restart using git

14 files changed, 249 insertions, 52 deletions

diff --git a/include/MSVMMaj.h b/include/MSVMMaj.h
index 021a341..5c3efb8 100644
--- a/include/MSVMMaj.h
+++ b/include/MSVMMaj.h
@@ -1,11 +1,12 @@
+#ifndef MSVMMAJ_H
+#define MSVMMAJ_H
 
-#define MAX_ITER 10000000
-#define MAX_LINE_LENGTH 1024
+#include "globals.h"
 
 /*
 	Model structure
 */
-struct Model {
+struct MajModel {
 	int weight_idx;
 	long K;
 	long n;
@@ -24,16 +25,19 @@ struct Model {
 	double *H;
 	double *R;
 	double *rho;
+	double training_error;
 	char *data_file;
 };
 
 /*
 	Data structure
 */
-struct Data {
+struct MajData {
 	long K;
 	long n;
 	long m;
 	long *y;
 	double *Z;
 };
+
+#endif
diff --git a/include/crossval.h b/include/crossval.h
new file mode 100644
index 0000000..0794622
--- /dev/null
+++ b/include/crossval.h
@@ -0,0 +1,13 @@
+#ifndef CROSSVAL_H
+#define CROSSVAL_H
+
+#include "globals.h"
+
+// forward delaration
+struct MajData;
+
+void msvmmaj_make_cv_split(long N, long folds, long *cv_idx);
+void msvmmaj_get_tt_split(struct MajData *full_data, struct MajData *train_data,
+		struct MajData *test_data, long *cv_idx, long fold_idx);
+
+#endif
diff --git a/include/globals.h b/include/globals.h
new file mode 100644
index 0000000..8420f76
--- /dev/null
+++ b/include/globals.h
@@ -0,0 +1,20 @@
+#ifndef GLOBALS_H
+#define GLOBALS_H
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#define MAX_LINE_LENGTH 1024
+
+#define Calloc(type, n) (type *)calloc((n), sizeof(type))
+#define Malloc(type, n) (type *)malloc((n)*sizeof(type))
+#define Memset(var, type, n) memset(var, 0, (n)*sizeof(type))
+
+#ifndef MIN_MAX_DEFINE
+#define MIN_MAX_DEFINE
+#define maximum(a, b) (a) > (b) ? (a) : (b)
+#define minimum(a, b) (a) < (b) ? (a) : (b)
+#endif
+
+#endif
diff --git a/include/libMSVMMaj.h b/include/libMSVMMaj.h
index 3e0f25c..21efc2f 100644
--- a/include/libMSVMMaj.h
+++ b/include/libMSVMMaj.h
@@ -1,31 +1,23 @@
-#include <stdio.h>
-#include <stdlib.h>
-#include <math.h>
-#include <cblas.h>
-#include <string.h>
+#ifndef LIBMSVMMAJ_H
+#define LIBMSVMMAJ_H
 
-#include "util.h"
-#include "matrix.h"
+#include "globals.h"
 
-void simplex_gen(long K, double *U);
-void category_matrix(struct Model *model, struct Data *data);
-void simplex_diff(struct Model *model, struct Data *dataset);
+// forward declarations 
+struct MajData;
+struct MajModel;
 
-void calculate_errors(struct Model *model, struct Data *data, double *ZV);
-void calculate_huber(struct Model *model);
+// function declarations
+void msvmmaj_simplex_gen(long K, double *U);
+void msvmmaj_category_matrix(struct MajModel *model, struct MajData *data);
+void msvmmaj_simplex_diff(struct MajModel *model, struct MajData *dataset);
 
-double get_msvmmaj_loss(struct Model *model, struct Data *data, double *ZV);
-void msvmmaj_update(struct Model *model, struct Data *data, 
-		double *B, double *ZAZ, double *ZAZV, double *ZAZVT);
-void step_doubling(struct Model *model);
+void msvmmaj_calculate_errors(struct MajModel *model, struct MajData *data, double *ZV);
+void msvmmaj_calculate_huber(struct MajModel *model);
 
-void main_loop(struct Model *model, struct Data *data);
+void msvmmaj_step_doubling(struct MajModel *model);
 
-int dposv(char UPLO, int N, int NRHS, double *A, int LDA, double *B, int LDB);
-
-void seed_model_V(struct Model *from_model, struct Model *to_model);
-void initialize_weights(struct Data *data, struct Model *model);
-
-void predict_labels(struct Data *data, struct Model *model, long *predy);
-double prediction_perf(struct Data *data, long *predy);
+void msvmmaj_seed_model_V(struct MajModel *from_model, struct MajModel *to_model);
+void msvmmaj_initialize_weights(struct MajData *data, struct MajModel *model);
 
