGraph.size(), AtomCount);
3092 int k=0;
3093 for(Graph::iterator runner = TotalGraph.begin(); runner != TotalGraph.end(); runner++) {
3094 KeySet test = (*runner).first;
3095 *out << "Fragment No." << (*runner).second.first << " with TEFactor " << (*runner).second.second << "." << endl;
3096 BondFragments->ListOfMolecules[k] = StoreFragmentFromKeySet(out, test, configuration);
3097 k++;
3098 }
3099 *out << k << "/" << BondFragments->NumberOfMolecules << " fragments generated from the keysets." << endl;
3100
3101 // ===== 9. Save fragments' configuration and keyset files et al to disk ===
3102 if (BondFragments->NumberOfMolecules != 0) {
3103 // create the SortIndex from BFS labels to order in the config file
3104 CreateMappingLabelsToConfigSequence(out, SortIndex);
3105
3106 *out << Verbose(1) << "Writing " << BondFragments->NumberOfMolecules << " possible bond fragmentation configs" << endl;
3107 if (BondFragments->OutputConfigForListOfFragments(out, configuration, SortIndex))
3108 *out << Verbose(1) << "All configs written." << endl;
3109 else
3110 *out << Verbose(1) << "Some config writing failed." << endl;
3111
3112 // store force index reference file
3113 BondFragments->StoreForcesFile(out, configuration->configpath, SortIndex);
3114
3115 // store keysets file
3116 StoreKeySetFile(out, TotalGraph, configuration->configpath);
3117
3118 // store Adjacency file
3119 StoreAdjacencyToFile(out, configuration->configpath);
3120
3121 // store Hydrogen saturation correction file
3122 BondFragments->AddHydrogenCorrection(out, configuration->configpath);
3123
3124 // store adaptive orders into file
3125 StoreOrderAtSiteFile(out, configuration->configpath);
3126
3127 // restore orbital and Stop values
3128 CalculateOrbitals(*configuration);
3129
3130 // free memory for bond part
3131 *out << Verbose(1) << "Freeing bond memory" << endl;
3132 delete(FragmentList); // remove bond molecule from memory
3133 Free((void **)&SortIndex, "molecule::FragmentMolecule: *SortIndex");
3134 } else
3135 *out << Verbose(1) << "FragmentList is zero on return, splitting failed." << endl;
3136 //} else
3137 // *out << Verbose(1) << "No fragments to store." << endl;
3138 *out << Verbose(0) << "End of bond fragmentation." << endl;
3139
3140 return ((int)(!FragmentationToDo)+1); // 1 - continue, 2 - stop (no fragmentation occured)
3141};
3142
3143
3144/** Picks from a global stack with all back edges the ones in the fragment.
3145 * \param *out output stream for debugging
3146 * \param **ListOfLocalAtoms array of father atom::nr to local atom::nr (reverse of atom::father)
3147 * \param *ReferenceStack stack with all the back egdes
3148 * \param *LocalStack stack to be filled
3149 * \return true - everything ok, false - ReferenceStack was empty
3150 */
3151bool molecule::PickLocalBackEdges(ofstream *out, atom **ListOfLocalAtoms, class StackClass<bond *> *&ReferenceStack, class StackClass<bond *> *&LocalStack)
3152{
3153 bool status = true;
3154 if (ReferenceStack->IsEmpty()) {
3155 cerr << "ReferenceStack is empty!" << endl;
3156 return false;
3157 }
3158 bond *Binder = ReferenceStack->PopFirst();
3159 bond *FirstBond = Binder; // mark the first bond, so that we don't loop through the stack indefinitely
3160 atom *Walker = NULL, *OtherAtom = NULL;
3161 ReferenceStack->Push(Binder);
3162
3163 do { // go through all bonds and push local ones
3164 Walker = ListOfLocalAtoms[Binder->leftatom->nr]; // get one atom in the reference molecule
3165 if (Walker != NULL) // if this Walker exists in the subgraph ...
3166 for(int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++) { // go through the local list of bonds
3167 OtherAtom = ListOfBondsPerAtom[Walker->nr][i]->GetOtherAtom(Walker);
3168 if (OtherAtom == ListOfLocalAtoms[Binder->rightatom->nr]) { // found the bond
3169 LocalStack->Push(ListOfBondsPerAtom[Walker->nr][i]);
3170 *out << Verbose(3) << "Found local edge " << *(ListOfBondsPerAtom[Walker->nr][i]) << "." << endl;
3171 break;
3172 }
3173 }
3174 Binder = ReferenceStack->PopFirst(); // loop the stack for next item
3175 *out << Verbose(3) << "Current candidate edge " << Binder << "." << endl;
3176 ReferenceStack->Push(Binder);
3177 } while (FirstBond != Binder);
3178
3179 return status;
3180};
3181
3182/** Stores pairs (Atom::nr, Atom::AdaptiveOrder) into file.
3183 * Atoms not present in the file get "-1".
3184 * \param *out output stream for debugging
3185 * \param *path path to file ORDERATSITEFILE
3186 * \return true - file writable, false - not writable
3187 */
3188bool molecule::StoreOrderAtSiteFile(ofstream *out, char *path)
3189{
3190 stringstream line;
3191 ofstream file;
3192
3193 line << path << "/" << FRAGMENTPREFIX << ORDERATSITEFILE;
3194 file.open(line.str().c_str());
3195 *out << Verbose(1) << "Writing OrderAtSite " << ORDERATSITEFILE << " ... " << endl;
3196 if (file != NULL) {
3197 atom *Walker = start;
3198 while (Walker->next != end) {
3199 Walker = Walker->next;
3200 file << Walker->nr << "\t" << (int)Walker->AdaptiveOrder << "\t" << (int)Walker->MaxOrder << endl;
3201 *out << Verbose(2) << "Storing: " << Walker->nr << "\t" << (int)Walker->AdaptiveOrder << "\t" << (int)Walker->MaxOrder << "." << endl;
3202 }
3203 file.close();
3204 *out << Verbose(1) << "done." << endl;
3205 return true;
3206 } else {
3207 *out << Verbose(1) << "failed to open file " << line.str() << "." << endl;
3208 return false;
3209 }
3210};
3211
3212/** Parses pairs(Atom::nr, Atom::AdaptiveOrder) from file and stores in molecule's Atom's.
3213 * Atoms not present in the file get "0".
3214 * \param *out output stream for debugging
3215 * \param *path path to file ORDERATSITEFILEe
3216 * \return true - file found and scanned, false - file not found
3217 * \sa ParseKeySetFile() and CheckAdjacencyFileAgainstMolecule() as this is meant to be used in conjunction with the two
3218 */
3219bool molecule::ParseOrderAtSiteFromFile(ofstream *out, char *path)
3220{
3221 unsigned char *OrderArray = (unsigned char *) Malloc(sizeof(unsigned char)*AtomCount, "molecule::ParseOrderAtSiteFromFile - *OrderArray");
3222 bool *MaxArray = (bool *) Malloc(sizeof(bool)*AtomCount, "molecule::ParseOrderAtSiteFromFile - *MaxArray");
3223 bool status;
3224 int AtomNr, value;
3225 stringstream line;
3226 ifstream file;
3227
3228 *out << Verbose(1) << "Begin of ParseOrderAtSiteFromFile" << endl;
3229 for(int i=AtomCount;i--;)
3230 OrderArray[i] = 0;
3231 line << path << "/" << FRAGMENTPREFIX << ORDERATSITEFILE;
3232 file.open(line.str().c_str());
3233 if (file != NULL) {
3234 for (int i=AtomCount;i--;) { // initialise with 0
3235 OrderArray[i] = 0;
3236 MaxArray[i] = 0;
3237 }
3238 while (!file.eof()) { // parse from file
3239 AtomNr = -1;
3240 file >> AtomNr;
3241 if (AtomNr != -1) { // test whether we really parsed something (this is necessary, otherwise last atom is set twice and to 0 on second time)
3242 file >> value;
3243 OrderArray[AtomNr] = value;
3244 file >> value;
3245 MaxArray[AtomNr] = value;
3246 //*out << Verbose(2) << "AtomNr " << AtomNr << " with order " << (int)OrderArray[AtomNr] << " and max order set to " << (int)MaxArray[AtomNr] << "." << endl;
3247 }
3248 }
3249 atom *Walker = start;
3250 while (Walker->next != end) { // fill into atom classes
3251 Walker = Walker->next;
3252 Walker->AdaptiveOrder = OrderArray[Walker->nr];
3253 Walker->MaxOrder = MaxArray[Walker->nr];
3254 *out << Verbose(2) << *Walker << " gets order " << (int)Walker->AdaptiveOrder << " and is " << (!Walker->MaxOrder ? "not " : " ") << "maxed." << endl;
3255 }
3256 file.close();
3257 *out << Verbose(1) << "done." << endl;
3258 status = true;
3259 } else {
3260 *out << Verbose(1) << "failed to open file " << line.str() << "." << endl;
3261 status = false;
3262 }
3263 Free((void **)&OrderArray, "molecule::ParseOrderAtSiteFromFile - *OrderArray");
3264 Free((void **)&MaxArray, "molecule::ParseOrderAtSiteFromFile - *MaxArray");
3265
3266 *out << Verbose(1) << "End of ParseOrderAtSiteFromFile" << endl;
3267 return status;
3268};
3269
3270/** Creates an 2d array of pointer with an entry for each atom and each bond it has.
3271 * Updates molecule::ListOfBondsPerAtom, molecule::NumberOfBondsPerAtom by parsing through
3272 * bond chain list, using molecule::AtomCount and molecule::BondCount.
3273 * Allocates memory, fills the array and exits
3274 * \param *out output stream for debugging
3275 */
3276void molecule::CreateListOfBondsPerAtom(ofstream *out)
3277{
3278 bond *Binder = NULL;
3279 atom *Walker = NULL;
3280 int TotalDegree;
3281 *out << Verbose(1) << "Begin of Creating ListOfBondsPerAtom: AtomCount = " << AtomCount << "\tBondCount = " << BondCount << "\tNoNonBonds = " << NoNonBonds << "." << endl;
3282
3283 // re-allocate memory
3284 *out << Verbose(2) << "(Re-)Allocating memory." << endl;
3285 if (ListOfBondsPerAtom != NULL) {
3286 for(int i=AtomCount;i--;)
3287 Free((void **)&ListOfBondsPerAtom[i], "molecule::CreateListOfBondsPerAtom: ListOfBondsPerAtom[i]");
3288 Free((void **)&ListOfBondsPerAtom, "molecule::CreateListOfBondsPerAtom: ListOfBondsPerAtom");
3289 }
3290 if (NumberOfBondsPerAtom != NULL)
3291 Free((void **)&NumberOfBondsPerAtom, "molecule::CreateListOfBondsPerAtom: NumberOfBondsPerAtom");
3292 ListOfBondsPerAtom = (bond ***) Malloc(sizeof(bond **)*AtomCount, "molecule::CreateListOfBondsPerAtom: ***ListOfBondsPerAtom");
3293 NumberOfBondsPerAtom = (int *) Malloc(sizeof(int)*AtomCount, "molecule::CreateListOfBondsPerAtom: *NumberOfBondsPerAtom");
3294
3295 // reset bond counts per atom
3296 for(int i=AtomCount;i--;)
3297 NumberOfBondsPerAtom[i] = 0;
3298 // count bonds per atom
3299 Binder = first;
3300 while (Binder->next != last) {
3301 Binder = Binder->next;
3302 NumberOfBondsPerAtom[Binder->leftatom->nr]++;
3303 NumberOfBondsPerAtom[Binder->rightatom->nr]++;
3304 }
3305 for(int i=AtomCount;i--;) {
3306 // allocate list of bonds per atom
3307 ListOfBondsPerAtom[i] = (bond **) Malloc(sizeof(bond *)*NumberOfBondsPerAtom[i], "molecule::CreateListOfBondsPerAtom: **ListOfBondsPerAtom[]");
3308 // clear the list again, now each NumberOfBondsPerAtom marks current free field
3309 NumberOfBondsPerAtom[i] = 0;
3310 }
3311 // fill the list
3312 Binder = first;
3313 while (Binder->next != last) {
3314 Binder = Binder->next;
3315 ListOfBondsPerAtom[Binder->leftatom->nr][NumberOfBondsPerAtom[Binder->leftatom->nr]++] = Binder;
3316 ListOfBondsPerAtom[Binder->rightatom->nr][NumberOfBondsPerAtom[Binder->rightatom->nr]++] = Binder;
3317 }
3318
3319 // output list for debugging
3320 *out << Verbose(3) << "ListOfBondsPerAtom for each atom:" << endl;
3321 Walker = start;
3322 while (Walker->next != end) {
3323 Walker = Walker->next;
3324 *out << Verbose(4) << "Atom " << Walker->Name << "/" << Walker->nr << " with " << NumberOfBondsPerAtom[Walker->nr] << " bonds: ";
3325 TotalDegree = 0;
3326 for (int j=0;j<NumberOfBondsPerAtom[Walker->nr];j++) {
3327 *out << *ListOfBondsPerAtom[Walker->nr][j] << "\t";
3328 TotalDegree += ListOfBondsPerAtom[Walker->nr][j]->BondDegree;
3329 }
3330 *out << " -- TotalDegree: " << TotalDegree << endl;
3331 }
3332 *out << Verbose(1) << "End of Creating ListOfBondsPerAtom." << endl << endl;
3333};
3334
3335/** Adds atoms up to \a BondCount distance from \a *Root and notes them down in \a **AddedAtomList.
