Select Git revision
maca_trans_chunker.c 3.15 KiB
#include<stdio.h>
#include<stdlib.h>
#include<string.h>
#include<unistd.h>
#include<getopt.h>
#include"context.h"
#include"feat_fct.h"
#include"feature_table.h"
#include"dico.h"
#include"beam.h"
#include"form2pos.h"
#include"simple_decoder_chunker.h"
/*#include"dnn_decoder.h"*/
#include"config2feat_vec.h"
void decode_chunker_help_message(context *ctx)
{
context_general_help_message(ctx);
context_beam_help_message(ctx);
context_conll_help_message(ctx);
fprintf(stderr, "INPUT\n");
context_input_help_message(ctx);
context_mcd_help_message(ctx);
context_model_help_message(ctx);
context_vocabs_help_message(ctx);
context_features_model_help_message(ctx);
context_f2p_filename_help_message(ctx);
}
void decode_chunker_check_options(context *ctx){
if(ctx->help
/*!ctx->conll_filename*/
/* || !ctx->perc_model_filename
|| !ctx->mcd_filename
|| !ctx->vocabs_filename
|| !ctx->features_model_filename*/
){
decode_chunker_help_message(ctx);
exit(1);
}
}
void decode_chunker_set_linguistic_resources_filenames(context *ctx)
{
char absolute_filename[500];
if(!ctx->perc_model_filename){
strcpy(absolute_filename, ctx->maca_data_path);
strcat(absolute_filename, DEFAULT_MODEL_CHUNKER_FILENAME);
ctx->perc_model_filename = strdup(absolute_filename);
}
if(!ctx->vocabs_filename){
strcpy(absolute_filename, ctx->maca_data_path);
strcat(absolute_filename, DEFAULT_VOCABS_CHUNKER_FILENAME);
ctx->vocabs_filename = strdup(absolute_filename);
}
/* if(!ctx->mcd_filename){
strcpy(absolute_filename, ctx->maca_data_path);
strcat(absolute_filename, DEFAULT_MULTI_COL_DESC_CHUNKER_FILENAME);
ctx->mcd_filename = strdup(absolute_filename);
}*/
if(!ctx->features_model_filename){
strcpy(absolute_filename, ctx->maca_data_path);
strcat(absolute_filename, DEFAULT_FEATURES_MODEL_CHUNKER_FILENAME);
ctx->features_model_filename = strdup(absolute_filename);
}
if(!ctx->f2p_filename){
strcpy(absolute_filename, ctx->maca_data_path);
strcat(absolute_filename, DEFAULT_F2P_FILENAME);
ctx->f2p_filename = strdup(absolute_filename);
ctx->f2p = form2pos_read(ctx->f2p_filename);
}
if(ctx->verbose){
fprintf(stderr, "perc_model_filename = %s\n", ctx->perc_model_filename);
fprintf(stderr, "vocabs_filename = %s\n", ctx->vocabs_filename);
fprintf(stderr, "mcd_filename = %s\n", ctx->mcd_filename);
fprintf(stderr, "perc_features_model_filename = %s\n", ctx->features_model_filename);
fprintf(stderr, "f2p_filename = %s\n", ctx->f2p_filename);
}
}
int main(int argc, char *argv[])
{
context *ctx = context_read_options(argc, argv);
decode_chunker_check_options(ctx);
decode_chunker_set_linguistic_resources_filenames(ctx);
ctx->features_model = feat_model_read(ctx->features_model_filename, feat_lib_build(), ctx->verbose);
ctx->vocabs = dico_vec_read(ctx->vocabs_filename, ctx->hash_ratio);
mcd_link_to_dico(ctx->mcd_struct, ctx->vocabs, ctx->verbose);
ctx->d_perceptron_features = dico_vec_get_dico(ctx->vocabs, (char *)"d_perceptron_features");
if(ctx->beam_width == 1)
simple_decoder_chunker(ctx);
context_free(ctx);
return 0;
}