+#endif
diff --git a/include/matrix.h b/include/matrix.h
index 4157475..5f0a441 100644
--- a/include/matrix.h
+++ b/include/matrix.h
@@ -1,3 +1,7 @@
+#ifndef MATRIX_H
+#define MATRIX_H
+
+#include "globals.h"
 
 void matrix_set(double *M, long cols, long i, long j, double val);
 void matrix_add(double *M, long cols, long i, long j, double val);
@@ -10,3 +14,5 @@ void matrix3_set(double *M, long N2, long N3, long i, long j, long k,
 double matrix3_get(double *M, long N2, long N3, long i, long j, long k);
 
 void print_matrix(double *M, long rows, long cols);
+
+#endif
diff --git a/include/msvmmaj_pred.h b/include/msvmmaj_pred.h
new file mode 100644
index 0000000..952389c
--- /dev/null
+++ b/include/msvmmaj_pred.h
@@ -0,0 +1,15 @@
+#ifndef MSVMMAJ_PRED_H
+#define MSVMMAJ_PRED_H
+
+#include "globals.h"
+
+// forward declarations
+struct MajData;
+struct MajModel;
+
+// function declarations
+void msvmmaj_predict_labels(struct MajData *data, struct MajModel *model,
+		long *predy);
+double msvmmaj_prediction_perf(struct MajData *data, long *perdy);
+
+#endif
diff --git a/include/msvmmaj_train.h b/include/msvmmaj_train.h
new file mode 100644
index 0000000..4fb198e
--- /dev/null
+++ b/include/msvmmaj_train.h
@@ -0,0 +1,19 @@
+#ifndef MSVMMAJ_TRAIN_H
+#define MSVMMAJ_TRAIN_H
+
+#include "globals.h"
+
+//forward declarations
+struct MajData;
+struct MajModel;
+
+// function declarations
+void msvmmaj_optimize(struct MajModel *model, struct MajData *data);
+
+double msvmmaj_get_loss(struct MajModel *model, struct MajData *data,
+		double *ZV);
+
+void msvmmaj_get_update(struct MajModel *model, struct MajData *data,
+		double *B, double *ZAZ, double *ZAZV, double *ZAZVT);
+
+#endif
diff --git a/include/msvmmaj_train_dataset.h b/include/msvmmaj_train_dataset.h
new file mode 100644
index 0000000..fdcdb4c
--- /dev/null
+++ b/include/msvmmaj_train_dataset.h
@@ -0,0 +1,61 @@
+#ifndef MSVMMAJ_TRAIN_DATASET_H
+#define MSVMMAJ_TRAIN_DATASET_H
+
+#include "globals.h"
+#include "types.h"
+
+struct Task {
+	KernelType kerneltype;
+	int weight_idx;
+	long folds;
+	long ID;
+	double p;
+	double kappa;
+	double lambda;
+	double epsilon;
+	double *kernel_param;
+	struct MajData *train_data;
+	struct MajData *test_data;
+	double performance;
+};
+
+struct Queue {
+	struct Task **tasks;
+	long N;
+	long i;
+};
+
+struct Training {
+	TrainType traintype;
+	long repeats;
+	long folds;
+	long Np;
+	long Nl;
+	long Nk;
+	long Ne;
+	long Nw;
+	int *weight_idxs;
+	double *ps;
+	double *lambdas;
+	double *kappas;
+	double *epsilons;
+	char *train_data_file;
+	char *test_data_file;
+};
+
+void make_queue(struct Training *training, struct Queue *queue,
+		struct MajData *train_data, struct MajData *test_data);
+
+struct Task *get_next_task(struct Queue *q);
+void start_training_tt(struct Queue *q);
+void start_training_cv(struct Queue *q);
+void free_queue(struct Queue *q);
+
+void consistency_repeats(struct Queue *q, long repeats, TrainType traintype);
+
+double cross_validation(struct MajModel *model, struct MajModel *seed_model,
+		struct MajData *data, long folds);
+
+void make_model_from_task(struct Task *task, struct MajModel *model);
+void copy_model(struct MajModel *from, struct MajModel *to);
+#endif
diff --git a/include/mylapack.h b/include/mylapack.h
new file mode 100644
index 0000000..4c79e0e
--- /dev/null
+++ b/include/mylapack.h
@@ -0,0 +1,11 @@
+#ifndef MYLAPACK_H
+#define MYLAPACK_H
+
+#include "globals.h"
+
+int dposv(char UPLO, int N, int NRHS, double *A, int LDA, double *B,
+		int LDB);
+int dsysv(char UPLO, int N, int NRHS, double *A, int LDA, int *IPIV,
+		double *B, int LDB, double *WORK, int LWORK);
+
+#endif
diff --git a/include/parallel.h b/include/parallel.h
new file mode 100644
index 0000000..8747347
--- /dev/null
+++ b/include/parallel.h
@@ -0,0 +1,13 @@
+
+struct Task {
+	enum KernelType kernel_type;
+	int weight_idx;
+	double epsilon;
+	double p;
+	double kappa;
+	double lambda;
+	double *kernel_param;
+	struct MajData **data;
+}
+
+
diff --git a/include/strutil.h b/include/strutil.h
new file mode 100644
index 0000000..66722ae
--- /dev/null
+++ b/include/strutil.h
@@ -0,0 +1,19 @@
+#ifndef STRUTIL_H
+#define STRUTIL_H
+
+#include "globals.h"
+#include "types.h"
+
+bool str_startswith(const char *str, const char *pre);
+bool str_endswith(const char *str, const char *suf);
+
+void next_line(FILE *fid, char *filename);
+void get_line(FILE *fid, char *filename, char *buffer);
+
+double get_fmt_double(FILE *fid, char *filename, const char *fmt);
+long get_fmt_long(FILE *fid, char *filename, const char *fmt);
+
+long all_doubles_str(char *buffer, long offset, double *all_doubles);
+long all_longs_str(char *buffer, long offset, long *all_longs);
+
+#endif
diff --git a/include/timer.h b/include/timer.h
new file mode 100644
index 0000000..8a737e0
--- /dev/null
+++ b/include/timer.h
@@ -0,0 +1,8 @@
+#ifndef TIMER_H
+#define TIMER_H
+
+#include "globals.h"
+
+double elapsed_time(clock_t s_time, clock_t e_time);
+
+#endif
diff --git a/include/types.h b/include/types.h
new file mode 100644
index 0000000..b4db8d8
--- /dev/null
+++ b/include/types.h
@@ -0,0 +1,21 @@
+#ifndef TYPES_H
+#define TYPES_H
+
+typedef enum {
+	false,
+	true
+} bool;
+
+typedef enum {
+	CV=0,
+	TT=1
+} TrainType;
+
+typedef enum {
+	K_LINEAR=0,
+	K_POLY=1,
+	K_RBF=2,
+	K_SIGMOID=3,
+} KernelType;
+
+#endif
diff --git a/include/util.h b/include/util.h
index 2cf36e8..facae79 100644
--- a/include/util.h
+++ b/include/util.h
@@ -1,33 +1,28 @@
-#include <stdarg.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <math.h>
-#include <string.h>
-#include <time.h>
-#include "MSVMMaj.h"
+#ifndef UTIL_H
+#define UTIL_H
 