3336 * Gray vertices are always enqueued in an StackClass<atom *> FIFO queue, the rest is usual BFS with adding vertices found was
3337 * white and putting into queue.
3338 * \param *out output stream for debugging
3339 * \param *Mol Molecule class to add atoms to
3340 * \param **AddedAtomList list with added atom pointers, index is atom father's number
3341 * \param **AddedBondList list with added bond pointers, index is bond father's number
3342 * \param *Root root vertex for BFS
3343 * \param *Bond bond not to look beyond
3344 * \param BondOrder maximum distance for vertices to add
3345 * \param IsAngstroem lengths are in angstroem or bohrradii
3346 */
3347void molecule::BreadthFirstSearchAdd(ofstream *out, molecule *Mol, atom **&AddedAtomList, bond **&AddedBondList, atom *Root, bond *Bond, int BondOrder, bool IsAngstroem)
3348{
3349 atom **PredecessorList = (atom **) Malloc(sizeof(atom *)*AtomCount, "molecule::BreadthFirstSearchAdd: **PredecessorList");
3350 int *ShortestPathList = (int *) Malloc(sizeof(int)*AtomCount, "molecule::BreadthFirstSearchAdd: *ShortestPathList");
3351 enum Shading *ColorList = (enum Shading *) Malloc(sizeof(enum Shading)*AtomCount, "molecule::BreadthFirstSearchAdd: *ColorList");
3352 class StackClass<atom *> *AtomStack = new StackClass<atom *>(AtomCount);
3353 atom *Walker = NULL, *OtherAtom = NULL;
3354 bond *Binder = NULL;
3355
3356 // add Root if not done yet
3357 AtomStack->ClearStack();
3358 if (AddedAtomList[Root->nr] == NULL) // add Root if not yet present
3359 AddedAtomList[Root->nr] = Mol->AddCopyAtom(Root);
3360 AtomStack->Push(Root);
3361
3362 // initialise each vertex as white with no predecessor, empty queue, color Root lightgray
3363 for (int i=AtomCount;i--;) {
3364 PredecessorList[i] = NULL;
3365 ShortestPathList[i] = -1;
3366 if (AddedAtomList[i] != NULL) // mark already present atoms (i.e. Root and maybe others) as visited
3367 ColorList[i] = lightgray;
3368 else
3369 ColorList[i] = white;
3370 }
3371 ShortestPathList[Root->nr] = 0;
3372
3373 // and go on ... Queue always contains all lightgray vertices
3374 while (!AtomStack->IsEmpty()) {
3375 // we have to pop the oldest atom from stack. This keeps the atoms on the stack always of the same ShortestPath distance.
3376 // e.g. if current atom is 2, push to end of stack are of length 3, but first all of length 2 would be popped. They again
3377 // append length of 3 (their neighbours). Thus on stack we have always atoms of a certain length n at bottom of stack and
3378 // followed by n+1 till top of stack.
3379 Walker = AtomStack->PopFirst(); // pop oldest added
3380 *out << Verbose(1) << "Current Walker is: " << Walker->Name << ", and has " << NumberOfBondsPerAtom[Walker->nr] << " bonds." << endl;
3381 for(int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++) {
3382 Binder = ListOfBondsPerAtom[Walker->nr][i];
3383 if (Binder != NULL) { // don't look at bond equal NULL
3384 OtherAtom = Binder->GetOtherAtom(Walker);
3385 *out << Verbose(2) << "Current OtherAtom is: " << OtherAtom->Name << " for bond " << *Binder << "." << endl;
3386 if (ColorList[OtherAtom->nr] == white) {
3387 if (Binder != Bond) // let other atom white if it's via Root bond. In case it's cyclic it has to be reached again (yet Root is from OtherAtom already black, thus no problem)
3388 ColorList[OtherAtom->nr] = lightgray;
3389 PredecessorList[OtherAtom->nr] = Walker; // Walker is the predecessor
3390 ShortestPathList[OtherAtom->nr] = ShortestPathList[Walker->nr]+1;
3391 *out << Verbose(2) << "Coloring OtherAtom " << OtherAtom->Name << " " << ((ColorList[OtherAtom->nr] == white) ? "white" : "lightgray") << ", its predecessor is " << Walker->Name << " and its Shortest Path is " << ShortestPathList[OtherAtom->nr] << " egde(s) long." << endl;
3392 if ((((ShortestPathList[OtherAtom->nr] < BondOrder) && (Binder != Bond))) ) { // Check for maximum distance
3393 *out << Verbose(3);
3394 if (AddedAtomList[OtherAtom->nr] == NULL) { // add if it's not been so far
3395 AddedAtomList[OtherAtom->nr] = Mol->AddCopyAtom(OtherAtom);
3396 *out << "Added OtherAtom " << OtherAtom->Name;
3397 AddedBondList[Binder->nr] = Mol->AddBond(AddedAtomList[Walker->nr], AddedAtomList[OtherAtom->nr], Binder->BondDegree);
3398 AddedBondList[Binder->nr]->Cyclic = Binder->Cyclic;
3399 AddedBondList[Binder->nr]->Type = Binder->Type;
3400 *out << " and bond " << *(AddedBondList[Binder->nr]) << ", ";
3401 } else { // this code should actually never come into play (all white atoms are not yet present in BondMolecule, that's why they are white in the first place)
3402 *out << "Not adding OtherAtom " << OtherAtom->Name;
3403 if (AddedBondList[Binder->nr] == NULL) {
3404 AddedBondList[Binder->nr] = Mol->AddBond(AddedAtomList[Walker->nr], AddedAtomList[OtherAtom->nr], Binder->BondDegree);
3405 AddedBondList[Binder->nr]->Cyclic = Binder->Cyclic;
3406 AddedBondList[Binder->nr]->Type = Binder->Type;
3407 *out << ", added Bond " << *(AddedBondList[Binder->nr]);
3408 } else
3409 *out << ", not added Bond ";
3410 }
3411 *out << ", putting OtherAtom into queue." << endl;
3412 AtomStack->Push(OtherAtom);
3413 } else { // out of bond order, then replace
3414 if ((AddedAtomList[OtherAtom->nr] == NULL) && (Binder->Cyclic))
3415 ColorList[OtherAtom->nr] = white; // unmark if it has not been queued/added, to make it available via its other bonds (cyclic)
3416 if (Binder == Bond)
3417 *out << Verbose(3) << "Not Queueing, is the Root bond";
3418 else if (ShortestPathList[OtherAtom->nr] >= BondOrder)
3419 *out << Verbose(3) << "Not Queueing, is out of Bond Count of " << BondOrder;
3420 if (!Binder->Cyclic)
3421 *out << ", is not part of a cyclic bond, saturating bond with Hydrogen." << endl;
3422 if (AddedBondList[Binder->nr] == NULL) {
3423 if ((AddedAtomList[OtherAtom->nr] != NULL)) { // .. whether we add or saturate
3424 AddedBondList[Binder->nr] = Mol->AddBond(AddedAtomList[Walker->nr], AddedAtomList[OtherAtom->nr], Binder->BondDegree);
3425 AddedBondList[Binder->nr]->Cyclic = Binder->Cyclic;
3426 AddedBondList[Binder->nr]->Type = Binder->Type;
3427 } else {
3428#ifdef ADDHYDROGEN
3429 if (!Mol->AddHydrogenReplacementAtom(out, Binder, AddedAtomList[Walker->nr], Walker, OtherAtom, ListOfBondsPerAtom[Walker->nr], NumberOfBondsPerAtom[Walker->nr], IsAngstroem))
3430 exit(1);
3431#endif
3432 }
3433 }
3434 }
3435 } else {
3436 *out << Verbose(3) << "Not Adding, has already been visited." << endl;
3437 // This has to be a cyclic bond, check whether it's present ...
3438 if (AddedBondList[Binder->nr] == NULL) {
3439 if ((Binder != Bond) && (Binder->Cyclic) && (((ShortestPathList[Walker->nr]+1) < BondOrder))) {
3440 AddedBondList[Binder->nr] = Mol->AddBond(AddedAtomList[Walker->nr], AddedAtomList[OtherAtom->nr], Binder->BondDegree);
3441 AddedBondList[Binder->nr]->Cyclic = Binder->Cyclic;
3442 AddedBondList[Binder->nr]->Type = Binder->Type;
3443 } else { // if it's root bond it has to broken (otherwise we would not create the fragments)
3444#ifdef ADDHYDROGEN
3445 if(!Mol->AddHydrogenReplacementAtom(out, Binder, AddedAtomList[Walker->nr], Walker, OtherAtom, ListOfBondsPerAtom[Walker->nr], NumberOfBondsPerAtom[Walker->nr], IsAngstroem))
3446 exit(1);
3447#endif
3448 }
3449 }
3450 }
3451 }
3452 }
3453 ColorList[Walker->nr] = black;
3454 *out << Verbose(1) << "Coloring Walker " << Walker->Name << " black." << endl;
3455 }
3456 Free((void **)&PredecessorList, "molecule::BreadthFirstSearchAdd: **PredecessorList");
3457 Free((void **)&ShortestPathList, "molecule::BreadthFirstSearchAdd: **ShortestPathList");
3458 Free((void **)&ColorList, "molecule::BreadthFirstSearchAdd: **ColorList");
3459 delete(AtomStack);
3460};
3461
3462/** Adds bond structure to this molecule from \a Father molecule.
3463 * This basically causes this molecule to become an induced subgraph of the \a Father, i.e. for every bond in Father
3464 * with end points present in this molecule, bond is created in this molecule.
3465 * Special care was taken to ensure that this is of complexity O(N), where N is the \a Father's molecule::AtomCount.
3466 * \param *out output stream for debugging
3467 * \param *Father father molecule
3468 * \return true - is induced subgraph, false - there are atoms with fathers not in \a Father
3469 * \todo not checked, not fully working probably
3470 */
3471bool molecule::BuildInducedSubgraph(ofstream *out, const molecule *Father)
3472{
3473 atom *Walker = NULL, *OtherAtom = NULL;
3474 bool status = true;
3475 atom **ParentList = (atom **) Malloc(sizeof(atom *)*Father->AtomCount, "molecule::BuildInducedSubgraph: **ParentList");
3476
3477 *out << Verbose(2) << "Begin of BuildInducedSubgraph." << endl;
3478
3479 // reset parent list
3480 *out << Verbose(3) << "Resetting ParentList." << endl;
3481 for (int i=Father->AtomCount;i--;)
3482 ParentList[i] = NULL;
3483
3484 // fill parent list with sons
3485 *out << Verbose(3) << "Filling Parent List." << endl;
3486 Walker = start;
3487 while (Walker->next != end) {
3488 Walker = Walker->next;
3489 ParentList[Walker->father->nr] = Walker;
3490 // Outputting List for debugging
3491 *out << Verbose(4) << "Son["<< Walker->father->nr <<"] of " << Walker->father << " is " << ParentList[Walker->father->nr] << "." << endl;
3492 }
3493
3494 // check each entry of parent list and if ok (one-to-and-onto matching) create bonds
3495 *out << Verbose(3) << "Creating bonds." << endl;
3496 Walker = Father->start;
3497 while (Walker->next != Father->end) {
3498 Walker = Walker->next;
3499 if (ParentList[Walker->nr] != NULL) {
3500 if (ParentList[Walker->nr]->father != Walker) {
3501 status = false;
3502 } else {
3503 for (int i=0;i<Father->NumberOfBondsPerAtom[Walker->nr];i++) {
3504 OtherAtom = Father->ListOfBondsPerAtom[Walker->nr][i]->GetOtherAtom(Walker);
3505 if (ParentList[OtherAtom->nr] != NULL) { // if otheratom is also a father of an atom on this molecule, create the bond
3506 *out << Verbose(4) << "Endpoints of Bond " << Father->ListOfBondsPerAtom[Walker->nr][i] << " are both present: " << ParentList[Walker->nr]->Name << " and " << ParentList[OtherAtom->nr]->Name << "." << endl;
3507 AddBond(ParentList[Walker->nr], ParentList[OtherAtom->nr], Father->ListOfBondsPerAtom[Walker->nr][i]->BondDegree);
3508 }
3509 }
3510 }
3511 }
3512 }
3513
3514 Free((void **)&ParentList, "molecule::BuildInducedSubgraph: **ParentList");
3515 *out << Verbose(2) << "End of BuildInducedSubgraph." << endl;
3516 return status;
3517};
3518
3519
3520/** Looks through a StackClass<atom *> and returns the likeliest removal candiate.
