This file is indexed.

/usr/include/chemps2/CASSCF.h is in libchemps2-dev 1.8.5-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
/*
   CheMPS2: a spin-adapted implementation of DMRG for ab initio quantum chemistry
   Copyright (C) 2013-2018 Sebastian Wouters

   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 2 of the License, or
   (at your option) any later version.

   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.

   You should have received a copy of the GNU General Public License along
   with this program; if not, write to the Free Software Foundation, Inc.,
   51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/

#ifndef CASSCF_CHEMPS2_H
#define CASSCF_CHEMPS2_H

#include <string>

#include "Hamiltonian.h"
#include "Irreps.h"
#include "TwoDM.h"
#include "ThreeDM.h"
#include "DMRG.h"
#include "Problem.h"
#include "Options.h"
#include "ConvergenceScheme.h"
#include "DMRGSCFindices.h"
#include "DMRGSCFunitary.h"
#include "DMRGSCFoptions.h"
#include "DMRGSCFwtilde.h"
#include "DMRGSCFmatrix.h"
#include "DMRGSCFintegrals.h"

namespace CheMPS2{
/** CASSCF class.
    \author Sebastian Wouters <sebastianwouters@gmail.com>
    \date June 18, 2013
    
    \section intro Introduction
    
    In methods which use a FCI solver, this solver can be replaced by DMRG. DMRG allows for an efficient extraction of the 2-RDM [CAS1, CAS2]. The 2-RDM of the active space is required in the complete active space self-consistent field (CASSCF) method to compute the gradient and the Hessian with respect to orbital rotations [CAS3]. It is therefore natural to introduce a CASSCF variant with DMRG as active space solver, called DMRG-SCF [CAS2, CAS4, CAS5], which allows to treat static correlation in large active spaces. In CheMPS2, I have implemented the augmented Hessian Newton-Raphson DMRG-SCF method, with exact Hessian [CAS5, CAS6]. It can be called with the function CheMPS2::CASSCF::doCASSCFnewtonraphson.
    
    \section equations_2step_dmrgscf Orbital gradient and Hessian

    The calculation of the orbital gradient and Hessian for DMRG-SCF is based on [CAS3]. The basic idea is to express the energy with the unitary group generators:
    \f{eqnarray*}{
      \hat{E}_{pq} & = & \sum\limits_{\sigma} \hat{a}^{\dagger}_{p \sigma} \hat{a}_{q \sigma} \\
      \left[ \hat{E}_{pq} , \hat{E}_{rs} \right] & = & \delta_{qr} \hat{E}_{ps} - \delta_{ps} \hat{E}_{rq} \\
      \hat{E}^{-}_{pq} & = & \hat{E}_{pq} - \hat{E}_{qp} \\
      \hat{T} & = & \sum\limits_{p<q} x_{pq} \hat{E}^{-}_{pq} \\
      E(\vec{x}) & = & \braket{0 | e^{\hat{T}} \hat{H} e^{-\hat{T}} | 0 } \\
      \left. \frac{\partial E(\vec{x})}{\partial x_{ij}} \right|_{0} & = & \braket{ 0 | \left[ \hat{E}_{ij}^{-}, \hat{H} \right] | 0 } \\
      \left. \frac{\partial^2 E(\vec{x})}{\partial x_{ij} \partial x_{kl}} \right|_{0} & = & \frac{1}{2} \braket{ 0 |  \left[ \hat{E}_{ij}^{-}, \left[ \hat{E}_{kl}^{-}, \hat{H} \right] \right] | 0 } + \frac{1}{2} \braket{ 0 |  \left[ \hat{E}_{kl}^{-}, \left[ \hat{E}_{ij}^{-}, \hat{H} \right] \right] | 0 }
    \f}
    The variables \f$x_{pq}\f$ only connect orbitals with the same irrep (\f$I_p=I_q\f$). Assuming that DMRG is exact, \f$x_{pq}\f$ in addition only connects orbitals when they belong to different occupation blocks: occupied, active, virtual. With some algebra, the derivatives can be rewritten. Real-valued symmetric one-electron integrals \f$h_{ij}\f$ and real-valued eightfold permutation symmetric two-electron integrals \f$(ij | kl)\f$ are assumed (chemical notation for the latter).
    \f{eqnarray*}{
      \Gamma^{2A}_{ijkl} & = & \sum\limits_{\sigma \tau} \braket{ 0 | \hat{a}^{\dagger}_{i \sigma} \hat{a}_{j \tau}^{\dagger} \hat{a}_{l \tau} \hat{a}_{k \sigma} | 0} \\
      \Gamma^1_{ij} & = & \sum\limits_{\sigma} \braket{ 0 | \hat{a}^{\dagger}_{i \sigma} \hat{a}_{j \sigma} | 0} \\
      \left. \frac{\partial E(\vec{x})}{\partial x_{ij}} \right|_{0} & = & 2 \left( F_{ij} - F_{ji} \right) \\
      F_{pq} & = & \sum\limits_{r} \Gamma_{pr}^{1} h_{qr} + \sum\limits_{rst} \Gamma^{2A}_{psrt} (qr | st) \\
      \left. \frac{\partial^2 E(\vec{x})}{\partial x_{ij} \partial x_{kl}} \right|_{0} & = & w_{ijkl} - w_{jikl} - w_{ijlk} + w_{jilk} \\
      w_{pqrs} & = & \delta_{qr} \left( F_{ps} + F_{sp} \right) + \tilde{w}_{pqrs} \\
      \tilde{w}_{pqrs} & = & 2 \Gamma^1_{pr} h_{qs} + 2 \sum\limits_{\alpha \beta} \left( \Gamma^{2A}_{r \alpha p \beta} (qs | \alpha \beta ) + \left( \Gamma^{2A}_{r \alpha \beta p} + \Gamma^{2A}_{r p \beta \alpha} \right) (q \alpha | s \beta ) \right)
    \f}
    In the calculation of \f$F_{pq}\f$, the indices \f$prst\f$ can only be occupied or active due to their appearance in the density matrices, and the only index which can be virtual is hence \f$q\f$. Moreover, due to the irrep symmetry of the integrals and density matrices, \f$F_{pq}\f$ is diagonal in the irreps: \f$I_p = I_q\f$. Alternatively, this can be understood by the fact that \f$x_{pq}\f$ only connects orbitals with the same irrep.