-#define Calloc(type, n) (type *)calloc((n), sizeof(type))
-#define Malloc(type, n) (type *)malloc((n)*sizeof(type))
-#define Memset(var, type, n) memset(var, 0, (n)*sizeof(type))
-#define maximum(a, b) a > b ? a : b
-#define minimum(a, b) a < b ? a : b
+#include "globals.h"
 
-void read_data(struct Data *dataset, char *data_file);
+// forward declarations
+struct MajData;
+struct MajModel;
 
-void read_model(struct Model *model, char *model_filename);
-void write_model(struct Model *model, char *output_filename);
+// function declarations
+void msvmmaj_read_data(struct MajData *dataset, char *data_file);
 
-void write_predictions(struct Data *data, long *predy, char *output_filename);
+void msvmmaj_read_model(struct MajModel *model, char *model_filename);
+void msvmmaj_write_model(struct MajModel *model, char *output_filename);
 
-int check_argv(int argc, char **argv, char *str);
-int check_argv_eq(int argc, char **argv, char *str);
+void msvmmaj_write_predictions(struct MajData *data, long *predy, 
+		char *output_filename);
 
-void set_print_string_function(void (*print_func)(const char *));
-void info(const char *fmt,...);
+int msvmmaj_check_argv(int argc, char **argv, char *str);
+int msvmmaj_check_argv_eq(int argc, char **argv, char *str);
 
-double rnd();
+void note(const char *fmt,...);
 
-void allocate_model(struct Model *model);
-void free_model(struct Model *model);
-void free_data(struct Data *data);
+void msvmmaj_allocate_model(struct MajModel *model);
+void msvmmaj_free_model(struct MajModel *model);
+void msvmmaj_free_data(struct MajData *data);
 
+#endif