3521 * \param *out output stream for debugging messages
3522 * \param *&Leaf KeySet to look through
3523 * \param *&ShortestPathList list of the shortest path to decide which atom to suggest as removal candidate in the end
3524 * \param index of the atom suggested for removal
3525 */
3526int molecule::LookForRemovalCandidate(ofstream *&out, KeySet *&Leaf, int *&ShortestPathList)
3527{
3528 atom *Runner = NULL;
3529 int SP, Removal;
3530
3531 *out << Verbose(2) << "Looking for removal candidate." << endl;
3532 SP = -1; //0; // not -1, so that Root is never removed
3533 Removal = -1;
3534 for (KeySet::iterator runner = Leaf->begin(); runner != Leaf->end(); runner++) {
3535 Runner = FindAtom((*runner));
3536 if (Runner->type->Z != 1) { // skip all those added hydrogens when re-filling snake stack
3537 if (ShortestPathList[(*runner)] > SP) { // remove the oldest one with longest shortest path
3538 SP = ShortestPathList[(*runner)];
3539 Removal = (*runner);
3540 }
3541 }
3542 }
3543 return Removal;
3544};
3545
3546/** Stores a fragment from \a KeySet into \a molecule.
3547 * First creates the minimal set of atoms from the KeySet, then creates the bond structure from the complete
3548 * molecule and adds missing hydrogen where bonds were cut.
3549 * \param *out output stream for debugging messages
3550 * \param &Leaflet pointer to KeySet structure
3551 * \param IsAngstroem whether we have Ansgtroem or bohrradius
3552 * \return pointer to constructed molecule
3553 */
3554molecule * molecule::StoreFragmentFromKeySet(ofstream *out, KeySet &Leaflet, bool IsAngstroem)
3555{
3556 atom *Runner = NULL, *FatherOfRunner = NULL, *OtherFather = NULL;
3557 atom **SonList = (atom **) Malloc(sizeof(atom *)*AtomCount, "molecule::StoreFragmentFromStack: **SonList");
3558 molecule *Leaf = new molecule(elemente);
3559 bool LonelyFlag = false;
3560 int size;
3561
3562// *out << Verbose(1) << "Begin of StoreFragmentFromKeyset." << endl;
3563
3564 Leaf->BondDistance = BondDistance;
3565 for(int i=NDIM*2;i--;)
3566 Leaf->cell_size[i] = cell_size[i];
3567
3568 // initialise SonList (indicates when we need to replace a bond with hydrogen instead)
3569 for(int i=AtomCount;i--;)
3570 SonList[i] = NULL;
3571
3572 // first create the minimal set of atoms from the KeySet
3573 size = 0;
3574 for(KeySet::iterator runner = Leaflet.begin(); runner != Leaflet.end(); runner++) {
3575 FatherOfRunner = FindAtom((*runner)); // find the id
3576 SonList[FatherOfRunner->nr] = Leaf->AddCopyAtom(FatherOfRunner);
3577 size++;
3578 }
3579
3580 // create the bonds between all: Make it an induced subgraph and add hydrogen
3581// *out << Verbose(2) << "Creating bonds from father graph (i.e. induced subgraph creation)." << endl;
3582 Runner = Leaf->start;
3583 while (Runner->next != Leaf->end) {
3584 Runner = Runner->next;
3585 LonelyFlag = true;
3586 FatherOfRunner = Runner->father;
3587 if (SonList[FatherOfRunner->nr] != NULL) { // check if this, our father, is present in list
3588 // create all bonds
3589 for (int i=0;i<NumberOfBondsPerAtom[FatherOfRunner->nr];i++) { // go through every bond of father
3590 OtherFather = ListOfBondsPerAtom[FatherOfRunner->nr][i]->GetOtherAtom(FatherOfRunner);
3591// *out << Verbose(2) << "Father " << *FatherOfRunner << " of son " << *SonList[FatherOfRunner->nr] << " is bound to " << *OtherFather;
3592 if (SonList[OtherFather->nr] != NULL) {
3593// *out << ", whose son is " << *SonList[OtherFather->nr] << "." << endl;
3594 if (OtherFather->nr > FatherOfRunner->nr) { // add bond (nr check is for adding only one of both variants: ab, ba)
3595// *out << Verbose(3) << "Adding Bond: ";
3596// *out <<
3597 Leaf->AddBond(Runner, SonList[OtherFather->nr], ListOfBondsPerAtom[FatherOfRunner->nr][i]->BondDegree);
3598// *out << "." << endl;
3599 //NumBonds[Runner->nr]++;
3600 } else {
3601// *out << Verbose(3) << "Not adding bond, labels in wrong order." << endl;
3602 }
3603 LonelyFlag = false;
3604 } else {
3605// *out << ", who has no son in this fragment molecule." << endl;
3606#ifdef ADDHYDROGEN
3607 //*out << Verbose(3) << "Adding Hydrogen to " << Runner->Name << " and a bond in between." << endl;
3608 if(!Leaf->AddHydrogenReplacementAtom(out, ListOfBondsPerAtom[FatherOfRunner->nr][i], Runner, FatherOfRunner, OtherFather, ListOfBondsPerAtom[FatherOfRunner->nr],NumberOfBondsPerAtom[FatherOfRunner->nr], IsAngstroem))
3609 exit(1);
3610#endif
3611 //NumBonds[Runner->nr] += ListOfBondsPerAtom[FatherOfRunner->nr][i]->BondDegree;
3612 }
3613 }
3614 } else {
3615 *out << Verbose(0) << "ERROR: Son " << Runner->Name << " has father " << FatherOfRunner->Name << " but its entry in SonList is " << SonList[FatherOfRunner->nr] << "!" << endl;
3616 }
3617 if ((LonelyFlag) && (size > 1)) {
3618 *out << Verbose(0) << *Runner << "has got bonds only to hydrogens!" << endl;
3619 }
3620#ifdef ADDHYDROGEN
3621 while ((Runner->next != Leaf->end) && (Runner->next->type->Z == 1)) // skip added hydrogen
3622 Runner = Runner->next;
3623#endif
3624 }
3625 Leaf->CreateListOfBondsPerAtom(out);
3626 //Leaflet->Leaf->ScanForPeriodicCorrection(out);
3627 Free((void **)&SonList, "molecule::StoreFragmentFromStack: **SonList");
3628// *out << Verbose(1) << "End of StoreFragmentFromKeyset." << endl;
3629 return Leaf;
3630};
3631
3632/** Creates \a MoleculeListClass of all unique fragments of the \a molecule containing \a Order atoms or vertices.
3633 * The picture to have in mind is that of a DFS "snake" of a certain length \a Order, i.e. as in the infamous
3634 * computer game, that winds through the connected graph representing the molecule. Color (white,
3635 * lightgray, darkgray, black) indicates whether a vertex has been discovered so far or not. Labels will help in
3636 * creating only unique fragments and not additional ones with vertices simply in different sequence.
3637 * The Predecessor is always the one that came before in discovering, needed on backstepping. And
3638 * finally, the ShortestPath is needed for removing vertices from the snake stack during the back-
3639 * stepping.
3640 * \param *out output stream for debugging
3641 * \param Order number of atoms in each fragment
3642 * \param *configuration configuration for writing config files for each fragment
3643 * \return List of all unique fragments with \a Order atoms
3644 */
3645/*
3646MoleculeListClass * molecule::CreateListOfUniqueFragmentsOfOrder(ofstream *out, int Order, config *configuration)
3647{
3648 atom **PredecessorList = (atom **) Malloc(sizeof(atom *)*AtomCount, "molecule::CreateListOfUniqueFragmentsOfOrder: **PredecessorList");
3649 int *ShortestPathList = (int *) Malloc(sizeof(int)*AtomCount, "molecule::CreateListOfUniqueFragmentsOfOrder: *ShortestPathList");
3650 int *Labels = (int *) Malloc(sizeof(int)*AtomCount, "molecule::CreateListOfUniqueFragmentsOfOrder: *Labels");
3651 enum Shading *ColorVertexList = (enum Shading *) Malloc(sizeof(enum Shading)*AtomCount, "molecule::CreateListOfUniqueFragmentsOfOrder: *ColorList");
3652 enum Shading *ColorEdgeList = (enum Shading *) Malloc(sizeof(enum Shading)*BondCount, "molecule::CreateListOfUniqueFragmentsOfOrder: *ColorBondList");
3653 StackClass<atom *> *RootStack = new StackClass<atom *>(AtomCount);
3654 StackClass<atom *> *TouchedStack = new StackClass<atom *>((int)pow(4,Order)+2); // number of atoms reached from one with maximal 4 bonds plus Root itself
3655 StackClass<atom *> *SnakeStack = new StackClass<atom *>(Order+1); // equal to Order is not possible, as then the StackClass<atom *> cannot discern between full and empty stack!
3656 MoleculeLeafClass *Leaflet = NULL, *TempLeaf = NULL;
3657 MoleculeListClass *FragmentList = NULL;
3658 atom *Walker = NULL, *OtherAtom = NULL, *Root = NULL, *Removal = NULL;
3659 bond *Binder = NULL;
3660 int RunningIndex = 0, FragmentCounter = 0;
3661
3662 *out << Verbose(1) << "Begin of CreateListOfUniqueFragmentsOfOrder." << endl;
3663
3664 // reset parent list
3665 *out << Verbose(3) << "Resetting labels, parent, predecessor, color and shortest path lists." << endl;
3666 for (int i=0;i<AtomCount;i++) { // reset all atom labels
3667 // initialise each vertex as white with no predecessor, empty queue, color lightgray, not labelled, no sons
3668 Labels[i] = -1;
3669 SonList[i] = NULL;
3670 PredecessorList[i] = NULL;
3671 ColorVertexList[i] = white;
3672 ShortestPathList[i] = -1;
3673 }
3674 for (int i=0;i<BondCount;i++)
3675 ColorEdgeList[i] = white;
3676 RootStack->ClearStack(); // clearstack and push first atom if exists
3677 TouchedStack->ClearStack();
3678 Walker = start->next;
3679 while ((Walker != end)
3680#ifdef ADDHYDROGEN
3681 && (Walker->type->Z == 1)
3682#endif
3683 ) { // search for first non-hydrogen atom
3684 *out << Verbose(4) << "Current Root candidate is " << Walker->Name << "." << endl;
3685 Walker = Walker->next;
3686 }
3687 if (Walker != end)
3688 RootStack->Push(Walker);
3689 else
3690 *out << Verbose(0) << "ERROR: Could not find an appropriate Root atom!" << endl;
3691 *out << Verbose(3) << "Root " << Walker->Name << " is on AtomStack, beginning loop through all vertices ..." << endl;
3692
3693 ///// OUTER LOOP ////////////
3694 while (!RootStack->IsEmpty()) {
3695 // get new root vertex from atom stack
3696 Root = RootStack->PopFirst();
3697 ShortestPathList[Root->nr] = 0;
3698 if (Labels[Root->nr] == -1)
3699 Labels[Root->nr] = RunningIndex++; // prevent it from getting again on AtomStack
3700 PredecessorList[Root->nr] = Root;
3701 TouchedStack->Push(Root);
3702 *out << Verbose(0) << "Root for this loop is: " << Root->Name << ".\n";
3703
3704 // clear snake stack
3705 SnakeStack->ClearStack();
3706 //SnakeStack->TestImplementation(out, start->next);
3707
3708 ///// INNER LOOP ////////////
3709 // Problems:
3710 // - what about cyclic bonds?
3711 Walker = Root;
3712 do {
3713 *out << Verbose(1) << "Current Walker is: " << Walker->Name;
3714 // initial setting of the new Walker: label, color, shortest path and put on stacks
3715 if (Labels[Walker->nr] == -1) { // give atom a unique, monotonely increasing number
3716 Labels[Walker->nr] = RunningIndex++;
3717 RootStack->Push(Walker);
3718 }
3719 *out << ", has label " << Labels[Walker->nr];
3720 if ((ColorVertexList[Walker->nr] == white) || ((Binder != NULL) && (ColorEdgeList[Binder->nr] == white))) { // color it if newly discovered and push on stacks (and if within reach!)