    In the calculation of \f$\tilde{w}_{pqrs}\f$, the indices \f$pr\alpha\beta\f$ can only be occupied or active due to their appearance in the density matrices, and the only indices which can be virtual are hence \f$qs\f$. Together with the remark for \f$F_{pq}\f$, this can save time for the two-electron integral rotation. Moreover, as \f$x_{pq}\f$ only connects orbitals with the same irrep, \f$I_p = I_q\f$ and \f$I_r = I_s\f$ in \f$\tilde{w}_{pqrs}\f$.

    By rewriting the density matrices, the calculation of \f$F_{pq}\f$ and \f$\tilde{w}_{pqrs}\f$ can be simplified. In the following, occ and act denote the doubly occupied and active orbital spaces, respectively.
    \f{eqnarray*}{
      \Gamma^1_{ij} & = & 2 \delta_{ij}^{occ} + \Gamma^{1,act}_{ij} \\
      \Gamma^{2A}_{ijkl} & = & \Gamma^{2A,act}_{ijkl} + 2 \delta_{ik}^{occ} \Gamma^{1,act}_{jl} - \delta_{il}^{occ} \Gamma^{1,act}_{jk} \\
      & + & 2 \delta_{jl}^{occ} \Gamma^{1,act}_{ik} - \delta_{jk}^{occ} \Gamma^{1,act}_{il} + 4 \delta_{ik}^{occ} \delta_{jl}^{occ} - 2 \delta_{il}^{occ} \delta_{jk}^{occ}
    \f}
    Define the following symmetric charge (Coulomb + exchange) matrices:
    \f{eqnarray*}{
      Q^{occ}_{ij} & = & \sum\limits_{s \in occ} \left[ 2 (ij | ss) - (is | js ) \right] \\
      Q^{act}_{ij} & = & \sum\limits_{st \in act} \frac{1}{2} \Gamma^{1,act}_{st} \left[ 2 (ij | st) - (is | jt ) \right] 
    \f}
    They can be calculated efficiently by (1) rotating the occupied and active density matrices from the current basis to the original basis, (2) contracting the rotated density matrices with the two-electron integrals in the original basis, and (3) rotating these contractions to the current basis. The constant part and the one-electron integrals of the active space Hamiltonian are:
    \f{eqnarray*}{
      \tilde{E}_{const} & = & E_{const} + \sum\limits_{s \in occ} \left( 2 h_{ss} + Q_{ss}^{occ} \right) \\
      \tilde{h}_{ij} & = & h_{ij} + Q_{ij}^{occ} 
    \f}
    The calculation of \f$F_{pq}\f$ boils down to:
    \f{eqnarray*}{
      p \in occ & : & F_{pq} = 2 \left( h_{qp} + Q^{occ}_{qp} + Q^{act}_{qp} \right) \\
      p \in act & : & F_{pq} = \sum\limits_{r \in act} \Gamma^{1,act}_{pr} \left[ h_{qr} + Q^{occ}_{qr} \right] +  \sum\limits_{rst \in act} \Gamma^{2A,act}_{psrt} (qr | st)
    \f}
    And the calculation of \f$\tilde{w}_{pqrs}\f$ (remember that \f$I_p = I_q\f$ and \f$I_r = I_s\f$):
    \f{eqnarray*}{
      (p,r) & \in & (occ,occ) : \tilde{w}_{pqrs} \\
                          & = & 4 \delta_{pr}^{occ} \left[ h_{qs} + Q^{occ}_{qs} + Q^{act}_{qs} \right] \\
