|
/* VERSION 1 19 MAY 1998*/ // IOANNIS KARYDIS, SIMON STONE, MATHEW SCOINES, KEITH PASSMORE @ Kingston University #include #include #include #include #include
char strings[20][100]; int counter; char sentence2[40]; int no_args; char outp[100]="\0"; void display(char *); void temp_man(char *templ, char *response) { char qtempl[100]="\0"; char *part1; char *part2; /*create array buffers to read template responses, seperate them at the square bracket and insert another buffer containing the response from the parse sentence function */ strcpy(qtempl,templ); part1 = strtok(qtempl, "["); part2 = strtok(NULL,"\0"); strcat(outp, part1); strcat(outp, " "); strcat(outp, response); strcat(outp, " "); strcat(outp, part2); strcat(outp, "\0");
/*return output to display function */ }
char * templ_database(char *sentence, int ident) { fstream tpl;//create a new file stream char buffer[100]="\0"; char templates[20][100]; char identt[30]; int no_r=0; //open kbase tpl file for input tpl.open("kbase.tpl",ios::in); /* allow identity of interrogatives to be read as character strings not integers */ if (ident == 1) strcpy(identt,"what"); if (ident == 2) strcpy(identt,"who"); if (ident == 3) strcpy(identt,"where"); if (ident == 4) strcpy(identt,"which"); if (ident == 5) strcpy(identt,"when"); if (ident == 6) strcpy(identt,"why"); if (ident == 7) strcpy(identt,"how"); if (ident == 8) strcpy(identt,"are"); if (ident == 9) strcpy(identt,"am"); if (ident == 10) strcpy(identt,"were"); if (ident == 11) strcpy(identt,"is"); if (ident == 12) strcpy(identt,"was"); if (ident == 13) strcpy(identt,"would"); if (ident == 14) strcpy(identt,"can"); if (ident == 15) strcpy(identt,"could"); if (ident == 16) strcpy(identt,"do"); if (ident == 17) strcpy(identt,"does"); if (ident == 18) strcpy(identt,"did"); if (ident == 19) strcpy(identt,"have"); if (ident == 20) strcpy(identt,"has"); if (ident == 21) strcpy(identt,"had"); if (ident == 22) strcpy(identt,"must"); if (ident == 23) strcpy(identt,"might"); if (ident == 24) strcpy(identt,"shall"); if (ident == 25) strcpy(identt,"should"); if (ident == 26) strcpy(identt,"like");
/*do this while not at the end of the template file */ while (!tpl.eof()) { while(buffer[0]!='*' && !tpl.eof()) //while the buffer is not equal to * and not end of file do this { //get the line and put it in the buffer up to the size of the buffer tpl.getline(buffer,sizeof(buffer)); }; /*if broken out of the loop above by reading an asterisk get the next line in to a buffer */ tpl.getline(buffer,sizeof(buffer)); //compare it to the listed interrogatives if (strcmp(buffer,identt) == 0) { /* read the number of lines in the selection of possible answers. This is written in the template file as an integer, beneath the asterisk */ tpl >> no_r; //this next line just moves the file pointer along to the end of the line // which the >> commnand does not do tpl.getline(buffer,sizeof(buffer)); //collect all of the lines for (counter=0;counter < no_r;counter++) { tpl.getline(templates[counter],sizeof(buffer)); } //use a random number to select one of the lines in the buffers as an answer temp_man(templates[rand() % no_r],sentence); tpl.close(); //this function constructs the output variable which is used by the display // function } }; return "I don't understand life..."; } void clean(char *sentence) //function to remove non alphabetical characters and make everything //lower case { int counter; for (counter = 0; counter < strlen(sentence); counter++) { sentence[counter] = removed(sentence[counter]); sentence[counter] = tolower(sentence[counter]); } for (counter = strlen(sentence) - 1; counter > 0 ; counter--) // Remove empty { // spaces if (sentence[counter] == ' ') { sentence[counter] = '\0'; } else { break; } } //copy sentence into sentence2 strcpy(sentence2, sentence); }
