move initializers for structs from header file to implementation, add constructors

src/ML-POD/fitpod_command.h

@@ -11,7 +11,6 @@
    See the README file in the top-level LAMMPS directory.
 ------------------------------------------------------------------------- */
 
-
 #ifdef COMMAND_CLASS
 // clang-format off
 CommandStyle(fitpod,FitPOD);
@@ -33,14 +32,17 @@ public:
 
  private:
   struct datastruct {
-    std::string file_format = "extxyz";
+    datastruct();
-    std::string file_extension = "xyz";
+    void copydatainfo(datastruct &data) const;
+
+    std::string file_format;
+    std::string file_extension;
     std::string data_path;
     std::vector<std::string> data_files;     // sorted file names
     std::vector<std::string> group_names;    // sorted group names
     std::vector<std::string> filenames;
-    std::string filenametag = "pod";
+    std::string filenametag;
-    std::string group_weight_type = "global";
+    std::string group_weight_type;
 
     std::vector<int> num_atom;
     std::vector<int> num_atom_cumsum;
@@ -52,86 +54,65 @@ private:
     int num_atom_max;
     int num_config_sum;
 
-    double *lattice=nullptr;
+    double *lattice;
-    double *energy=nullptr;
+    double *energy;
-    double *stress=nullptr;
+    double *stress;
-    double *position=nullptr;
+    double *position;
-    double *force=nullptr;
+    double *force;
-    int *atomtype=nullptr;
+    int *atomtype;
     // Group weights will have same size as energy.
-    double *we=nullptr;
+    double *we;
-    double *wf=nullptr;
+    double *wf;
 
-    int training = 1;
+    int training;
-    int normalizeenergy = 1;
+    int normalizeenergy;
-    int training_analysis = 1;
+    int training_analysis;
-    int test_analysis = 1;
+    int test_analysis;
-    int training_calculation = 0;
+    int training_calculation;
-    int test_calculation = 0;
+    int test_calculation;
-    int randomize = 1;
+    int randomize;
-    int precision = 8;
+    int precision;
-    double fraction = 1.0;
+    double fraction;
 
     std::unordered_map<std::string, double> we_map;
     std::unordered_map<std::string, double> wf_map;
 
-    double fitting_weights[12] = {100.0, 1.0, 0.0, 1, 1, 0, 0, 1, 1, 1, 1, 1e-10};
+    double fitting_weights[12];
-
-    void copydatainfo(datastruct &data) const
-    {
-      data.data_path = data_path;
-      data.file_format = file_format;
-      data.file_extension = file_extension;
-      data.data_files = data_files;
-      data.filenametag = filenametag;
-      data.filenames = filenames;
-      data.training_analysis = training_analysis;
-      data.test_analysis = test_analysis;
-      data.training_calculation = training_calculation;
-      data.test_calculation = test_calculation;
-      data.fraction = fraction;
-      data.randomize = randomize;
-      data.precision = precision;
-      data.training = training;
-      data.normalizeenergy = normalizeenergy;
-      for (int i = 0; i < 12; i++)
-        data.fitting_weights[i] = fitting_weights[i];
-      data.we_map = we_map;
-      data.wf_map = wf_map;
-    }
   };
 
   struct neighborstruct {
-    int *alist=nullptr;
+    neighborstruct();
-    int *pairnum=nullptr;
-    int *pairnum_cumsum=nullptr;
-    int *pairlist=nullptr;
-    double *y=nullptr;
 
-    //int natom;
+    int *alist;
-    //int nalist;
+    int *pairnum;
-    int natom_max = 0;
+    int *pairnum_cumsum;
-    int sze = 0;
+    int *pairlist;
-    int sza = 0;
+    double *y;
-    int szy = 0;
+
-    int szp = 0;
+    int natom_max;
+    int sze;
+    int sza;
+    int szy;
+    int szp;
   };
 
   struct descriptorstruct {
-    double *bd=nullptr;  // base descriptors
+    descriptorstruct();
-    double *pd=nullptr;  // multi-environment descriptors (probabilities)
+
-    double *gd=nullptr;  // global descriptors
+    double *bd;     // base descriptors
-    double *gdd=nullptr; // derivatives of global descriptors and peratom descriptors
+    double *pd;     // multi-environment descriptors (probabilities)
-    double *A=nullptr;  // least-square matrix for all descriptors
+    double *gd;     // global descriptors
-    double *b=nullptr;  // least-square vector for all descriptors
+    double *gdd;    // derivatives of global descriptors and peratom descriptors
-    double *c=nullptr;  // coefficents of descriptors
+    double *A;      // least-square matrix for all descriptors
-    int szd = 0;
+    double *b;      // least-square vector for all descriptors
-    int nCoeffAll = 0; // number of global descriptors
+    double *c;      // coefficents of descriptors
-    int nClusters = 0; // number of environment clusters
+    int szd;
+    int nCoeffAll;    // number of global descriptors
+    int nClusters;    // number of environment clusters
   };
 