                          & + & 4 \left[ 4 (qp | sr) - ( qs | pr ) - ( qr | sp ) \right] \\
      (p,r) & \in & (act,act) : \tilde{w}_{pqrs} = 2 \Gamma^{1,act}_{rp} \left[ h_{qs} + Q^{occ}_{qs} \right] \\
                          & + & 2 \sum\limits_{\alpha\beta \in act} \left[ \Gamma^{2A,act}_{r \alpha p \beta} (qs | \alpha \beta ) + \left( \Gamma^{2A,act}_{r \alpha \beta p} + \Gamma^{2A,act}_{r p \beta \alpha} \right) (q \alpha | s \beta ) \right] \\
      (p,r) & \in & (act,occ) : \tilde{w}_{pqrs} \\
                          & = & 2 \sum\limits_{\alpha \in act} \Gamma^{1,act}_{\alpha p} \left[ 4 (q \alpha | s r) - (qs | \alpha r) - (qr | s \alpha) \right] \\
      (p,r) & \in & (occ,act) : \tilde{w}_{pqrs} \\
                          & = & 2 \sum\limits_{\beta \in act} \Gamma^{1,act}_{r \beta} \left[ 4 (q p | s \beta) -  (qs | p \beta) - (q \beta | sp) \right]
    \f}

    \section origalgo Augmented Hessian Newton-Raphson DMRG-SCF
    
    The CASSCF energy is a function of \f$\vec{x}\f$. Up to second order, the energy is given by 
    \f[
    E(\vec{x}) = \braket{0 | e^{\hat{T}(\vec{x})} \hat{H} e^{-\hat{T}(\vec{x})} | 0 } \approx E(0) + \vec{x}^T \vec{g} + \frac{1}{2} \vec{x}^T \mathbf{H} \vec{x}.
    \f]
    The vector \f$\vec{g}\f$ is the gradient and the matrix \f$\mathbf{H}\f$ the Hessian for orbital rotations [CAS3]. They have been described in the previous section. The minimum of \f$E(\vec{x})\f$ is found at \f$\vec{x} = - \mathbf{H}^{-1} \vec{g}\f$. The variables \f$\vec{x}\f$ parametrize an additional orbital rotation \f$\mathbf{U}_{add} = \exp(\mathbf{X}(\vec{x}))\f$, with \f$\mathbf{X}(\vec{x}) = -\mathbf{X}^T(\vec{x})\f$ a real-valued skew-symmetric matrix. The additional orbital rotation \f$\mathbf{U}_{add}\f$ transforms the current orbitals \f$\mathbf{U}(n)\f$ to the new orbitals
    \f[
    \mathbf{U}(n+1) = \mathbf{U}_{add} \mathbf{U}(n) = \exp(\mathbf{X}(\vec{x}(n))) \mathbf{U}(n).
    \f]
    This updating scheme is called the Newton-Raphson method [CAS3]. If the Hessian is positive definite, these updates are stable. For saddle points in the energy landscape, the Hessian has negative eigenvalues, and these updates can be unstable. It is therefore better to use the augmented Hessian Newton-Raphson method [CAS7]:
    \f[
    \left[ \begin{array}{cc} \mathbf{H} & \vec{g} \\ \vec{g}^T & 0 \end{array} \right] \left[ \begin{array}{c} \vec{x} \\ 1 \end{array} \right] = \alpha \left[ \begin{array}{c} \vec{x} \\ 1 \end{array} \right].
    \f]
    The eigenvector with smallest algebraic eigenvalue determines a stable update \f$\vec{x}\f$, as is well explained in Ref. [CAS7].
    