void display(char *output) { /* format web page send output through cgi bin */
cout << "Content-Type: text/HTML \n\n"; cout << "<HTML>\n"; cout << "<HEAD>\n"; cout << "<TITLE>Response</TITLE>\n"; cout << "</HEAD>\n"; cout << "<body bgcolor=\"#FFFFFF\">\n"; cout << "<p align=\"center\"><img src=\"led.gif\" width=\"378\" height=\"64\"></p>\n"; cout << "<p align=\"center\"><font size=\"4\">A ChatterBot by:</font></p>\n"; cout << "<p align=\"center\"><a href=\"mailto:[email protected]\">Keith\n"; cout << "Passmore</a><br>\n"; cout << "<a href=\"mailto:[email protected]\">Simon Stone</a><br>\n"; cout << "<a href=\"mailto:[email protected]\">Ioannis Karydis</a><br>\n"; cout << "<a href=\"mailto:[email protected]\">Mathew Scoines</a></p>\n"; cout << "<p align=\"center\"><a href=\"http://www.kingston.ac.uk/~k967325/arthur.htm\">CLICK HERE TO READ HOW ARTHUR WAS DEVELOPED</a></p>\n"; cout << "<ISINDEX prompt='Please type your question: '>\n\n"; cout << output << "\n"; cout << "</body>\n"; cout << "</html>\n"; } int search_database(char *sentence) { //open database file fstream db; //declare no_r as an integer int no_r; char responses[20][100]; int counter; char buffer[80]; db.open("kbase.dat",ios::in); /*open the database */ clean(sentence); //clean the sentence //sentence2 is returned by clean() sentence2 and then copied into sentence strcpy(sentence,sentence2); /*do this while the end of the database file has not been reached */ while (!db.eof()) { /* while the first character of the array buffer is not an asterisk and it is not the end of the file, do this */ while(buffer[0]!='*' && !db.eof()) { //take the next line into the buffer db.getline(buffer,sizeof(buffer)); /* if the buffer is empty i.e. the line pulled in is blank then insert the characters 'EEEEEE' just so that the line is not empty. */ if (strcmp(buffer, "")==0) strcpy(buffer, "EEEE"); }; //otherwise take the next line into the buffer db.getline(buffer,sizeof(buffer)); //if the buffer is empty insert the characters 'EEEEE'. if (strcmp(buffer, "")==0) strcpy(buffer, "EEEE"); /* compare the contents of the buffer to the sentence, looking to match the buffer to any part of the sentence */ if (strstr(sentence,buffer) != NULL) { /* if there is a match do this take the number from the database file that indicates the number of choices of response available */ db >> no_r; //nove the line pointer onto the next line as the //>> command does not do this db.getline(buffer,sizeof(buffer)); //collect all of the responses available for (counter=0;counter < no_r;counter++) { db.getline(responses[counter],sizeof(buffer)); } display(responses[rand() % no_r]); //randomly select one of the responses db.close(); //close the database return 0; //return the response to main which sends it to display } };
return 1; }
void response(char *pattern) { char rsentence[255]; rsentence[0]='\0'; //create an array for response sentence called rsentence //declare counter, counter2 and 'i' as integers int counter=0, counter2=0, i=0; //create a flag to use later in the program to signal an effect int ok = 0; //allow for an integer value for the identity of the question int ident_question; //declare rand_num as an integer int rand_num; char rpattern[20] = "000000000000000000"; /*initialize a pattern for responses called rpattern to zero. Compare pattern from parse() to check for valid pattern. In this case; interrogative, verb, pronoun as the first three digits */ ok = 0; if (strncmp(pattern, "643", 3) ==0) { // set counter to count to the length of the pattern for (counter=0;counter<strlen(pattern);counter++) { /* look for a match between the integer value of the pattern at that point in the pattern indicated by the counter and the integer value of an interrogative defined in the header file */ if (pattern[counter]==Interrogative) { /* if there is a match to the interrogative,look for a match to the individual interrogative */ if (strcmp(strings[counter],"what")==0) ident_question=What; if (strcmp(strings[counter],"when")==0) ident_question=When; if (strcmp(strings[counter],"where")==0) ident_question=Where; if (strcmp(strings[counter],"why")==0) ident_question=Why; if (strcmp(strings[counter],"which")==0) ident_question=Which; if (strcmp(strings[counter],"who")==0) ident_question=Who; if (strcmp(strings[counter],"how")==0) ident_question=How; } } //define the rules for the response to the type of questions listed below if (ident_question == What || ident_question == How || ident_question == Who || ident_question == Why || ident_question == Where || ident_question == Which) { for (counter=0;counter<strlen(pattern);counter++) //count through the pattern using the integer values for pronouns { for (counter = 0; counter < strlen(rpattern); counter++) {if (pattern[counter]==Pro) /*if the pattern is a pronoun, look through the list of pronouns, compare the strings(words)as they are counted through the sentence. pronoun 1 is 'you' if the sentence contains 'you'. The rsentence would contain 'I' */ if ((strcmp(strings[counter],pronoun[1]))==0) strcat(rsentence,"I "); // if the pronoun is 'I', the return sentence will contain 'you' if((strcmp(strings[counter],pronoun[0]))==0) strcat(rsentence,"you "); //if the pronoun is 'he' the return sentence (rsentence) will contain 'he' if ((strcmp(strings[counter],pronoun[3]))==0) strcat(rsentence,"he "); //if the pronoun is 'she' the return sentence will contain 'she' if ((strcmp(strings[counter],pronoun[4]))==0) strcat(rsentence,"she "); //if the pronoun is 'it' the return sentence will contain 'it' if ((strcmp(strings[counter], pronoun[6]))==0) strcat(rsentence,"it "); //if the pronoun is 'they' the return sentence will contain 'they' if ((strcmp(strings[counter],pronoun[2]))==0) strcat(rsentence,"they "); //if the pronoun is 'we' return sentence will contain 'you' if ((strcmp(strings[counter],pronoun[8]))==0) strcat(rsentence,"you "); } } for (counter=0;counter<strlen(rpattern);counter++) { /* this line breaks the program out of the loop if the verb is 'does' or 'do'. The program resumes at the line if pattern[0]=='6' */ if (strcmp(strings[counter], verb[19]) == 0 || strcmp(strings[counter], verb[7])==0) { break; } /* if the pronoun in the incoming sentence is 'he', 'she' or 'it' the return sentence will be appended with 'is' */ if (strcmp(strings[counter],pronoun[3])==0 || strcmp(strings[counter],pronoun[4])==0 || strcmp(strings[counter], pronoun[6])==0) { strcat(rsentence,"is "); } /* if the pronoun in the incoming sentence is 'they' or 'we', then append the return sentence with 'are' */ if (strcmp(strings[counter], pronoun[2])==0 || strcmp(strings[counter], pronoun[8])==0) { strcat(rsentence, "are "); } /* if the incoming sentence pronoun is 'you' then append the return sentence with 'am' */ if (strcmp(strings[counter], pronoun[1])==0) { strcat(rsentence, "am "); } /* if the pronoun in the incoming sentence is 'i' then append the return sentence with 'are' */ if (strcmp(strings[counter], pronoun[0]) == 0) { strcat(rsentence, "are "); } } strcat(rsentence, " "); //if the first digit of the pattern in the incoming sentence is '6' or '4' if (pattern[0]=='6' || pattern[0]=='4') { //and the length of the pattern is greater than 3 digits if (strlen(pattern) > 3) { //set counter to zero