cff2fann.c

#include<stdio.h>
#include<stdlib.h>
#include<string.h>
#include<unistd.h>
#include<getopt.h>
#include"context.h"
#include"util.h"
#include"cf_file.h"


void cff2fann_help_message(context *ctx)
{
  context_general_help_message(ctx);
  context_mode_help_message(ctx);
  context_sent_nb_help_message(ctx);

  fprintf(stderr, "INPUT\n");
  context_conll_help_message(ctx);
  fprintf(stderr, "IN TEST MODE\n");
  context_vocabs_help_message(ctx);

  fprintf(stderr, "OUTPUT\n");
  context_cff_help_message(ctx);
  fprintf(stderr, "IN TRAIN MODE\n");
  context_vocabs_help_message(ctx);

}

void cff2fann_check_options(context *ctx)
{
  if(!ctx->input_filename
     || ctx->help
     /* || !ctx->mcd_filename */
     /* || !(ctx->cff_filename || ctx->fann_filename) */
     ){
    cff2fann_help_message(ctx);
    exit(1);
  }
}

void one_hot_print(FILE *f, int val, int dim)
{
  int i;
  for(i=0; i < dim; i++)
    fprintf(f, "%d ", (i == val)? 1  : 0);
}

void cff2fann(context *ctx)
{
  char buffer[10000];
  char *token;
  int col_nb;
  int feat_type;
  mcd *m = ctx->mcd_struct;
  FILE *f = myfopen(ctx->input_filename, "r");
  int val;
  dico *vocab;
  
  while(fgets(buffer, 10000, f)){
    /* printf("%s", buffer); */
    /* printf("\n"); */
    token = strtok(buffer, "\t");
    col_nb = 0;
    while(token){
      /* printf("col = %d token = %s max = %d\n", col_nb, token, max_array[col_nb]); */
      val = atoi(token);
      vocab = dico_vec_get_dico(ctx->vocabs, "d_perceptron_features");
      printf("!!! %s\n", dico_int2string(vocab, val));
      if(col_nb == 0){
        one_hot_print(stderr, val, ctx->mvt_nb); 
        printf("\n");
      }
      else{
	feat_type = feat_model_get_type_feat_n(ctx->features_model, col_nb - 1);
	/* printf("feat_type = %d\n", feat_type); */
	int mcd_col = m->wf2col[feat_type];
	/* printf("representation = %d\n", m->representation[mcd_col]); */
	if(m->representation[mcd_col] == MCD_REPRESENTATION_EMB){
	  /* printf("it is an embedding val = %d, file = %s\n", val, m->filename[mcd_col]); */
	  word_emb_print(stderr, m->word_emb_array[mcd_col], val);
	  printf("\n");
	}
	if(m->representation[mcd_col] == MCD_REPRESENTATION_VOCAB){
	  /* printf("it is a vocab\n"); */
	  one_hot_print(stderr, val, m->dico_array[mcd_col]->nbelem); 
	  printf("\n");
	}
      }
      col_nb++;
      token = strtok(NULL , "\t");
    }
  }
}

int main(int argc, char *argv[])
{
  context *ctx;
  int nb_feat;
  int nb_class;

  ctx = context_read_options(argc, argv);
  cff2fann_check_options(ctx);

  ctx->vocabs = dico_vec_read(ctx->vocabs_filename, ctx->hash_ratio);

  ctx->features_model = feat_model_read(ctx->features_model_filename, ctx->verbose);

  look_for_number_of_features_and_classes(ctx->input_filename, &nb_feat, &nb_class);
  ctx->mvt_nb = nb_class;

  mcd_link_to_dico(ctx->mcd_struct, ctx->vocabs, 1);

  printf("Coucou\n");
  
  cff2fann(ctx);
  return 0;
}