3721 if ((Binder != NULL) && (ColorEdgeList[Binder->nr] == white)) {
3722 // Binder ought to be set still from last neighbour search
3723 *out << ", coloring bond " << *Binder << " black";
3724 ColorEdgeList[Binder->nr] = black; // mark this bond as used
3725 }
3726 if (ShortestPathList[Walker->nr] == -1) {
3727 ShortestPathList[Walker->nr] = ShortestPathList[PredecessorList[Walker->nr]->nr]+1;
3728 TouchedStack->Push(Walker); // mark every atom for lists cleanup later, whose shortest path has been changed
3729 }
3730 if ((ShortestPathList[Walker->nr] < Order) && (ColorVertexList[Walker->nr] != darkgray)) { // if not already on snake stack
3731 SnakeStack->Push(Walker);
3732 ColorVertexList[Walker->nr] = darkgray; // mark as dark gray of on snake stack
3733 }
3734 }
3735 *out << ", SP of " << ShortestPathList[Walker->nr] << " and its color is " << GetColor(ColorVertexList[Walker->nr]) << "." << endl;
3736
3737 // then check the stack for a newly stumbled upon fragment
3738 if (SnakeStack->ItemCount() == Order) { // is stack full?
3739 // store the fragment if it is one and get a removal candidate
3740 Removal = StoreFragmentFromStack(out, Root, Walker, Leaflet, SnakeStack, ShortestPathList, SonList, Labels, &FragmentCounter, configuration);
3741 // remove the candidate if one was found
3742 if (Removal != NULL) {
3743 *out << Verbose(2) << "Removing item " << Removal->Name << " with SP of " << ShortestPathList[Removal->nr] << " from snake stack." << endl;
3744 SnakeStack->RemoveItem(Removal);
3745 ColorVertexList[Removal->nr] = lightgray; // return back to not on snake stack but explored marking
3746 if (Walker == Removal) { // if the current atom is to be removed, we also have to take a step back
3747 Walker = PredecessorList[Removal->nr];
3748 *out << Verbose(2) << "Stepping back to " << Walker->Name << "." << endl;
3749 }
3750 }
3751 } else
3752 Removal = NULL;
3753
3754 // finally, look for a white neighbour as the next Walker
3755 Binder = NULL;
3756 if ((Removal == NULL) || (Walker != PredecessorList[Removal->nr])) { // don't look, if a new walker has been set above
3757 *out << Verbose(2) << "Snake has currently " << SnakeStack->ItemCount() << " item(s)." << endl;
3758 OtherAtom = NULL; // this is actually not needed, every atom has at least one neighbour
3759 if (ShortestPathList[Walker->nr] < Order) {
3760 for(int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++) {
3761 Binder = ListOfBondsPerAtom[Walker->nr][i];
3762 *out << Verbose(2) << "Current bond is " << *Binder << ": ";
3763 OtherAtom = Binder->GetOtherAtom(Walker);
3764 if ((Labels[OtherAtom->nr] != -1) && (Labels[OtherAtom->nr] < Labels[Root->nr])) { // we don't step up to labels bigger than us
3765 *out << "Label " << Labels[OtherAtom->nr] << " is smaller than Root's " << Labels[Root->nr] << "." << endl;
3766 //ColorVertexList[OtherAtom->nr] = lightgray; // mark as explored
3767 } else { // otherwise check its colour and element
3768 if (
3769#ifdef ADDHYDROGEN
3770 (OtherAtom->type->Z != 1) &&
3771#endif
3772 (ColorEdgeList[Binder->nr] == white)) { // skip hydrogen, look for unexplored vertices
3773 *out << "Moving along " << GetColor(ColorEdgeList[Binder->nr]) << " bond " << Binder << " to " << ((ColorVertexList[OtherAtom->nr] == white) ? "unexplored" : "explored") << " item: " << OtherAtom->Name << "." << endl;
3774 // i find it currently rather sensible to always set the predecessor in order to find one's way back
3775 //if (PredecessorList[OtherAtom->nr] == NULL) {
3776 PredecessorList[OtherAtom->nr] = Walker;
3777 *out << Verbose(3) << "Setting Predecessor of " << OtherAtom->Name << " to " << PredecessorList[OtherAtom->nr]->Name << "." << endl;
3778 //} else {
3779 // *out << Verbose(3) << "Predecessor of " << OtherAtom->Name << " is " << PredecessorList[OtherAtom->nr]->Name << "." << endl;
3780 //}
3781 Walker = OtherAtom;
3782 break;
3783 } else {
3784 if (OtherAtom->type->Z == 1)
3785 *out << "Links to a hydrogen atom." << endl;
3786 else
3787 *out << "Bond has not white but " << GetColor(ColorEdgeList[Binder->nr]) << " color." << endl;
3788 }
3789 }
3790 }
3791 } else { // means we have stepped beyond the horizon: Return!
3792 Walker = PredecessorList[Walker->nr];
3793 OtherAtom = Walker;
3794 *out << Verbose(3) << "We have gone too far, stepping back to " << Walker->Name << "." << endl;
3795 }
3796 if (Walker != OtherAtom) { // if no white neighbours anymore, color it black
3797 *out << Verbose(2) << "Coloring " << Walker->Name << " black." << endl;
3798 ColorVertexList[Walker->nr] = black;
3799 Walker = PredecessorList[Walker->nr];
3800 }
3801 }
3802 } while ((Walker != Root) || (ColorVertexList[Root->nr] != black));
3803 *out << Verbose(2) << "Inner Looping is finished." << endl;
3804
3805 // if we reset all AtomCount atoms, we have again technically O(N^2) ...
3806 *out << Verbose(2) << "Resetting lists." << endl;
3807 Walker = NULL;
3808 Binder = NULL;
3809 while (!TouchedStack->IsEmpty()) {
3810 Walker = TouchedStack->PopLast();
3811 *out << Verbose(3) << "Re-initialising entries of " << *Walker << "." << endl;
3812 for(int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++)
3813 ColorEdgeList[ListOfBondsPerAtom[Walker->nr][i]->nr] = white;
3814 PredecessorList[Walker->nr] = NULL;
3815 ColorVertexList[Walker->nr] = white;
3816 ShortestPathList[Walker->nr] = -1;
3817 }
3818 }
3819 *out << Verbose(1) << "Outer Looping over all vertices is done." << endl;
3820
3821 // copy together
3822 *out << Verbose(1) << "Copying all fragments into MoleculeList structure." << endl;
3823 FragmentList = new MoleculeListClass(FragmentCounter, AtomCount);
3824 RunningIndex = 0;
3825 while ((Leaflet != NULL) && (RunningIndex < FragmentCounter)) {
3826 FragmentList->ListOfMolecules[RunningIndex++] = Leaflet->Leaf;
3827 Leaflet->Leaf = NULL; // prevent molecule from being removed
3828 TempLeaf = Leaflet;
3829 Leaflet = Leaflet->previous;
3830 delete(TempLeaf);
3831 };
3832
3833 // free memory and exit
3834 Free((void **)&PredecessorList, "molecule::CreateListOfUniqueFragmentsOfOrder: **PredecessorList");
3835 Free((void **)&ShortestPathList, "molecule::CreateListOfUniqueFragmentsOfOrder: *ShortestPathList");
3836 Free((void **)&Labels, "molecule::CreateListOfUniqueFragmentsOfOrder: *Labels");
3837 Free((void **)&ColorVertexList, "molecule::CreateListOfUniqueFragmentsOfOrder: *ColorList");
3838 delete(RootStack);
3839 delete(TouchedStack);
3840 delete(SnakeStack);
3841
3842 *out << Verbose(1) << "End of CreateListOfUniqueFragmentsOfOrder." << endl;
3843 return FragmentList;
3844};
3845*/
3846
3847/** Structure containing all values in power set combination generation.
3848 */
3849struct UniqueFragments {
3850 config *configuration;
3851 atom *Root;
3852 Graph *Leaflet;
3853 KeySet *FragmentSet;
3854 int ANOVAOrder;
3855 int FragmentCounter;
3856 int CurrentIndex;
3857 double TEFactor;
3858 int *ShortestPathList;
3859 bool **UsedList;
3860 bond **BondsPerSPList;
3861 int *BondsPerSPCount;
3862};
3863
3864/** From a given set of Bond sorted by Shortest Path distance, create all possible fragments of size \a SetDimension.
3865 * -# loops over every possible combination (2^dimension of edge set)
3866 * -# inserts current set, if there's still space left
3867 * -# yes: calls SPFragmentGenerator with structure, created new edge list and size respective to root dist
3868ance+1
3869 * -# no: stores fragment into keyset list by calling InsertFragmentIntoGraph
3870 * -# removes all items added into the snake stack (in UniqueFragments structure) added during level (root
3871distance) and current set
3872 * \param *out output stream for debugging
3873 * \param FragmentSearch UniqueFragments structure with all values needed
3874 * \param RootDistance current shortest path level, whose set of edges is represented by **BondsSet
3875 * \param SetDimension Number of possible bonds on this level (i.e. size of the array BondsSet[])
3876 * \param SubOrder remaining number of allowed vertices to add
3877 */
3878void molecule::SPFragmentGenerator(ofstream *out, struct UniqueFragments *FragmentSearch, int RootDistance, bond **BondsSet, int SetDimension, int SubOrder)
3879{
3880 atom *OtherWalker = NULL;
3881 int verbosity = 0; //FragmentSearch->ANOVAOrder-SubOrder;
3882 int NumCombinations;
3883 bool bit;
3884 int bits, TouchedIndex, SubSetDimension, SP, Added;
3885 int Removal;
3886 int SpaceLeft;
3887 int *TouchedList = (int *) Malloc(sizeof(int)*(SubOrder+1), "molecule::SPFragmentGenerator: *TouchedList");
3888 bond *Binder = NULL;
3889 bond **BondsList = NULL;
3890 KeySetTestPair TestKeySetInsert;
3891
3892 NumCombinations = 1 << SetDimension;
3893
3894 // Hier muessen von 1 bis NumberOfBondsPerAtom[Walker->nr] alle Kombinationen
3895 // von Endstuecken (aus den Bonds) hinzugefᅵᅵgt werden und fᅵᅵr verbleibende ANOVAOrder
3896 // rekursiv GraphCrawler in der nᅵᅵchsten Ebene aufgerufen werden
3897
3898 *out << Verbose(1+verbosity) << "Begin of SPFragmentGenerator." << endl;
3899 *out << Verbose(1+verbosity) << "We are " << RootDistance << " away from Root, which is " << *FragmentSearch->Root << ", SubOrder is " << SubOrder << ", SetDimension is " << SetDimension << " and this means " << NumCombinations-1 << " combination(s)." << endl;
3900
3901 // initialised touched list (stores added atoms on this level)
3902 *out << Verbose(1+verbosity) << "Clearing touched list." << endl;
3903 for (TouchedIndex=SubOrder+1;TouchedIndex--;) // empty touched list
3904 TouchedList[TouchedIndex] = -1;
3905 TouchedIndex = 0;
3906
3907 // create every possible combination of the endpieces
3908 *out << Verbose(1+verbosity) << "Going through all combinations of the power set." << endl;
3909 for (int i=1;i<NumCombinations;i++) { // sweep through all power set combinations (skip empty set!)
3910 // count the set bit of i
3911 bits = 0;
3912 for (int j=SetDimension;j--;)
3913 bits += (i & (1 << j)) >> j;
3914
3915 *out << Verbose(1+verbosity) << "Current set is " << Binary(i | (1 << SetDimension)) << ", number of bits is " << bits << "." << endl;
3916 if (bits <= SubOrder) { // if not greater than additional atoms allowed on stack, continue
3917 // --1-- add this set of the power set of bond partners to the snake stack
3918 Added = 0;
3919 for (int j=0;j<SetDimension;j++) { // pull out every bit by shifting
3920 bit = ((i & (1 << j)) != 0); // mask the bit for the j-th bond
3921 if (bit) { // if bit is set, we add this bond partner
3922 OtherWalker = BondsSet[j]->rightatom; // rightatom is always the one more distant, i.e. the one to add
3923 //*out << Verbose(1+verbosity) << "Current Bond is " << ListOfBondsPerAtom[Walker->nr][i] << ", checking on " << *OtherWalker << "." << endl;
3924 *out << Verbose(2+verbosity) << "Adding " << *OtherWalker << " with nr " << OtherWalker->nr << "." << endl;
3925 TestKeySetInsert = FragmentSearch->FragmentSet->insert(OtherWalker->nr);
3926 if (TestKeySetInsert.second) {
3927 TouchedList[TouchedIndex++] = OtherWalker->nr; // note as added
3928 Added++;
3929 } else {
3930 *out << Verbose(2+verbosity) << "This was item was already present in the keyset." << endl;
3931 }
3932 //FragmentSearch->UsedList[OtherWalker->nr][i] = true;
3933 //}
3934 } else {
3935 *out << Verbose(2+verbosity) << "Not adding." << endl;
3936 }
3937 }
3938
3939 SpaceLeft = SubOrder - Added ;// SubOrder - bits; // due to item's maybe being already present, this does not work anymore
3940 if (SpaceLeft > 0) {
3941 *out << Verbose(1+verbosity) << "There's still some space left on stack: " << SpaceLeft << "." << endl;
3942 if (SubOrder > 1) { // Due to Added above we have to check extra whether we're not already reaching beyond the desired Order
3943 // --2-- look at all added end pieces of this combination, construct bond subsets and sweep through a power set of these by recursion
3944 SP = RootDistance+1; // this is the next level
3945 // first count the members in the subset
3946 SubSetDimension = 0;
3947 Binder = FragmentSearch->BondsPerSPList[2*SP]; // start node for this level
3948 while (Binder->next != FragmentSearch->BondsPerSPList[2*SP+1]) { // compare to end node of this level
3949 Binder = Binder->next;
3950 for (int k=TouchedIndex;k--;) {
3951 if (Binder->Contains(TouchedList[k])) // if we added this very endpiece
3952 SubSetDimension++;
3953 }
3954 }
3955 // then allocate and fill the list
3956 BondsList = (bond **) Malloc(sizeof(bond *)*SubSetDimension, "molecule::SPFragmentGenerator: **BondsList");
3957 SubSetDimension = 0;
3958 Binder = FragmentSearch->BondsPerSPList[2*SP];
3959 while (Binder->next != FragmentSearch->BondsPerSPList[2*SP+1]) {
3960 Binder = Binder->next;
3961 for (int k=0;k<TouchedIndex;k++) {
3962 if (Binder->leftatom->nr == TouchedList[k]) // leftatom is always the close one
3963 BondsList[SubSetDimension++] = Binder;
3964 }
3965 }
3966 *out << Verbose(2+verbosity) << "Calling subset generator " << SP << " away from root " << *FragmentSearch->Root << " with sub set dimension " << SubSetDimension << "." << endl;
3967 SPFragmentGenerator(out, FragmentSearch, SP, BondsList, SubSetDimension, SubOrder-bits);
3968 Free((void **)&BondsList, "molecule::SPFragmentGenerator: **BondsList");
3969 }
3970 } else {
3971 // --2-- otherwise store the complete fragment
3972 *out << Verbose(1+verbosity) << "Enough items on stack for a fragment!" << endl;
3973 // store fragment as a KeySet
3974 *out << Verbose(2) << "Found a new fragment[" << FragmentSearch->FragmentCounter << "], local nr.s are: ";
3975 for(KeySet::iterator runner = FragmentSearch->FragmentSet->begin(); runner != FragmentSearch->FragmentSet->end(); runner++)
3976 *out << (*runner) << " ";
3977 *out << endl;
3978 //if (!CheckForConnectedSubgraph(out, FragmentSearch->FragmentSet))
3979 //*out << Verbose(0) << "ERROR: The found fragment is not a connected subgraph!" << endl;
3980 InsertFragmentIntoGraph(out, FragmentSearch);
3981 //Removal = LookForRemovalCandidate(out, FragmentSearch->FragmentSet, FragmentSearch->ShortestPathList);
3982 //Removal = StoreFragmentFromStack(out, FragmentSearch->Root, FragmentSearch->Leaflet, FragmentSearch->FragmentStack, FragmentSearch->ShortestPathList, &FragmentSearch->FragmentCounter, FragmentSearch->configuration);
3983 }
3984
3985 // --3-- remove all added items in this level from snake stack
3986 *out << Verbose(1+verbosity) << "Removing all items that were added on this SP level " << RootDistance << "." << endl;
3987 for(int j=0;j<TouchedIndex;j++) {
3988 Removal = TouchedList[j];
3989 *out << Verbose(2+verbosity) << "Removing item nr. " << Removal << " from snake stack." << endl;
3990 FragmentSearch->FragmentSet->erase(Removal);
3991 TouchedList[j] = -1;
3992 }
3993 *out << Verbose(2) << "Remaining local nr.s on snake stack are: ";
3994 for(KeySet::iterator runner = FragmentSearch->FragmentSet->begin(); runner != FragmentSearch->FragmentSet->end(); runner++)
3995 *out << (*runner) << " ";
3996 *out << endl;
3997 TouchedIndex = 0; // set Index to 0 for list of atoms added on this level
3998 } else {
3999 *out << Verbose(2+verbosity) << "More atoms to add for this set (" << bits << ") than space left on stack " << SubOrder << ", skipping this set." << endl;
4000 }
4001 }
4002 Free((void **)&TouchedList, "molecule::SPFragmentGenerator: *TouchedList");
4003 *out << Verbose(1+verbosity) << "End of SPFragmentGenerator, " << RootDistance << " away from Root " << *FragmentSearch->Root << " and SubOrder is " << SubOrder << "." << endl;
4004};
4005
4006/** For a given keyset \a *Fragment, checks whether it is connected in the current molecule.
4007 * \param *out output stream for debugging
4008 * \param *Fragment Keyset of fragment's vertices
4009 * \return true - connected, false - disconnected
4010 * \note this is O(n^2) for it's just a bug checker not meant for permanent use!
4011 */
4012bool molecule::CheckForConnectedSubgraph(ofstream *out, KeySet *Fragment)
4013{
4014 atom *Walker = NULL, *Walker2 = NULL;
4015 bool BondStatus = false;
4016 int size;
4017
4018 *out << Verbose(1) << "Begin of CheckForConnectedSubgraph" << endl;
4019 *out << Verbose(2) << "Disconnected atom: ";
4020
4021 // count number of atoms in graph
4022 size = 0;
4023 for(KeySet::iterator runner = Fragment->begin(); runner != Fragment->end(); runner++)
4024 size++;
4025 if (size > 1)
4026 for(KeySet::iterator runner = Fragment->begin(); runner != Fragment->end(); runner++) {
4027 Walker = FindAtom(*runner);
4028 BondStatus = false;
4029 for(KeySet::iterator runners = Fragment->begin(); runners != Fragment->end(); runners++) {
4030 Walker2 = FindAtom(*runners);
4031 for (int i=0;i<NumberOfBondsPerAtom[Walker->nr]; i++) {
4032 if (ListOfBondsPerAtom[Walker->nr][i]->GetOtherAtom(Walker) == Walker2) {
4033 BondStatus = true;
4034 break;
4035 }
4036 if (BondStatus)
4037 break;
4038 }
4039 }
4040 if (!BondStatus) {
4041 *out << (*Walker) << endl;
4042 return false;
4043 }
4044 }
4045 else {
4046 *out << "none." << endl;
4047 return true;
4048 }
4049 *out << "none." << endl;
4050
4051 *out << Verbose(1) << "End of CheckForConnectedSubgraph" << endl;
4052
4053 return true;
4054}
4055
4056/** Creates a list of all unique fragments of certain vertex size from a given graph \a Fragment for a given root vertex in the context of \a this molecule.
4057 * -# initialises UniqueFragments structure
4058 * -# fills edge list via BFS
4059 * -# creates the fragment by calling recursive function SPFragmentGenerator with UniqueFragments structure, 0 as
4060 root distance, the edge set, its dimension and the current suborder
4061 * -# Free'ing structure
4062 * Note that we may use the fact that the atoms are SP-ordered on the atomstack. I.e. when popping always the last, we first get all
4063 * with SP of 2, then those with SP of 3, then those with SP of 4 and so on.
4064 * \param *out output stream for debugging
4065 * \param Order bond order (limits BFS exploration and "number of digits" in power set generation
4066 * \param FragmentSearch UniqueFragments structure containing TEFactor, root atom and so on
4067 * \param RestrictedKeySet Restricted vertex set to use in context of molecule
4068 * \return number of inserted fragments
4069 * \note ShortestPathList in FragmentSearch structure is probably due to NumberOfAtomsSPLevel and SP not needed anymore
4070 */
4071int molecule::PowerSetGenerator(ofstream *out, int Order, struct UniqueFragments &FragmentSearch, KeySet RestrictedKeySet)
4072{
4073 int SP, AtomKeyNr;
4074 atom *Walker = NULL, *OtherWalker = NULL, *Predecessor = NULL;
4075 bond *Binder = NULL;
4076 bond *CurrentEdge = NULL;
4077 bond **BondsList = NULL;
4078 int RootKeyNr = FragmentSearch.Root->GetTrueFather()->nr;
4079 int Counter = FragmentSearch.FragmentCounter;
4080 int RemainingWalkers;
4081
4082 *out << endl;
4083 *out << Verbose(0) << "Begin of PowerSetGenerator with order " << Order << " at Root " << *FragmentSearch.Root << "." << endl;
4084
4085 // prepare Label and SP arrays of the BFS search
4086 FragmentSearch.ShortestPathList[FragmentSearch.Root->nr] = 0;
4087
4088 // prepare root level (SP = 0) and a loop bond denoting Root
4089 for (int i=1;i<Order;i++)
4090 FragmentSearch.BondsPerSPCount[i] = 0;
4091 FragmentSearch.BondsPerSPCount[0] = 1;
4092 Binder = new bond(FragmentSearch.Root, FragmentSearch.Root);
4093 add(Binder, FragmentSearch.BondsPerSPList[1]);
4094
4095 // do a BFS search to fill the SP lists and label the found vertices
4096 // Actually, we should construct a spanning tree vom the root atom and select all edges therefrom and put them into
4097 // according shortest path lists. However, we don't. Rather we fill these lists right away, as they do form a spanning
4098 // tree already sorted into various SP levels. That's why we just do loops over the depth (CurrentSP) and breadth
4099 // (EdgeinSPLevel) of this tree ...
4100 // In another picture, the bonds always contain a direction by rightatom being the one more distant from root and hence
4101 // naturally leftatom forming its predecessor, preventing the BFS"seeker" from continuing in the wrong direction.
4102 *out << endl;
4103 *out << Verbose(0) << "Starting BFS analysis ..." << endl;
4104 for (SP = 0; SP < (Order-1); SP++) {
4105 *out << Verbose(1) << "New SP level reached: " << SP << ", creating new SP list with " << FragmentSearch.BondsPerSPCount[SP] << " item(s)";
4106 if (SP > 0) {
4107 *out << ", old level closed with " << FragmentSearch.BondsPerSPCount[SP-1] << " item(s)." << endl;
4108 FragmentSearch.BondsPerSPCount[SP] = 0;
4109 } else
4110 *out << "." << endl;
4111
4112 RemainingWalkers = FragmentSearch.BondsPerSPCount[SP];
4113 CurrentEdge = FragmentSearch.BondsPerSPList[2*SP]; /// start of this SP level's list
4114 while (CurrentEdge->next != FragmentSearch.BondsPerSPList[2*SP+1]) { /// end of this SP level's list
4115 CurrentEdge = CurrentEdge->next;
4116 RemainingWalkers--;
4117 Walker = CurrentEdge->rightatom; // rightatom is always the one more distant
4118 Predecessor = CurrentEdge->leftatom; // ... and leftatom is predecessor
4119 AtomKeyNr = Walker->nr;
4120 *out << Verbose(0) << "Current Walker is: " << *Walker << " with nr " << Walker->nr << " and SP of " << SP << ", with " << RemainingWalkers << " remaining walkers on this level." << endl;
4121 // check for new sp level
4122 // go through all its bonds
4123 *out << Verbose(1) << "Going through all bonds of Walker." << endl;
4124 for (int i=0;i<NumberOfBondsPerAtom[AtomKeyNr];i++) {
4125 Binder = ListOfBondsPerAtom[AtomKeyNr][i];
4126 OtherWalker = Binder->GetOtherAtom(Walker);
4127 if ((RestrictedKeySet.find(OtherWalker->nr) != RestrictedKeySet.end())
4128 #ifdef ADDHYDROGEN
4129 && (OtherWalker->type->Z != 1)
4130 #endif
4131 ) { // skip hydrogens and restrict to fragment
4132 *out << Verbose(2) << "Current partner is " << *OtherWalker << " with nr " << OtherWalker->nr << " in bond " << *Binder << "." << endl;
4133 // set the label if not set (and push on root stack as well)
4134 if ((OtherWalker != Predecessor) && (OtherWalker->GetTrueFather()->nr > RootKeyNr)) { // only pass through those with label bigger than Root's
4135 FragmentSearch.ShortestPathList[OtherWalker->nr] = SP+1;
4136 *out << Verbose(3) << "Set Shortest Path to " << FragmentSearch.ShortestPathList[OtherWalker->nr] << "." << endl;
4137 // add the bond in between to the SP list
4138 Binder = new bond(Walker, OtherWalker); // create a new bond in such a manner, that bond::rightatom is always the one more distant
4139 add(Binder, FragmentSearch.BondsPerSPList[2*(SP+1)+1]);
4140 FragmentSearch.BondsPerSPCount[SP+1]++;
4141 *out << Verbose(3) << "Added its bond to SP list, having now " << FragmentSearch.BondsPerSPCount[SP+1] << " item(s)." << endl;
4142 } else {
4143 if (OtherWalker != Predecessor)
4144 *out << Verbose(3) << "Not passing on, as index of " << *OtherWalker << " " << OtherWalker->GetTrueFather()->nr << " is smaller than that of Root " << RootKeyNr << "." << endl;
4145 else
4146 *out << Verbose(3) << "This is my predecessor " << *Predecessor << "." << endl;
4147 }
4148 } else *out << Verbose(2) << "Is not in the restricted keyset or skipping hydrogen " << *OtherWalker << "." << endl;
4149 }
4150 }
4151 }
4152
4153 // outputting all list for debugging
4154 *out << Verbose(0) << "Printing all found lists." << endl;
4155 for(int i=1;i<Order;i++) { // skip the root edge in the printing
4156 Binder = FragmentSearch.BondsPerSPList[2*i];
4157 *out << Verbose(1) << "Current SP level is " << i << "." << endl;
4158 while (Binder->next != FragmentSearch.BondsPerSPList[2*i+1]) {
4159 Binder = Binder->next;
4160 *out << Verbose(2) << *Binder << endl;
4161 }
4162 }
4163
4164 // creating fragments with the found edge sets (may be done in reverse order, faster)
4165 SP = -1; // the Root <-> Root edge must be subtracted!
4166 for(int i=Order;i--;) { // sum up all found edges
4167 Binder = FragmentSearch.BondsPerSPList[2*i];
4168 while (Binder->next != FragmentSearch.BondsPerSPList[2*i+1]) {
4169 Binder = Binder->next;
4170 SP ++;
4171 }
4172 }
4173 *out << Verbose(0) << "Total number of edges is " << SP << "." << endl;
4174 if (SP >= (Order-1)) {
4175 // start with root (push on fragment stack)
4176 *out << Verbose(0) << "Starting fragment generation with " << *FragmentSearch.Root << ", local nr is " << FragmentSearch.Root->nr << "." << endl;
4177 FragmentSearch.FragmentSet->clear();
4178 *out << Verbose(0) << "Preparing subset for this root and calling generator." << endl;
4179 // prepare the subset and call the generator
4180 BondsList = (bond **) Malloc(sizeof(bond *)*FragmentSearch.BondsPerSPCount[0], "molecule::PowerSetGenerator: **BondsList");
4181 BondsList[0] = FragmentSearch.BondsPerSPList[0]->next; // on SP level 0 there's only the root bond
4182
4183 SPFragmentGenerator(out, &FragmentSearch, 0, BondsList, FragmentSearch.BondsPerSPCount[0], Order);
4184
4185 Free((void **)&BondsList, "molecule::PowerSetGenerator: **BondsList");
4186 } else {
4187 *out << Verbose(0) << "Not enough total number of edges to build " << Order << "-body fragments." << endl;
4188 }
4189
4190 // as FragmentSearch structure is used only once, we don't have to clean it anymore
4191 // remove root from stack
4192 *out << Verbose(0) << "Removing root again from stack." << endl;
4193 FragmentSearch.FragmentSet->erase(FragmentSearch.Root->nr);
4194
4195 // free'ing the bonds lists
4196 *out << Verbose(0) << "Free'ing all found lists. and resetting index lists" << endl;
4197 for(int i=Order;i--;) {
4198 *out << Verbose(1) << "Current SP level is " << i << ": ";
4199 Binder = FragmentSearch.BondsPerSPList[2*i];
4200 while (Binder->next != FragmentSearch.BondsPerSPList[2*i+1]) {
4201 Binder = Binder->next;
4202 // *out << "Removing atom " << Binder->leftatom->nr << " and " << Binder->rightatom->nr << "." << endl; // make sure numbers are local
4203 FragmentSearch.ShortestPathList[Binder->leftatom->nr] = -1;
4204 FragmentSearch.ShortestPathList[Binder->rightatom->nr] = -1;
4205 }
4206 // delete added bonds
4207 cleanup(FragmentSearch.BondsPerSPList[2*i], FragmentSearch.BondsPerSPList[2*i+1]);
4208 // also start and end node
4209 *out << "cleaned." << endl;
4210 }
4211
4212 // return list
4213 *out << Verbose(0) << "End of PowerSetGenerator." << endl;
4214 return (FragmentSearch.FragmentCounter - Counter);
4215};
4216
4217/** Corrects the nuclei position if the fragment was created over the cell borders.
4218 * Scans all bonds, checks the distance, if greater than typical, we have a candidate for the correction.
4219 * We remove the bond whereafter the graph probably separates. Then, we translate the one component periodically
4220 * and re-add the bond. Looping on the distance check.
4221 * \param *out ofstream for debugging messages
4222 */
4223void molecule::ScanForPeriodicCorrection(ofstream *out)
4224{
4225 bond *Binder = NULL;
4226 bond *OtherBinder = NULL;
4227 atom *Walker = NULL;
4228 atom *OtherWalker = NULL;
4229 double *matrix = ReturnFullMatrixforSymmetric(cell_size);
4230 enum Shading *ColorList = NULL;
4231 double tmp;
4232 Vector Translationvector;
4233 //class StackClass<atom *> *CompStack = NULL;
4234 class StackClass<atom *> *AtomStack = new StackClass<atom *>(AtomCount);
4235 bool flag = true;
4236
4237 *out << Verbose(2) << "Begin of ScanForPeriodicCorrection." << endl;
4238
4239 ColorList = (enum Shading *) Malloc(sizeof(enum Shading)*AtomCount, "molecule::ScanForPeriodicCorrection: *ColorList");
4240 while (flag) {
4241 // remove bonds that are beyond bonddistance
4242 for(int i=NDIM;i--;)
4243 Translationvector.x[i] = 0.;
4244 // scan all bonds
4245 Binder = first;
4246 flag = false;
4247 while ((!flag) && (Binder->next != last)) {
4248 Binder = Binder->next;
4249 for (int i=NDIM;i--;) {
4250 tmp = fabs(Binder->leftatom->x.x[i] - Binder->rightatom->x.x[i]);
4251 //*out << Verbose(3) << "Checking " << i << "th distance of " << *Binder->leftatom << " to " << *Binder->rightatom << ": " << tmp << "." << endl;
4252 if (tmp > BondDistance) {
4253 OtherBinder = Binder->next; // note down binding partner for later re-insertion
4254 unlink(Binder); // unlink bond
4255 *out << Verbose(2) << "Correcting at bond " << *Binder << "." << endl;
4256 flag = true;
4257 break;
4258 }
4259 }
4260 }
4261 if (flag) {
4262 // create translation vector from their periodically modified distance
4263 for (int i=NDIM;i--;) {
4264 tmp = Binder->leftatom->x.x[i] - Binder->rightatom->x.x[i];
4265 if (fabs(tmp) > BondDistance)
4266 Translationvector.x[i] = (tmp < 0) ? +1. : -1.;
4267 }
4268 Translationvector.MatrixMultiplication(matrix);
4269 //*out << Verbose(3) << "Translation vector is ";
4270 Translationvector.Output(out);
4271 *out << endl;
4272 // apply to all atoms of first component via BFS
4273 for (int i=AtomCount;i--;)
4274 ColorList[i] = white;
4275 AtomStack->Push(Binder->leftatom);
4276 while (!AtomStack->IsEmpty()) {
4277 Walker = AtomStack->PopFirst();
4278 //*out << Verbose (3) << "Current Walker is: " << *Walker << "." << endl;
4279 ColorList[Walker->nr] = black; // mark as explored
4280 Walker->x.AddVector(&Translationvector); // translate
4281 for (int i=0;i<NumberOfBondsPerAtom[Walker->nr];i++) { // go through all binding partners
4282 if (ListOfBondsPerAtom[Walker->nr][i] != Binder) {
4283 OtherWalker = ListOfBondsPerAtom[Walker->nr][i]->GetOtherAtom(Walker);
4284 if (ColorList[OtherWalker->nr] == white) {
4285 AtomStack->Push(OtherWalker); // push if yet unexplored
4286 }
4287 }
4288 }
4289 }
4290 // re-add bond
4291 link(Binder, OtherBinder);
4292 } else {
4293 *out << Verbose(3) << "No corrections for this fragment." << endl;
4294 }
4295 //delete(CompStack);
4296 }
4297
4298 // free allocated space from ReturnFullMatrixforSymmetric()
4299 delete(AtomStack);
4300 Free((void **)&ColorList, "molecule::ScanForPeriodicCorrection: *ColorList");
4301 Free((void **)&matrix, "molecule::ScanForPeriodicCorrection: *matrix");
4302 *out << Verbose(2) << "End of ScanForPeriodicCorrection." << endl;
4303};
4304
4305/** Blows the 6-dimensional \a cell_size array up to a full NDIM by NDIM matrix.
4306 * \param *symm 6-dim array of unique symmetric matrix components
4307 * \return allocated NDIM*NDIM array with the symmetric matrix
4308 */
4309double * molecule::ReturnFullMatrixforSymmetric(double *symm)
4310{
4311 double *matrix = (double *) Malloc(sizeof(double)*NDIM*NDIM, "molecule::ReturnFullMatrixforSymmetric: *matrix");
4312 matrix[0] = symm[0];
4313 matrix[1] = symm[1];
4314 matrix[2] = symm[3];
4315 matrix[3] = symm[1];
4316 matrix[4] = symm[2];
4317 matrix[5] = symm[4];
4318 matrix[6] = symm[3];
4319 matrix[7] = symm[4];
4320 matrix[8] = symm[5];
4321 return matrix;
4322};
4323
4324bool KeyCompare::operator() (const KeySet SubgraphA, const KeySet SubgraphB) const
4325{
4326 //cout << "my check is used." << endl;
4327 if (SubgraphA.size() < SubgraphB.size()) {
4328 return true;
4329 } else {
4330 if (SubgraphA.size() > SubgraphB.size()) {
4331 return false;
4332 } else {
4333 KeySet::iterator IteratorA = SubgraphA.begin();
4334 KeySet::iterator IteratorB = SubgraphB.begin();
4335 while ((IteratorA != SubgraphA.end()) && (IteratorB != SubgraphB.end())) {
4336 if ((*IteratorA) < (*IteratorB))
4337 return true;
4338 else if ((*IteratorA) > (*IteratorB)) {
4339 return false;
4340 } // else, go on to next index
4341 IteratorA++;
4342 IteratorB++;
4343 } // end of while loop
4344 }// end of check in case of equal sizes
4345 }
4346 return false; // if we reach this point, they are equal
4347};
4348
4349//bool operator < (KeySet SubgraphA, KeySet SubgraphB)
4350//{
4351// return KeyCompare(SubgraphA, SubgraphB);
4352//};
4353
4354/** Checking whether KeySet is not already present in Graph, if so just adds factor.
4355 * \param *out output stream for debugging
4356 * \param &set KeySet to insert
4357 * \param &graph Graph to insert into
4358 * \param *counter pointer to unique fragment count
4359 * \param factor energy factor for the fragment
4360 */
4361inline void InsertFragmentIntoGraph(ofstream *out, struct UniqueFragments *Fragment)
4362{
4363 GraphTestPair testGraphInsert;
4364
4365 testGraphInsert = Fragment->Leaflet->insert(GraphPair (*Fragment->FragmentSet,pair<int,double>(Fragment->FragmentCounter,Fragment->TEFactor))); // store fragment number and current factor
4366 if (testGraphInsert.second) {
4367 *out << Verbose(2) << "KeySet " << Fragment->FragmentCounter << " successfully inserted." << endl;
4368 Fragment->FragmentCounter++;
4369 } else {
4370 *out << Verbose(2) << "KeySet " << Fragment->FragmentCounter << " failed to insert, present fragment is " << ((*(testGraphInsert.first)).second).first << endl;
4371 ((*(testGraphInsert.first)).second).second += Fragment->TEFactor; // increase the "created" counter
4372 *out << Verbose(2) << "New factor is " << ((*(testGraphInsert.first)).second).second << "." << endl;
4373 }
4374};
4375//void inline InsertIntoGraph(ofstream *out, KeyStack &stack, Graph &graph, int *counter, double factor)
4376//{
4377// // copy stack contents to set and call overloaded function again
4378// KeySet set;
4379// for(KeyStack::iterator runner = stack.begin(); runner != stack.begin(); runner++)
4380// set.insert((*runner));
4381// InsertIntoGraph(out, set, graph, counter, factor);
4382//};
4383
4384/** Inserts each KeySet in \a graph2 into \a graph1.
4385 * \param *out output stream for debugging
4386 * \param graph1 first (dest) graph
4387 * \param graph2 second (source) graph
4388 * \param *counter keyset counter that gets increased
4389 */
4390inline void InsertGraphIntoGraph(ofstream *out, Graph &graph1, Graph &graph2, int *counter)
4391{
4392 GraphTestPair testGraphInsert;
4393
4394 for(Graph::iterator runner = graph2.begin(); runner != graph2.end(); runner++) {
4395 testGraphInsert = graph1.insert(GraphPair ((*runner).first,pair<int,double>((*counter)++,((*runner).second).second))); // store fragment number and current factor
4396 if (testGraphInsert.second) {
4397 *out << Verbose(2) << "KeySet " << (*counter)-1 << " successfully inserted." << endl;
4398 } else {
4399 *out << Verbose(2) << "KeySet " << (*counter)-1 << " failed to insert, present fragment is " << ((*(testGraphInsert.first)).second).first << endl;
4400 ((*(testGraphInsert.first)).second).second += (*runner).second.second;
4401 *out << Verbose(2) << "New factor is " << (*(testGraphInsert.first)).second.second << "." << endl;
4402 }
4403 }
4404};
4405
4406
4407/** Performs BOSSANOVA decomposition at selected sites, increasing the cutoff by one at these sites.