    As a final remark in this section, I would like to say that orbitals have gauge freedom. One can always multiply them with a phase factor. It is therefore possible to choose the orbital gauges so that all \f$\mathbf{U}\f$ are always special orthogonal: \f$\det(\mathbf{U})=+1\f$.

    \section diis Direct inversion of the iterative subspace (DIIS)

    When the update norm \f$\|\vec{x}\|_2\f$ is small enough, the convergence can be accelerated by the direct inversion of the iterative subspace (DIIS) [CAS5, CAS8, CAS9, CAS10]. For a given set of orbitals \f$\mathbf{U}(n)\f$, the update \f$\vec{x}(n)\f$ is calculated with the augmented Hessian Newton-Raphson method. This update defines the next set of orbitals:
    \f[
    \mathbf{U}(n+1) = \mathbf{U}_{add} \mathbf{U}(n) = \exp(\mathbf{X}(\vec{x}(n))) \mathbf{U}(n).
    \f]
    In DIIS, the error vector \f$\vec{x}(n)\f$ and the state vector \f$\mathbf{Y}(n+1) = \log(\mathbf{U}(n+1))\f$ are added to a list. The error
    \f[
    e = \left\| \sum\limits_{i = 1}^{\kappa} c_i \vec{x}(n - \kappa + i) \right\|_2
    \f]
    is minimized under the constraint \f$\sum_{i} c_i = 1\f$. \f$\kappa\f$ is the size of the list memory, i.e. the number of retained vectors. The minimization of the error \f$e\f$ can be performed with Lagrangian calculus:
    \f[
    \left[ \begin{array}{cc} \mathbf{B} & \vec{1} \\ \vec{1}^T & 0 \end{array} \right] \left[ \begin{array}{c} \vec{c} \\ \lambda \end{array} \right] = \left[ \begin{array}{c} \vec{0} \\ 1 \end{array} \right],
    \f]
    where \f$2\lambda\f$ is the Lagrangian multiplier and
    \f[
    \left[\mathbf{B}\right]_{ij} = \vec{x}^T(n - \kappa + i) \vec{x}(n - \kappa + j) = \left[\mathbf{B}\right]_{ji}.
    \f]
    The new state vector is then defined as
    \f[
    \mathbf{Y}_{new} = \sum\limits_{i = 1}^{\kappa} c_i \mathbf{Y}(n+1 - \kappa + i).
    \f]
    The new state vector \f$\mathbf{Y}_{new}\f$ is calculated by the function CheMPS2::DIIS::calculateParam. The current orbitals are then set to \f$\mathbf{U}(n+1) = \exp(\mathbf{Y}_{new})\f$.
    
    \section biblio References

    [CAS1]  D. Zgid and M. Nooijen, Journal of Chemical Physics 128, 144115 (2008). http://dx.doi.org/10.1063/1.2883980 \n
    [CAS2]  D. Ghosh, J. Hachmann, T. Yanai and G.K.-L. Chan, Journal of Chemical Physics 128, 144117 (2008). http://dx.doi.org/10.1063/1.2883976 \n
    [CAS3]  P.E.M. Siegbahn, J. Almlof, A. Heiberg and B.O. Roos, Journal of Chemical Physics 74, 2384-2396 (1981). http://dx.doi.org/10.1063/1.441359 \n
    [CAS4]  D. Zgid and M. Nooijen, Journal of Chemical Physics 128, 144116 (2008). http://dx.doi.org/10.1063/1.2883981 \n
    [CAS5]  T. Yanai, Y. Kurashige, D. Ghosh and G.K.-L. Chan, International Journal of Quantum Chemistry 109, 2178-2190 (2009). http://dx.doi.org/10.1002/qua.22099 \n
    [CAS6]  S. Wouters, W. Poelmans, P.W. Ayers and D. Van Neck, Computer Physics Communications 185, 1501-1514 (2014). http://dx.doi.org/10.1016/j.cpc.2014.01.019 \n
    [CAS7]  A. Banerjee, N. Adams, J. Simons and R. Shepard, Journal of Physical Chemistry 89, 52-57 (1985). http://dx.doi.org/10.1021/j100247a015 \n
    [CAS8]  P. Pulay, Chemical Physics Letters 73, 393-398 (1980). http://dx.doi.org/10.1016/0009-2614(80)80396-4 \n
    [CAS9]  C.D. Sherrill, Programming DIIS, http://vergil.chemistry.gatech.edu/notes/diis/node3.html (2000). \n
    [CAS10] T. Rohwedder and R. Schneider, Journal of Mathematical Chemistry 49, 1889-1914 (2011). http://dx.doi.org/10.1007/s10910-011-9863-y \n
*/
   class CASSCF{

