Skip to content
Snippets Groups Projects
Commit fcfe6c1c authored by Alexis Nasr's avatar Alexis Nasr
Browse files

modified some english tokenization rules

parent 61309227
Branches
No related tags found
No related merge requests found
...@@ -29,7 +29,8 @@ extern char *token; ...@@ -29,7 +29,8 @@ extern char *token;
[0-9]+\.[0-9]+ {maca_tokenizer_segment(yytext, yytext);} [0-9]+\.[0-9]+ {maca_tokenizer_segment(yytext, yytext);}
don't printf("do\nnot\n"); it's printf("it\nis\n");
isn't printf("is\nnot\n");
don’t printf("do\nnot"); don’t printf("do\nnot");
doesn't printf("does\nnot"); doesn't printf("does\nnot");
doesn’t printf("does\nnot"); doesn’t printf("does\nnot");
...@@ -37,8 +38,8 @@ won't printf("will\nnot"); ...@@ -37,8 +38,8 @@ won't printf("will\nnot");
won’t printf("will\nnot"); won’t printf("will\nnot");
cannot printf("can\nnot"); cannot printf("can\nnot");
wanna printf("want\nto"); wanna printf("want\nto");
's printf("\n's"); 's {maca_tokenizer_segment((char *)"'s", yytext);}
’s printf("\n's"); ’s {maca_tokenizer_segment((char *)"’s", yytext);}
\n+ printf("\n"); \n+ printf("\n");
. {maca_tokenizer_add_char_to_token(yytext[0]);} . {maca_tokenizer_add_char_to_token(yytext[0]);}
......
...@@ -13,19 +13,11 @@ ...@@ -13,19 +13,11 @@
void cff2fann_help_message(context *ctx) void cff2fann_help_message(context *ctx)
{ {
context_general_help_message(ctx); context_general_help_message(ctx);
context_mode_help_message(ctx);
context_sent_nb_help_message(ctx);
fprintf(stderr, "INPUT\n"); printf("\nInput:\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); context_cff_help_message(ctx);
fprintf(stderr, "IN TRAIN MODE\n"); context_features_model_help_message(ctx);
context_vocabs_help_message(ctx); context_vocabs_help_message(ctx);
} }
void cff2fann_check_options(context *ctx) void cff2fann_check_options(context *ctx)
...@@ -36,9 +28,8 @@ void cff2fann_check_options(context *ctx) ...@@ -36,9 +28,8 @@ void cff2fann_check_options(context *ctx)
if(!ctx->cff_filename if(!ctx->cff_filename
|| ctx->help || ctx->help
|| !ctx->mcd_filename || !ctx->vocabs_filename
|| !ctx->features_model_filename || !ctx->features_model_filename
/* || !(ctx->cff_filename || ctx->fann_filename) */
){ ){
cff2fann_help_message(ctx); cff2fann_help_message(ctx);
exit(1); exit(1);
...@@ -63,7 +54,7 @@ void print_header(mcd *m, feat_model *fm) ...@@ -63,7 +54,7 @@ void print_header(mcd *m, feat_model *fm)
for(i=0; i <fm->nbelem; i++){ for(i=0; i <fm->nbelem; i++){
fd = fm->array[i]; fd = fm->array[i];
if(fd->nbelem > 1){ if(fd->nbelem > 1){
printf("feature %d is a complex feature, skipping it\n", i); fprintf(stderr, "feature %d is a complex feature, skipping it\n", i);
} }
else{ else{
sfd = fd->array[0]; sfd = fd->array[0];
...@@ -76,7 +67,7 @@ void print_header(mcd *m, feat_model *fm) ...@@ -76,7 +67,7 @@ void print_header(mcd *m, feat_model *fm)
for(i=0; i <fm->nbelem; i++){ for(i=0; i <fm->nbelem; i++){
fd = fm->array[i]; fd = fm->array[i];
if(fd->nbelem > 1){ if(fd->nbelem > 1){
printf("feature %d is a complex feature, skipping it\n", i); fprintf(stderr, "feature %d is a complex feature, skipping it\n", i);
} }
else{ else{
sfd = fd->array[0]; sfd = fd->array[0];
......
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment