/*pos_tran.cpp
to translate moby part-of-speech into a Prolog lexicon
ftp://ftp.dcs.shef.ac.uk/share/ilash/Moby/moby.tar.Z
info at http://www.clres.com/dict.html
*/
#include<stdio.h>

void WritePos(FILE* out,char * W, char POS_M,bool ParseError){
char POS_Pl[32];
	strcpy(POS_Pl,((POS_M=='N')?"n":
			(POS_M=='p')?"n_pl":
			(POS_M=='h')?"np":
			(POS_M=='V')?"v":
			(POS_M=='t')?"v_tr":
			(POS_M=='i')?"v_in":
			(POS_M=='A')?"adj":
			(POS_M=='v')?"adv":
			(POS_M=='C')?"conj": //conjuction
			(POS_M=='P')?"prep":
			(POS_M=='!')?"intj": //interjection
			(POS_M=='r')?"prn": //pronoun
			(POS_M=='D')?"art_d": //Definite Article
			(POS_M=='I')?"art_i": //Indefinite Article
			(POS_M=='o')?"nom":"error") //Nominative
			);
	if(!ParseError)fprintf(out,"%s([%s]).\n",POS_Pl,W);
	else fprintf(out,"/*PARSEERROR:%s([%s]).*/\n",POS_Pl,W);
	if (!strcmp(POS_Pl,"error"))
		fprintf(out,"/*The invalid part-of-speech code for %s is %c*/\n",W,POS_M);
	return;
}

main(int argc, char**argv){
FILE * infile,*outfile;
char t,i,word[256];
int l;
bool InWord;//in text of word vs. in part of speech
bool ProblemWord;//funny character -- eliminate word
	if (argc != 3){
         	printf("\tusage: %s inputfile outputfile\n",argv[0]);
         	printf("\tThis program translates the moby part-of-speech file into a Prolog lexicon\n");
		exit(0);
	}
	if ((infile=fopen(argv[1],"r")) == NULL){
	 	printf("invalid input file\n");
	 	exit(1);
	}
	if ((outfile=fopen(argv[2],"w")) == NULL){
		printf("invalid output file\n");
		exit(1);
	}
//	for (i=0;i<126;i++){ //limited for debugging
	while(!feof(infile)){ //unlimited for SERIOUS work
	 	word[0]=0;
	 	InWord = true;
	 	ProblemWord=false;
	 	while ((t=fgetc(infile))!= 13){
	 		//printf("%c\t%d\n",t,t);
	 		if (t == -41){
	 		 	InWord= false;
	 		}
	 		else if (InWord){ //concatenate t onto word
	 			l=strlen(word);
	 			if(t>='A' && t<='Z') t = t | 0x20;
	 			word[l]=(t==' ')?',':t;
	 			if(t=='\'' || t<' ') ProblemWord = true;
	 			word[l+1]=0;
	 		}
	 		else WritePos(outfile,word,t,ProblemWord);
	 	}
	}
}