4408 * -# constructs a complete keyset of the molecule
4409 * -# In a loop over all possible roots from the given rootstack
4410 * -# increases order of root site
4411 * -# calls PowerSetGenerator with this order, the complete keyset and the rootkeynr
4412 * -# for all consecutive lower levels PowerSetGenerator is called with the suborder, the higher order keyset
4413as the restricted one and each site in the set as the root)
4414 * -# these are merged into a fragment list of keysets
4415 * -# All fragment lists (for all orders, i.e. from all destination fields) are merged into one list for return
4416 * Important only is that we create all fragments, it is not important if we create them more than once
4417 * as these copies are filtered out via use of the hash table (KeySet).
4418 * \param *out output stream for debugging
4419 * \param Fragment&*List list of already present keystacks (adaptive scheme) or empty list
4420 * \param &RootStack stack with all root candidates (unequal to each atom in complete molecule if adaptive scheme is applied)
4421 * \param *MinimumRingSize minimum ring size for each atom (molecule::Atomcount)
4422 * \return pointer to Graph list
4423 */
4424void molecule::FragmentBOSSANOVA(ofstream *out, Graph *&FragmentList, KeyStack &RootStack, int *MinimumRingSize)
4425{
4426 Graph ***FragmentLowerOrdersList = NULL;
4427 int NumLevels, NumMolecules, TotalNumMolecules = 0, *NumMoleculesOfOrder = NULL;
4428 int counter = 0, Order;
4429 int UpgradeCount = RootStack.size();
4430 KeyStack FragmentRootStack;
4431 int RootKeyNr, RootNr;
4432 struct UniqueFragments FragmentSearch;
4433
4434 *out << Verbose(0) << "Begin of FragmentBOSSANOVA." << endl;
4435
4436 // FragmentLowerOrdersList is a 2D-array of pointer to MoleculeListClass objects, one dimension represents the ANOVA expansion of a single order (i.e. 5)
4437 // with all needed lower orders that are subtracted, the other dimension is the BondOrder (i.e. from 1 to 5)
4438 NumMoleculesOfOrder = (int *) Malloc(sizeof(int)*UpgradeCount, "molecule::FragmentBOSSANOVA: *NumMoleculesOfOrder");
4439 FragmentLowerOrdersList = (Graph ***) Malloc(sizeof(Graph **)*UpgradeCount, "molecule::FragmentBOSSANOVA: ***FragmentLowerOrdersList");
4440
4441 // initialise the fragments structure
4442 FragmentSearch.ShortestPathList = (int *) Malloc(sizeof(int)*AtomCount, "molecule::PowerSetGenerator: *ShortestPathList");
4443 FragmentSearch.FragmentCounter = 0;
4444 FragmentSearch.FragmentSet = new KeySet;
4445 FragmentSearch.Root = FindAtom(RootKeyNr);
4446 for (int i=AtomCount;i--;) {
4447 FragmentSearch.ShortestPathList[i] = -1;
4448 }
4449
4450 // Construct the complete KeySet which we need for topmost level only (but for all Roots)
4451 atom *Walker = start;
4452 KeySet CompleteMolecule;
4453 while (Walker->next != end) {
4454 Walker = Walker->next;
4455 CompleteMolecule.insert(Walker->GetTrueFather()->nr);
4456 }
4457
4458 // this can easily be seen: if Order is 5, then the number of levels for each lower order is the total sum of the number of levels above, as
4459 // each has to be split up. E.g. for the second level we have one from 5th, one from 4th, two from 3th (which in turn is one from 5th, one from 4th),
4460 // hence we have overall four 2th order levels for splitting. This also allows for putting all into a single array (FragmentLowerOrdersList[])
4461 // with the order along the cells as this: 5433222211111111 for BondOrder 5 needing 16=pow(2,5-1) cells (only we use bit-shifting which is faster)
4462 RootNr = 0; // counts through the roots in RootStack
4463 while ((RootNr < UpgradeCount) && (!RootStack.empty())) {
4464 RootKeyNr = RootStack.front();
4465 RootStack.pop_front();
4466 Walker = FindAtom(RootKeyNr);
4467 // check cyclic lengths
4468 //if ((MinimumRingSize[Walker->GetTrueFather()->nr] != -1) && (Walker->GetTrueFather()->AdaptiveOrder+1 > MinimumRingSize[Walker->GetTrueFather()->nr])) {
4469 // *out << Verbose(0) << "Bond order " << Walker->GetTrueFather()->AdaptiveOrder << " of Root " << *Walker << " greater than or equal to Minimum Ring size of " << MinimumRingSize << " found is not allowed." << endl;
4470 //} else
4471 {
4472 // increase adaptive order by one
4473 Walker->GetTrueFather()->AdaptiveOrder++;
4474 Order = Walker->AdaptiveOrder = Walker->GetTrueFather()->AdaptiveOrder;
4475
4476 // initialise Order-dependent entries of UniqueFragments structure
4477 FragmentSearch.BondsPerSPList = (bond **) Malloc(sizeof(bond *)*Order*2, "molecule::PowerSetGenerator: ***BondsPerSPList");
4478 FragmentSearch.BondsPerSPCount = (int *) Malloc(sizeof(int)*Order, "molecule::PowerSetGenerator: *BondsPerSPCount");
4479 for (int i=Order;i--;) {
4480 FragmentSearch.BondsPerSPList[2*i] = new bond(); // start node
4481 FragmentSearch.BondsPerSPList[2*i+1] = new bond(); // end node
4482 FragmentSearch.BondsPerSPList[2*i]->next = FragmentSearch.BondsPerSPList[2*i+1]; // intertwine these two
4483 FragmentSearch.BondsPerSPList[2*i+1]->previous = FragmentSearch.BondsPerSPList[2*i];
4484 FragmentSearch.BondsPerSPCount[i] = 0;
4485 }
4486
4487 // allocate memory for all lower level orders in this 1D-array of ptrs
4488 NumLevels = 1 << (Order-1); // (int)pow(2,Order);
4489 FragmentLowerOrdersList[RootNr] = (Graph **) Malloc(sizeof(Graph *)*NumLevels, "molecule::FragmentBOSSANOVA: **FragmentLowerOrdersList[]");
4490 for (int i=0;i<NumLevels;i++)
4491 FragmentLowerOrdersList[RootNr][i] = NULL;
4492
4493 // create top order where nothing is reduced
4494 *out << Verbose(0) << "==============================================================================================================" << endl;
4495 *out << Verbose(0) << "Creating KeySets of Bond Order " << Order << " for " << *Walker << ", " << (RootStack.size()-RootNr) << " Roots remaining." << endl; // , NumLevels is " << NumLevels << "
4496
4497 // Create list of Graphs of current Bond Order (i.e. F_{ij})
4498 FragmentLowerOrdersList[RootNr][0] = new Graph;
4499 FragmentSearch.TEFactor = 1.;
4500 FragmentSearch.Leaflet = FragmentLowerOrdersList[RootNr][0]; // set to insertion graph
4501 FragmentSearch.Root = Walker;
4502 NumMoleculesOfOrder[RootNr] = PowerSetGenerator(out, Walker->AdaptiveOrder, FragmentSearch, CompleteMolecule);
4503 *out << Verbose(1) << "Number of resulting KeySets is: " << NumMoleculesOfOrder[RootNr] << "." << endl;
4504 if (NumMoleculesOfOrder[RootNr] != 0) {
4505 NumMolecules = 0;
4506
4507 // we don't have to dive into suborders! These keysets are all already created on lower orders!
4508 // this was all ancient stuff, when we still depended on the TEFactors (and for those the suborders were needed)
4509
4510// if ((NumLevels >> 1) > 0) {
4511// // create lower order fragments
4512// *out << Verbose(0) << "Creating list of unique fragments of lower Bond Order terms to be subtracted." << endl;
4513// Order = Walker->AdaptiveOrder;
4514// for (int source=0;source<(NumLevels >> 1);source++) { // 1-terms don't need any more splitting, that's why only half is gone through (shift again)
4515// // step down to next order at (virtual) boundary of powers of 2 in array
4516// while (source >= (1 << (Walker->AdaptiveOrder-Order))) // (int)pow(2,Walker->AdaptiveOrder-Order))
4517// Order--;
4518// *out << Verbose(0) << "Current Order is: " << Order << "." << endl;
4519// for (int SubOrder=Order-1;SubOrder>0;SubOrder--) {
4520// int dest = source + (1 << (Walker->AdaptiveOrder-(SubOrder+1)));
4521// *out << Verbose(0) << "--------------------------------------------------------------------------------------------------------------" << endl;
4522// *out << Verbose(0) << "Current SubOrder is: " << SubOrder << " with source " << source << " to destination " << dest << "." << endl;
4523//
4524// // every molecule is split into a list of again (Order - 1) molecules, while counting all molecules
4525// //*out << Verbose(1) << "Splitting the " << (*FragmentLowerOrdersList[RootNr][source]).size() << " molecules of the " << source << "th cell in the array." << endl;
4526// //NumMolecules = 0;
4527// FragmentLowerOrdersList[RootNr][dest] = new Graph;
4528// for(Graph::iterator runner = (*FragmentLowerOrdersList[RootNr][source]).begin();runner != (*FragmentLowerOrdersList[RootNr][source]).end(); runner++) {
4529// for (KeySet::iterator sprinter = (*runner).first.begin();sprinter != (*runner).first.end(); sprinter++) {
4530// Graph TempFragmentList;
4531// FragmentSearch.TEFactor = -(*runner).second.second;
4532// FragmentSearch.Leaflet = &TempFragmentList; // set to insertion graph
4533// FragmentSearch.Root = FindAtom(*sprinter);
4534// NumMoleculesOfOrder[RootNr] += PowerSetGenerator(out, SubOrder, FragmentSearch, (*runner).first);
4535// // insert new keysets FragmentList into FragmentLowerOrdersList[Walker->AdaptiveOrder-1][dest]
4536// *out << Verbose(1) << "Merging resulting key sets with those present in destination " << dest << "." << endl;
4537// InsertGraphIntoGraph(out, *FragmentLowerOrdersList[RootNr][dest], TempFragmentList, &NumMolecules);
4538// }
4539// }
4540// *out << Verbose(1) << "Number of resulting molecules for SubOrder " << SubOrder << " is: " << NumMolecules << "." << endl;
4541// }
4542// }
4543// }
4544 } else {
4545 Walker->GetTrueFather()->MaxOrder = true;
4546// *out << Verbose(1) << "Hence, we don't dive into SubOrders ... " << endl;
4547 }
4548 // now, we have completely filled each cell of FragmentLowerOrdersList[] for the current Walker->AdaptiveOrder
4549 //NumMoleculesOfOrder[Walker->AdaptiveOrder-1] = NumMolecules;
4550 TotalNumMolecules += NumMoleculesOfOrder[RootNr];
4551// *out << Verbose(1) << "Number of resulting molecules for Order " << (int)Walker->GetTrueFather()->AdaptiveOrder << " is: " << NumMoleculesOfOrder[RootNr] << "." << endl;
4552 RootStack.push_back(RootKeyNr); // put back on stack
4553 RootNr++;
4554
4555 // free Order-dependent entries of UniqueFragments structure for next loop cycle
4556 Free((void **)&FragmentSearch.BondsPerSPCount, "molecule::PowerSetGenerator: *BondsPerSPCount");
4557 for (int i=Order;i--;) {
4558 delete(FragmentSearch.BondsPerSPList[2*i]);
4559 delete(FragmentSearch.BondsPerSPList[2*i+1]);
4560 }
4561 Free((void **)&FragmentSearch.BondsPerSPList, "molecule::PowerSetGenerator: ***BondsPerSPList");
4562 }
4563 }
4564 *out << Verbose(0) << "==============================================================================================================" << endl;
4565 *out << Verbose(1) << "Total number of resulting molecules is: " << TotalNumMolecules << "." << endl;
4566 *out << Verbose(0) << "==============================================================================================================" << endl;
4567
4568 // cleanup FragmentSearch structure
4569 Free((void **)&FragmentSearch.ShortestPathList, "molecule::PowerSetGenerator: *ShortestPathList");
4570 delete(FragmentSearch.FragmentSet);
4571
4572 // now, FragmentLowerOrdersList is complete, it looks - for BondOrder 5 - as this (number is the ANOVA Order of the terms therein)
4573 // 5433222211111111
4574 // 43221111
4575 // 3211
4576 // 21
4577 // 1
4578
4579 // Subsequently, we combine all into a single list (FragmentList)
4580
4581 *out << Verbose(0) << "Combining the lists of all orders per order and finally into a single one." << endl;
4582 if (FragmentList == NULL) {
4583 FragmentList = new Graph;
4584 counter = 0;
4585 } else {
4586 counter = FragmentList->size();
4587 }
4588 RootNr = 0;
4589 while (!RootStack.empty()) {
4590 RootKeyNr = RootStack.front();
4591 RootStack.pop_front();
4592 Walker = FindAtom(RootKeyNr);
4593 NumLevels = 1 << (Walker->AdaptiveOrder - 1);
4594 for(int i=0;i<NumLevels;i++) {
4595 if (FragmentLowerOrdersList[RootNr][i] != NULL) {
4596 InsertGraphIntoGraph(out, *FragmentList, (*FragmentLowerOrdersList[RootNr][i]), &counter);
4597 delete(FragmentLowerOrdersList[RootNr][i]);
4598 }
4599 }
4600 Free((void **)&FragmentLowerOrdersList[RootNr], "molecule::FragmentBOSSANOVA: **FragmentLowerOrdersList[]");
4601 RootNr++;
4602 }
4603 Free((void **)&FragmentLowerOrdersList, "molecule::FragmentBOSSANOVA: ***FragmentLowerOrdersList");
4604 Free((void **)&NumMoleculesOfOrder, "molecule::FragmentBOSSANOVA: *NumMoleculesOfOrder");
4605
4606 *out << Verbose(0) << "End of FragmentBOSSANOVA." << endl;
4607};
4608
4609/** Comparison function for GSL heapsort on distances in two molecules.
