/*710_tran.cpp
to translate text710.txt into a Prolog lexicon
taken from the Oxford Text Archive
ftp://ota.ox.ac.uk/pub/ota/public/dicts/
*/
#include<stdio.h>
#include<string.h>
#define RecSize 129

void WritePos(FILE* out,char * W, char POS_M,bool ParseError){
char POS_Pl[32];
	strcpy(POS_Pl,((POS_M=='G')?"v":	//anomalous verb
			(POS_M=='H')?"v_tr":
			(POS_M=='I')?"v_in":
			(POS_M=='J')?"v":	//tr & intr verb
			(POS_M=='K')?"n":       //count noun
			(POS_M=='L')?"n":	//uncountable noun
			(POS_M=='M')?"n":	//both countable & non
			(POS_M=='N')?"n":	//Proper Noun
			(POS_M=='O')?"adj":
			(POS_M=='P')?"/*adv":
			(POS_M=='Q')?"prn": //pronoun
			(POS_M=='R')?"art_d": //Definite Article
			(POS_M=='S')?"art_i": //Indefinite Article
			(POS_M=='T')?"prep":
			(POS_M=='U')?"/*prefix":
			(POS_M=='V')?"conj": //conjuction
			(POS_M=='W')?"/*intj": //interjection
			(POS_M=='X')?"/*particle":
			(POS_M=='Y')?"/*abbreviation":
			(POS_M=='Z')?"/*not classified":"error") //Nominative
			);
	if(!ParseError)fprintf(out,"%s([%s]).%s\n",POS_Pl,W,(POS_Pl[0]=='/')?"*/":"");
	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;
}

void terminatestr(char* str, int length){
int i;
	for (i=length;i>=0;i--){
	 	if(str[i]!=' ' && str[i]!=0){
	 	 	str[i+1]=0;
	 	 	break;
	 	}
	}
	return;
}


main(int argc, char**argv){
FILE * infile,*outfile;
char t,i,j,word[24],line[RecSize];
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 COMPUTER-USABLE DICTIONARY FILE BASED ON\n\
THE OXFORD ADVANCED LEARNER'S DICTIONARY OF CURRENT ENGLISH 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<127;i++) line[i]=0;
	for(i=0;i<24;i++)word[i]=0;
	
	
//	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);
	 	}*/
	 	if(!fread(line,1,RecSize,infile)) break;//end while
		ProblemWord= false;
		strncpy(word,line,23);
		terminatestr(word,23);
		for(j=0;j<strlen(word);j++){
		 	if((t =word[j])>='A' && t<='Z') word[j] = t | 0x20;//lowercase
		 	if(t=='\''||t=='`'||t=='"'||t=='.'||t=='~') ProblemWord = true;
		 	if(t==' ') word[j]=',';
		}		
		for(j=46;j<70;j+=4){
			if(line[j]==' ')break;
			//printf("%s %d %c\n",word,j,line[j]);		 	
			WritePos(outfile,word,line[j],ProblemWord);
		}
			 	
	}
}