      public:
      
         //! Constructor
         /** \param ham_in Hamiltonian containing the matrix elements of the Hamiltonian for which a CASSCF calculation is desired
             \param docc  Array containing the number of doubly occupied HF orbitals per irrep
             \param socc  Array containing the number of singly occupied HF orbitals per irrep
             \param nocc  Array containing the number of doubly occupied (inactive) orbitals per irrep
             \param ndmrg Array containing the number of active orbitals per irrep
             \param nvirt Array containing the number of virtual (secondary) orbitals per irrep
             \param tmp_folder Temporary work folder for the DMRG renormalized operators and the ERI rotations */
         CASSCF( Hamiltonian * ham_in, int * docc, int * socc, int * nocc, int * ndmrg, int * nvirt, const string tmp_folder=CheMPS2::defaultTMPpath );
         
         //! Destructor
         virtual ~CASSCF();
         
         //! Get the number of irreps
         /** \return The number of irreps */
         int get_num_irreps();
         
         //! Do the CASSCF cycles with the augmented Hessian Newton-Raphson method
         /** \param Nelectrons Total number of electrons in the system: occupied HF orbitals + active space
             \param TwoS Twice the targeted spin
             \param Irrep Desired wave-function irrep
             \param OptScheme The optimization scheme to run the inner DMRG loop. If NULL: use FCI instead of DMRG.
             \param rootNum Denotes the targeted state in state-specific CASSCF; 1 means ground state, 2 first excited state etc.
             \param scf_options Contains the DMRGSCF options
             \return The converged DMRGSCF energy */
         double solve( const int Nelectrons, const int TwoS, const int Irrep, ConvergenceScheme * OptScheme, const int rootNum, DMRGSCFoptions * scf_options );

         //! Calculate the caspt2 correction energy for a converged casscf wavefunction
         /** \param Nelectrons Total number of electrons in the system: occupied HF orbitals + active space
             \param TwoS Twice the targeted spin
             \param Irrep Desired wave-function irrep
             \param OptScheme The optimization scheme to run the inner DMRG loop. If NULL: use FCI instead of DMRG.
             \param rootNum Denotes the targeted state in state-specific CASSCF; 1 means ground state, 2 first excited state etc.
             \param scf_options Contains the DMRGSCF options
             \param IPEA The CASPT2 IPEA shift from Ghigo, Roos and Malmqvist, Chemical Physics Letters 396, 142-149 (2004)
             \param IMAG The CASPT2 imaginary shift from Forsberg and Malmqvist, Chemical Physics Letters 274, 196-204 (1997)
             \param PSEUDOCANONICAL If true, use the exact DMRG 4-RDM in the pseudocanonical basis. If false, use the cumulant approximated DMRG 4-RDM in the unrotated basis.
             \param CHECKPOINT If true, write checkpoints to disk and read them back in again in order to perform the contraction of the generalized Fock operator with the 4-RDM in multiple runs.
             \param CUMULANT If true, a cumulant approximation is used for the 4-RDM and CHECKPOINT is overwritten to false. If false, the full 4-RDM is used.
             \return The CASPT2 variational correction energy */
         double caspt2( const int Nelectrons, const int TwoS, const int Irrep, ConvergenceScheme * OptScheme, const int rootNum, DMRGSCFoptions * scf_options, const double IPEA, const double IMAG, const bool PSEUDOCANONICAL, const bool CHECKPOINT = false, const bool CUMULANT = false );