-  int save_descriptors = 0;
+  int save_descriptors;
-  int compute_descriptors = 0;
+  int compute_descriptors;
   datastruct traindata;
   datastruct testdata;
   datastruct envdata;
@@ -159,9 +140,12 @@ private:
   void matrix33_inverse(double *invA, double *A1, double *A2, double *A3);
 
   double squareDistance(const double *a, const double *b, int DIMENSIONS);
-  void assignPointsToClusters(double *points, double *centroids, int *assignments, int *clusterSizes, int NUM_POINTS, int NUM_CLUSTERS, int DIMENSION);
+  void assignPointsToClusters(double *points, double *centroids, int *assignments,
-  void updateCentroids(double *points, double *centroids, int *assignments, int *clusterSizes, int NUM_POINTS, int NUM_CLUSTERS, int DIMENSIONS);
+                              int *clusterSizes, int NUM_POINTS, int NUM_CLUSTERS, int DIMENSION);
-  void KmeansClustering(double *points, double *centroids, int *assignments, int *clusterSizes, int NUM_POINTS, int NUM_CLUSTERS, int DIMENSIONS, int MAX_ITER);
+  void updateCentroids(double *points, double *centroids, int *assignments, int *clusterSizes,
+                       int NUM_POINTS, int NUM_CLUSTERS, int DIMENSIONS);
+  void KmeansClustering(double *points, double *centroids, int *assignments, int *clusterSizes,
+                        int NUM_POINTS, int NUM_CLUSTERS, int DIMENSIONS, int MAX_ITER);
 
   void savedata2textfile(std::string filename, std::string text, double *A, int n, int m, int dim);
   void savematrix2binfile(std::string filename, double *A, int nrows, int ncols);
@@ -169,14 +153,20 @@ private:
 
   // functions for reading input files and fitting
 
-  int read_data_file(double *fitting_weights, std::string &file_format, std::string &file_extension, std::string &env_path,
+  int read_data_file(double *fitting_weights, std::string &file_format, std::string &file_extension,
-    std::string &test_path, std::string &training_path, std::string &filenametag, const std::string &data_file, std::string &group_weight_type,
+                     std::string &env_path, std::string &test_path, std::string &training_path,
-    std::unordered_map<std::string, double> &we_map, std::unordered_map<std::string, double> &wf_map);
+                     std::string &filenametag, const std::string &data_file,
-  void get_exyz_files(std::vector<std::string> &, std::vector<std::string> &, const std::string &, const std::string &);
+                     std::string &group_weight_type,
+                     std::unordered_map<std::string, double> &we_map,
+                     std::unordered_map<std::string, double> &wf_map);
+  void get_exyz_files(std::vector<std::string> &, std::vector<std::string> &, const std::string &,
+                      const std::string &);
   int get_number_atom_exyz(std::vector<int> &num_atom, int &num_atom_sum, std::string file);
-  int get_number_atoms(std::vector<int>& num_atom, std::vector<int> &num_atom_sum, std::vector<int>& num_config, std::vector<std::string> training_files);
+  int get_number_atoms(std::vector<int> &num_atom, std::vector<int> &num_atom_sum,
-  void read_exyz_file(double *lattice, double *stress, double *energy, double *we, double *wf, double *pos, double *forces,
+                       std::vector<int> &num_config, std::vector<std::string> training_files);
-    int *atomtype, std::string file, std::vector<std::string> species, double we_group, double wf_group);
+  void read_exyz_file(double *lattice, double *stress, double *energy, double *we, double *wf,
+                      double *pos, double *forces, int *atomtype, std::string file,
+                      std::vector<std::string> species, double we_group, double wf_group);
   void get_data(datastruct &data, const std::vector<std::string> &species);
   std::vector<int> linspace(int start_in, int end_in, int num_in);
   std::vector<int> shuffle(int start_in, int end_in, int num_in);
@@ -188,7 +178,8 @@ private:
   int podneighborlist(int *neighlist, int *numneigh, double *r, double rcutsq, int nx, int N,
                       int dim);
   int podfullneighborlist(double *y, int *alist, int *neighlist, int *numneigh, int *numneighsum,
-    double *x, double *a1, double *a2, double *a3, double rcut, int *pbc, int nx);
+                          double *x, double *a1, double *a2, double *a3, double rcut, int *pbc,
+                          int nx);
   void estimate_memory_neighborstruct(const datastruct &data, int *pbc, double rcut, int nelements);
   void allocate_memory_neighborstruct();
   void allocate_memory_descriptorstruct(int nd);
@@ -204,8 +195,6 @@ private:
   double energyforce_calculation_fastpod(double *force, const datastruct &data, int ci);
   void energyforce_calculation(const datastruct &data);
 };
-
 }    // namespace LAMMPS_NS
-
 #endif
 #endif