4610 * \param *a
4611 * \param *b
4612 * \return <0, \a *a less than \a *b, ==0 if equal, >0 \a *a greater than \a *b
4613 */
4614inline int CompareDoubles (const void * a, const void * b)
4615{
4616 if (*(double *)a > *(double *)b)
4617 return -1;
4618 else if (*(double *)a < *(double *)b)
4619 return 1;
4620 else
4621 return 0;
4622};
4623
4624/** Determines whether two molecules actually contain the same atoms and coordination.
4625 * \param *out output stream for debugging
4626 * \param *OtherMolecule the molecule to compare this one to
4627 * \param threshold upper limit of difference when comparing the coordination.
4628 * \return NULL - not equal, otherwise an allocated (molecule::AtomCount) permutation map of the atom numbers (which corresponds to which)
4629 */
4630int * molecule::IsEqualToWithinThreshold(ofstream *out, molecule *OtherMolecule, double threshold)
4631{
4632 int flag;
4633 double *Distances = NULL, *OtherDistances = NULL;
4634 Vector CenterOfGravity, OtherCenterOfGravity;
4635 size_t *PermMap = NULL, *OtherPermMap = NULL;
4636 int *PermutationMap = NULL;
4637 atom *Walker = NULL;
4638 bool result = true; // status of comparison
4639
4640 *out << Verbose(3) << "Begin of IsEqualToWithinThreshold." << endl;
4641 /// first count both their atoms and elements and update lists thereby ...
4642 //*out << Verbose(0) << "Counting atoms, updating list" << endl;
4643 CountAtoms(out);
4644 OtherMolecule->CountAtoms(out);
4645 CountElements();
4646 OtherMolecule->CountElements();
4647
4648 /// ... and compare:
4649 /// -# AtomCount
4650 if (result) {
4651 if (AtomCount != OtherMolecule->AtomCount) {
4652 *out << Verbose(4) << "AtomCounts don't match: " << AtomCount << " == " << OtherMolecule->AtomCount << endl;
4653 result = false;
4654 } else *out << Verbose(4) << "AtomCounts match: " << AtomCount << " == " << OtherMolecule->AtomCount << endl;
4655 }
4656 /// -# ElementCount
4657 if (result) {
4658 if (ElementCount != OtherMolecule->ElementCount) {
4659 *out << Verbose(4) << "ElementCount don't match: " << ElementCount << " == " << OtherMolecule->ElementCount << endl;
4660 result = false;
4661 } else *out << Verbose(4) << "ElementCount match: " << ElementCount << " == " << OtherMolecule->ElementCount << endl;
4662 }
4663 /// -# ElementsInMolecule
4664 if (result) {
4665 for (flag=MAX_ELEMENTS;flag--;) {
4666 //*out << Verbose(5) << "Element " << flag << ": " << ElementsInMolecule[flag] << " <-> " << OtherMolecule->ElementsInMolecule[flag] << "." << endl;
4667 if (ElementsInMolecule[flag] != OtherMolecule->ElementsInMolecule[flag])
4668 break;
4669 }
4670 if (flag < MAX_ELEMENTS) {
4671 *out << Verbose(4) << "ElementsInMolecule don't match." << endl;
4672 result = false;
4673 } else *out << Verbose(4) << "ElementsInMolecule match." << endl;
4674 }
4675 /// then determine and compare center of gravity for each molecule ...
4676 if (result) {
4677 *out << Verbose(5) << "Calculating Centers of Gravity" << endl;
4678 DetermineCenter(CenterOfGravity);
4679 OtherMolecule->DetermineCenter(OtherCenterOfGravity);
4680 *out << Verbose(5) << "Center of Gravity: ";
4681 CenterOfGravity.Output(out);
4682 *out << endl << Verbose(5) << "Other Center of Gravity: ";
4683 OtherCenterOfGravity.Output(out);
4684 *out << endl;
4685 if (CenterOfGravity.Distance(&OtherCenterOfGravity) > threshold) {
4686 *out << Verbose(4) << "Centers of gravity don't match." << endl;
4687 result = false;
4688 }
4689 }
4690
4691 /// ... then make a list with the euclidian distance to this center for each atom of both molecules
4692 if (result) {
4693 *out << Verbose(5) << "Calculating distances" << endl;
4694 Distances = (double *) Malloc(sizeof(double)*AtomCount, "molecule::IsEqualToWithinThreshold: Distances");
4695 OtherDistances = (double *) Malloc(sizeof(double)*AtomCount, "molecule::IsEqualToWithinThreshold: OtherDistances");
4696 Walker = start;
4697 while (Walker->next != end) {
4698 Walker = Walker->next;
4699 Distances[Walker->nr] = CenterOfGravity.Distance(&Walker->x);
4700 }
4701 Walker = OtherMolecule->start;
4702 while (Walker->next != OtherMolecule->end) {
4703 Walker = Walker->next;
4704 OtherDistances[Walker->nr] = OtherCenterOfGravity.Distance(&Walker->x);
4705 }
4706
4707 /// ... sort each list (using heapsort (o(N log N)) from GSL)
4708 *out << Verbose(5) << "Sorting distances" << endl;
4709 PermMap = (size_t *) Malloc(sizeof(size_t)*AtomCount, "molecule::IsEqualToWithinThreshold: *PermMap");
4710 OtherPermMap = (size_t *) Malloc(sizeof(size_t)*AtomCount, "molecule::IsEqualToWithinThreshold: *OtherPermMap");
4711 gsl_heapsort_index (PermMap, Distances, AtomCount, sizeof(double), CompareDoubles);
4712 gsl_heapsort_index (OtherPermMap, OtherDistances, AtomCount, sizeof(double), CompareDoubles);
4713 PermutationMap = (int *) Malloc(sizeof(int)*AtomCount, "molecule::IsEqualToWithinThreshold: *PermutationMap");
4714 *out << Verbose(5) << "Combining Permutation Maps" << endl;
4715 for(int i=AtomCount;i--;)
4716 PermutationMap[PermMap[i]] = (int) OtherPermMap[i];
4717
4718 /// ... and compare them step by step, whether the difference is individiually(!) below \a threshold for all
4719 *out << Verbose(4) << "Comparing distances" << endl;
4720 flag = 0;
4721 for (int i=0;i<AtomCount;i++) {
4722 *out << Verbose(5) << "Distances: |" << Distances[PermMap[i]] << " - " << OtherDistances[OtherPermMap[i]] << "| = " << fabs(Distances[PermMap[i]] - OtherDistances[OtherPermMap[i]]) << " ?<? " << threshold << endl;
4723 if (fabs(Distances[PermMap[i]] - OtherDistances[OtherPermMap[i]]) > threshold)
4724 flag = 1;
4725 }
4726 Free((void **)&PermMap, "molecule::IsEqualToWithinThreshold: *PermMap");
4727 Free((void **)&OtherPermMap, "molecule::IsEqualToWithinThreshold: *OtherPermMap");
4728
4729 /// free memory
4730 Free((void **)&Distances, "molecule::IsEqualToWithinThreshold: Distances");
4731 Free((void **)&OtherDistances, "molecule::IsEqualToWithinThreshold: OtherDistances");
4732 if (flag) { // if not equal
4733 Free((void **)&PermutationMap, "molecule::IsEqualToWithinThreshold: *PermutationMap");
4734 result = false;
4735 }
4736 }
4737 /// return pointer to map if all distances were below \a threshold
4738 *out << Verbose(3) << "End of IsEqualToWithinThreshold." << endl;
4739 if (result) {
4740 *out << Verbose(3) << "Result: Equal." << endl;
4741 return PermutationMap;
4742 } else {
4743 *out << Verbose(3) << "Result: Not equal." << endl;
4744 return NULL;
4745 }
4746};
4747
4748/** Returns an index map for two father-son-molecules.
4749 * The map tells which atom in this molecule corresponds to which one in the other molecul with their fathers.
4750 * \param *out output stream for debugging
4751 * \param *OtherMolecule corresponding molecule with fathers
4752 * \return allocated map of size molecule::AtomCount with map
4753 * \todo make this with a good sort O(n), not O(n^2)
4754 */
4755int * molecule::GetFatherSonAtomicMap(ofstream *out, molecule *OtherMolecule)
4756{
4757 atom *Walker = NULL, *OtherWalker = NULL;
4758 *out << Verbose(3) << "Begin of GetFatherAtomicMap." << endl;
4759 int *AtomicMap = (int *) Malloc(sizeof(int)*AtomCount, "molecule::GetAtomicMap: *AtomicMap"); //Calloc
4760 for (int i=AtomCount;i--;)
4761 AtomicMap[i] = -1;
4762 if (OtherMolecule == this) { // same molecule
4763 for (int i=AtomCount;i--;) // no need as -1 means already that there is trivial correspondence
4764 AtomicMap[i] = i;
4765 *out << Verbose(4) << "Map is trivial." << endl;
4766 } else {
4767 *out << Verbose(4) << "Map is ";
4768 Walker = start;
4769 while (Walker->next != end) {
4770 Walker = Walker->next;
4771 if (Walker->father == NULL) {
4772 AtomicMap[Walker->nr] = -2;
4773 } else {
4774 OtherWalker = OtherMolecule->start;
4775 while (OtherWalker->next != OtherMolecule->end) {
4776 OtherWalker = OtherWalker->next;
4777 //for (int i=0;i<AtomCount;i++) { // search atom
4778 //for (int j=0;j<OtherMolecule->AtomCount;j++) {
4779 //*out << Verbose(4) << "Comparing father " << Walker->father << " with the other one " << OtherWalker->father << "." << endl;
4780 if (Walker->father == OtherWalker)
4781 AtomicMap[Walker->nr] = OtherWalker->nr;
4782 }
4783 }
4784 *out << AtomicMap[Walker->nr] << "\t";
4785 }
4786 *out << endl;
4787 }
4788 *out << Verbose(3) << "End of GetFatherAtomicMap." << endl;
4789 return AtomicMap;
4790};
4791
4792/** Stores the temperature evaluated from velocities in molecule::Trajectories.
4793 * We simply use the formula equivaleting temperature and kinetic energy:
4794 * \f$k_B T = \sum_i m_i v_i^2\f$
4795 * \param *out output stream for debugging
4796 * \param startstep first MD step in molecule::Trajectories
4797 * \param endstep last plus one MD step in molecule::Trajectories
4798 * \param *output output stream of temperature file
4799 * \return file written (true), failure on writing file (false)
4800 */
4801bool molecule::OutputTemperatureFromTrajectories(ofstream *out, int startstep, int endstep, ofstream *output)
4802{
4803 double temperature;
4804 atom *Walker = NULL;
4805 // test stream
4806 if (output == NULL)
4807 return false;
4808 else
4809 *output << "# Step Temperature [K] Temperature [a.u.]" << endl;
4810 for (int step=startstep;step < endstep; step++) { // loop over all time steps
4811 temperature = 0.;
4812 Walker = start;
4813 while (Walker->next != end) {
4814 Walker = Walker->next;
4815 for (int i=NDIM;i--;)
4816 temperature += Walker->type->mass * Trajectories[Walker].U.at(step).x[i]* Trajectories[Walker].U.at(step).x[i];
4817 }
4818 *output << step << "\t" << temperature*AtomicEnergyToKelvin << "\t" << temperature << endl;
4819 }
4820 return true;
4821};
Note: See TracBrowser for help on using the repository browser.