         //! CASSCF unitary rotation remove call
         /* \param filename File to delete */
         static void deleteStoredUnitary( const string filename=CheMPS2::DMRGSCF_unitary_storage_name ){ delete_file( filename ); }

         //! CASSCF DIIS vectors remove call
         /* \param filename File to delete */
         static void deleteStoredDIIS( const string filename=CheMPS2::DMRGSCF_diis_storage_name ){ delete_file( filename ); }

         //! Build the F-matrix (Eq. (11) in the Siegbahn paper [CAS3])
         /** \param localFmat Matrix where the result should be stored
             \param localTmat Matrix which contains the one-electron integrals
             \param localJKocc Matrix which contains the Coulomb and exchange interaction due to the frozen core orbitals
             \param localJKact Matrix which contains the Coulomb and exchange interaction due to the active space
             \param localIdx Orbital index bookkeeper for the CASSCF calculations
             \param theInts The rotated two-electron integrals (at most 2 virtual indices)
             \param local2DM The DMRG 2-RDM
             \param local1DM The DMRG 1-RDM */
         static void buildFmat( DMRGSCFmatrix * localFmat, const DMRGSCFmatrix * localTmat, const DMRGSCFmatrix * localJKocc, const DMRGSCFmatrix * localJKact, const DMRGSCFindices * localIdx, const DMRGSCFintegrals * theInts, double * local2DM, double * local1DM );

         //! Build the Wtilde-matrix (Eq. (20b) in the Siegbahn paper [CAS3])
         /** \param localwtilde Where the result should be stored
             \param localTmat Matrix which contains the one-electron integrals
             \param localJKocc Matrix which contains the Coulomb and exchange interaction due to the frozen core orbitals
             \param localJKact Matrix which contains the Coulomb and exchange interaction due to the active space
             \param localIdx Orbital index bookkeeper for the CASSCF calculations
             \param theInts The rotated two-electron integrals (at most 2 virtual indices)
             \param local2DM The DMRG 2-RDM
             \param local1DM The DMRG 1-RDM */
         static void buildWtilde( DMRGSCFwtilde * localwtilde, const DMRGSCFmatrix * localTmat, const DMRGSCFmatrix * localJKocc, const DMRGSCFmatrix * localJKact, const DMRGSCFindices * localIdx, const DMRGSCFintegrals * theInts, double * local2DM, double * local1DM );

         //! Calculate the augmented Hessian Newton-Raphson update for the orthogonal orbital rotation matrix
         /** \param localFmat Matrix which contains the Fock operator (Eq. (11) in the Siegbahn paper [CAS3])
             \param localwtilde Object which contains the second order derivative of the energy with respect to the unitary (Eq. (20b) in the Siegbahn paper [CAS3])
             \param localIdx Orbital index bookkeeper for the CASSCF calculations
             \param localUmat The unitary matrix for CASSCF calculations (in this function it is used to fetch the orbital ordering convention of the skew-symmetric parametrization)
             \param theupdate Where the augmented Hessian Newton-Raphson update will be stored
             \param updateNorm Pointer to one double to store the update norm
             \param gradNorm Pointer to one double to store the gradient norm */
         static void augmentedHessianNR( DMRGSCFmatrix * localFmat, DMRGSCFwtilde * localwtilde, const DMRGSCFindices * localIdx, const DMRGSCFunitary * localUmat, double * theupdate, double * updateNorm, double * gradNorm );

         //! Copy over the DMRG 2-RDM
         /** \param theDMRG2DM The 2-RDM from the DMRG object
             \param LAS The total number of DMRG orbitals
             \param two_dm The CASSCF 2-RDM */
         static void copy2DMover( TwoDM * theDMRG2DM, const int LAS, double * two_dm );

         //! Construct the 1-RDM from the 2-RDM
         /** \param num_elec The number of DMRG active space electrons
             \param LAS The total number of DMRG orbitals
             \param one_dm The CASSCF 1-RDM
             \param two_dm The CASSCF 2-RDM */
         static void setDMRG1DM( const int num_elec, const int LAS, double * one_dm, double * two_dm );

         //! Copy a one-orbital quantity from array format to DMRGSCFmatrix format
         /** \param origin Array to copy
             \param result DMRGSCFmatrix to store the copy
             \param idx Object which handles the index conventions for CASSCF
             \param one_rdm If true, the occupied orbitals get occupation 2 */
         static void copy_active( double * origin, DMRGSCFmatrix * result, const DMRGSCFindices * idx, const bool one_rdm );

         //! Copy a one-orbital quantity from DMRGSCFmatrix format to array format
         /** \param origin DMRGSCFmatrix to copy
             \param result Array to store the copy
             \param idx Object which handles the index conventions for CASSCF */
         static void copy_active( const DMRGSCFmatrix * origin, double * result, const DMRGSCFindices * idx );

         //! From an Edmiston-Ruedenberg active space rotation, fetch the eigenvectors and store them in eigenvecs
         /** \param umat The Edmiston-Ruedenberg active space rotation
             \param idx Object which handles the index conventions for CASSCF
             \param eigenvecs Where the eigenvectors are stored */
         static void fillLocalizedOrbitalRotations( DMRGSCFunitary * umat, DMRGSCFindices * idx, double * eigenvecs );

         //! Block-diagonalize Mat
         /** \param space Can be 'O', 'A', or 'V' and denotes which block of Mat should be considered
             \param Mat Matrix to block-diagonalize
             \param Umat The unitary rotation will be updated so that Mat is block-diagonal in the orbitals 'space'
             \param work1 Workspace
             \param work2 Workspace
             \param idx Object which handles the index conventions for CASSCF
             \param invert If true, the eigenvectors are sorted from large to small instead of the other way around
             \param two_dm   If not NULL, this 4-index array will be rotated to the new eigenvecs if space == 'A'
             \param three_dm If not NULL, this 6-index array will be rotated to the new eigenvecs if space == 'A'
             \param contract If not NULL, this 6-index array will be rotated to the new eigenvecs if space == 'A' */
         static void block_diagonalize( const char space, const DMRGSCFmatrix * Mat, DMRGSCFunitary * Umat, double * work1, double * work2, const DMRGSCFindices * idx, const bool invert, double * two_dm, double * three_dm, double * contract );

         //! Construct the Fock matrix
         /** \param Fock Matrix to store the Fock operator in
             \param Tmat Matrix with the one-electron integrals
             \param Qocc Matrix with the Coulomb and exchange contributions of the occupied (inactive) orbitals
             \param Qact Matrix with the Coulomb and exchange contributions of the active space orbitals
             \param idx Object which handles the index conventions for CASSCF */
         static void construct_fock( DMRGSCFmatrix * Fock, const DMRGSCFmatrix * Tmat, const DMRGSCFmatrix * Qocc, const DMRGSCFmatrix * Qact, const DMRGSCFindices * idx );

         //! Return the RMS deviation from block-diagonal
         /** \param matrix Matrix to be assessed
             \param idx Object which handles the index conventions for CASSCF
             \return RMS deviation from block-diagonal */
         static double deviation_from_blockdiag( DMRGSCFmatrix * matrix, const DMRGSCFindices * idx );

         //! Write the checkpoint file for the contraction of the generalized Fock operator with the 4-RDM to disk
         /** \param f4rdm_file The filename
             \param hamorb1 The next hamiltonian orbital 1
             \param hamorb2 The next hamiltonian orbital 2
             \param tot_dmrg_power6 The size of the array contract
             \param contract The current partial contraction */
         static void write_f4rdm_checkpoint( const string f4rdm_file, int * hamorb1, int * hamorb2, const int tot_dmrg_power6, double * contract );

         //! Read the checkpoint file for the contraction of the generalized Fock operator with the 4-RDM from disk
         /** \param f4rdm_file The filename
             \param hamorb1 The next hamiltonian orbital 1
             \param hamorb2 The next hamiltonian orbital 2
             \param tot_dmrg_power6 The size of the array contract
             \param contract The current partial contraction
             \return Whether the file was found and read */
         static bool read_f4rdm_checkpoint( const string f4rdm_file, int * hamorb1, int * hamorb2, const int tot_dmrg_power6, double * contract );

         //! Build the contraction of the fock matrix with the 4-RDM
         /** \param fockmx Array of size ham->getL() x ham->getL() containing the Fock matrix elements
             \param dmrgsolver DMRG object which is solved, and for which the 2-RDM and 3-RDM have been calculated as well
             \param ham Active space Hamiltonian, which is needed for the size of the active space and the orbital irreps
             \param next_orb1 The next first  orbital for which the contraction should be continued
             \param next_orb2 The next second orbital for which the contraction should be continued
             \param work Work array of size ham->getL()**6
             \param result On entry, contains the partial contraction corresponding to (next_orb1, next_orb2). On exit, contains the full contraction.
             \param CHECKPOINT Whether or not the standard CheMPS2 F.4-RDM checkpoint should be created/updated to continue the contraction at later times 
             \param PSEUDOCANONICAL Whether or not pseudocanonical orbitals are used in the active space */
         static void fock_dot_4rdm( double * fockmx, CheMPS2::DMRG * dmrgsolver, CheMPS2::Hamiltonian * ham, int next_orb1, int next_orb2, double * work, double * result, const bool CHECKPOINT, const bool PSEUDOCANONICAL );

      private:

         // CASSCF tmp folder
         string tmp_folder;

         // Index convention handler
         DMRGSCFindices * iHandler;

         // Unitary matrix storage and manipulator
         DMRGSCFunitary * unitary;

         // Whether CheMPS2::CASSCF::solve has been successfully terminated
         bool successful_solve;

         // The original Hamiltonian
         double NUCL_ORIG;
         const TwoIndex  * TMAT_ORIG;
         const FourIndex * VMAT_ORIG;

         // Irreps controller
         Irreps SymmInfo;

         // The number of orbitals
         int L;

         // The number of irreps
         int num_irreps;

         // Number of DMRG orbitals
         int nOrbDMRG;

         // Space for the DMRG 1DM
         double * DMRG1DM;

         // Space for the DMRG 2DM
         double * DMRG2DM;

         // Helper function to check HF
         void checkHF( int * docc, int * socc );

         // Fill Econst and Tmat of HamDMRG
         void fillConstAndTmatDMRG( Hamiltonian * HamDMRG ) const;

         // Calculate the gradient, return function is the gradient 2-norm
         static double construct_gradient( DMRGSCFmatrix * Fmatrix, const DMRGSCFindices * idx, double * gradient );

         // Add hessian * origin to target
         static void add_hessian( DMRGSCFmatrix * Fmatrix, DMRGSCFwtilde * Wtilde, const DMRGSCFindices * idx, double * origin, double * target );

         // Construct the diagonal of the hessian
         static void diag_hessian( DMRGSCFmatrix * Fmatrix, const DMRGSCFwtilde * Wtilde, const DMRGSCFindices * idx, double * diagonal );

         // Apply augmented hessian
         static void DGEMM_WRAP( double prefactor, char transA, char transB, double * A, double * B, double * C, int m, int n, int k, int lda, int ldb, int ldc );
         static void DGEMV_WRAP( double prefactor, double * matrix, double * result, double * vector, int rowdim, int coldim, int ldmat, int incres, int incvec );
         static void augmented_hessian( DMRGSCFmatrix * Fmatrix, DMRGSCFwtilde * Wtilde, const DMRGSCFindices * idx, double * origin, double * target, double * gradient, const int linsize );

         // Rotate an active space object
         static void rotate_active_space_object( const int num_indices, double * object, double * work, double * rotation, const int LAS, const int NJUMP, const int NROTATE );

         // Fmat function as defined by Eq. (11) in the Siegbahn paper.
         DMRGSCFmatrix * theFmatrix;

         // The Coulomb and exchange interaction with the occupied and active electrons respectively
         DMRGSCFmatrix * theQmatOCC;
         DMRGSCFmatrix * theQmatACT;
         DMRGSCFmatrix * theQmatWORK;
         DMRGSCFmatrix * theTmatrix;
         void rotateOldToNew(DMRGSCFmatrix * myMatrix);
         void buildTmatrix();
         void constructCoulombAndExchangeMatrixInOrigIndices( DMRGSCFmatrix * density, DMRGSCFmatrix * result );
         void buildQmatOCC();
         void buildQmatACT();

         // Function to get coefficients of certain Slater determinants for Fe2. Important to figure out diatomic D(inf)h symmetries when calculating them in D2h symmetry.
         void coeff_fe2( DMRG * theDMRG );

         static void delete_file( const string filename );

   };
}

#endif