and do the following counter = 0; do /* this do statement is repeated until all words in the pattern have been compared */ { /* if the word at position [counter] is 'does' then set counter2 to counter +1. This means that only words in the sentence that come after the word 'does' will be looked at for comparison */ if ((strcmp(strings[counter],verb[19]))==0) { counter2 = counter + 1; //do this while counter2 is less than the number of verbs in the header file do { //if another word in the pattern is a verb if (pattern[counter2] == Verb) { //add the verb to the sentence strcat(rsentence, strings[counter2]); if (strcmp(strings[counter2], verb[7]) != 0 && strcmp(strings[counter2], verb[20])!=0) //if the verb is not 'do' or 'go' add an 's' strcat(rsentence, "s"); //else add 'es' else strcat (rsentence, "es"); //add an empty space strcat(rsentence, " "); //break out of the loop ok = 1;//this flag is now set to one break; } //increment counter2 counter2++; } while (counter2 < NO_VERBS - 1); }//this brace closes the 'if the verb is does' if statement //set a loop to count through the number of verbs for(i=0; i < NO_VERBS - 1; i++) { //this loop deals with all the second verbs in a sentence if the first verb // is not 'does' //find the first verb in a sentence if((strcmp(strings[counter],verb[i]))==0 ) { //set counter2 so that all the rest of the word after that are looked at counter2 = counter + 1;
//do all this while counter two is less than the number of verbs in the header //file do { //if a second verb is found if (pattern[counter2] == Verb) { //if that verb is 'shall', 'should', 'could', or 'can' if (i==17 || i==16 || i==9 || i==10) { //add verb to rsentence and add space to rsentence strcat(rsentence, verb[i]); strcat(rsentence," "); } //if not, if flag is not set to '1' add the next verb encountered if (ok!=1) { strcat(rsentence, strings[counter2]); } strcat(rsentence, " "); //set counter to 500 in order to break loop counter2 = 500; break; } counter2++; } while (counter2 < NO_VERBS - 1); } } counter++; } while (counter < strlen(pattern)); } } //this code is used to identify a prepositional or article in the //incoming sentence. It then appends the remainder of the incoming sentence to the //response sentence after the rules pertaining to the response have been applied //to the pronoun and verb and possibly interrogative. for (counter = 0; counter < 8; counter++) { for (i = 0; i < 8; i++) { if (strcmp(strings[counter], prepositional[i]) == 0) { strcat(rsentence, prepositional[i]); //this counter starts at the word after the prepositional and counts //to the end of the incoming sentence for (counter2 = counter + 1; counter2 < no_args - 1; counter2++) { strcat(rsentence, " "); strcat(rsentence, strings[counter2]); } counter = 10; //these values are inserted into these counters //in order to break the loops for i and counter respectively i = 10; break;//break is used to break the last loop using counter2 } else if (strcmp(strings[counter], article[i]) == 0) { strcat(rsentence, article[i]); for (counter2 = counter + 1; counter2 < no_args - 1; counter2++) { strcat(rsentence, " "); strcat(rsentence, strings[counter2]); } counter = 10; i = 10; break; } } } } } //this code is for sentences starting verb, pronoun. else if (strncmp(pattern, "43",2) == 0) { for (counter=0;counter<strlen(pattern);counter++) //this code identifies all the verbs that can be used to start questions //in this pattern { if (pattern[counter]==Verb) { if (strcmp(strings[counter],"are")==0) ident_question=Are; if (strcmp(strings[counter],"am")==0) ident_question=Am; if (strcmp(strings[counter],"were")==0) ident_question=Were; if (strcmp(strings[counter],"is")==0) ident_question=Is; if (strcmp(strings[counter],"was")==0) ident_question=Was; if (strcmp(strings[counter],"would")==0) ident_question=Would; if (strcmp(strings[counter],"can")==0) ident_question=Can; if (strcmp(strings[counter],"could")==0) ident_question=Could; if (strcmp(strings[counter],"do")==0) ident_question=Do; if (strcmp(strings[counter],"does")==0) ident_question=Does; if (strcmp(strings[counter],"did")==0) ident_question=Did; if (strcmp(strings[counter],"have")==0) ident_question=Have; if (strcmp(strings[counter],"has")==0) ident_question=Has; if (strcmp(strings[counter],"had")==0) ident_question=Had; if (strcmp(strings[counter],"must")==0) ident_question=Must; if (strcmp(strings[counter],"might")==0) ident_question=Might; if (strcmp(strings[counter],"shall")==0) ident_question=Shall; if (strcmp(strings[counter],"should")==0) ident_question=Should; } } if(ident_question == Are ||ident_question == Am ||ident_question == Were ||ident_question == Is ||ident_question == Was ||ident_question == Would ||ident_question == Can ||ident_question == Could ||ident_question == Do ||ident_question == Does ||ident_question == Did ||ident_question == Have ||ident_question == Has ||ident_question == Had ||ident_question == Must ||ident_question == Might ||ident_question == Shall ||ident_question == Should) { for (counter=0;counter<strlen(pattern);counter++) //sort incoming pronouns to the appropriate output responses in the same way //as above { for (counter = 0; counter < strlen(rpattern); counter++) { if (pattern[counter]==Pro) if ((strcmp(strings[counter],pronoun[1]))==0) strcat(rsentence,"I "); if((strcmp(strings[counter],pronoun[0]))==0) strcat(rsentence,"you "); if ((strcmp(strings[counter],pronoun[3]))==0) strcat(rsentence,"he "); if ((strcmp(strings[counter],pronoun[4]))==0) strcat(rsentence,"she "); if ((strcmp(strings[counter], pronoun[6]))==0) strcat(rsentence,"it "); if ((strcmp(strings[counter],pronoun[2]))==0) strcat(rsentence,"they "); if ((strcmp(strings[counter],pronoun[8]))==0) strcat(rsentence,"you "); } } // same code as above for sorting responses if (strcmp(strings[0], verb[0])==0 || strcmp(strings[0], verb[1])==0 || strcmp(strings[0], verb[5]) == 0) { if (strcmp(strings[1], pronoun[3]) == 0 || strcmp(strings[1], pronoun[4]) == 0 || strcmp(strings[1], pronoun[6]) == 0) { strcat (rsentence, "is "); } if (strcmp(strings[1], pronoun[2]) == 0 || strcmp(strings[1], pronoun[8]) == 0) { strcat(rsentence, "are "); } if (strcmp(strings[1], pronoun[1]) == 0) { strcat(rsentence, "am "); } if (strcmp(strings[1], pronoun[0]) == 0) { strcat(rsentence, "are "); } } else { strcat (rsentence, strings[0]); } } } //take results of the response sentence to be inserted into the templates templ_database(rsentence,ident_question); display(outp); }
void parse() { int counter2, counter3; char pattern[20]="000000000000000000\0"; char output[255]; //breakdown incoming sentence into numerical values for the bit pattern //the counter in each case has its upper limit set to the number of words in //the appropriate lexical array for (counter2=0; counter2 <= counter; counter2++) { for (counter3=0; counter3<3; counter3++) { if (strcmp(adverb[counter3],strings[counter2])==0) { pattern[counter2] = Adv; } } for (counter3=0; counter3<11; counter3++) { if (strcmp(noun[counter3],strings[counter2])==0) { pattern[counter2] = Noun; } } for (counter3=0; counter3<7; counter3++) { if (strcmp(interrogative[counter3],strings[counter2])==0) { pattern[counter2] = Interrogative; } } for (counter3=0; counter3< NO_VERBS + 1; counter3++) { if (strcmp(verb[counter3],strings[counter2])==0) { pattern[counter2] = Verb; } } for (counter3=0; counter3<10; counter3++) { if (strcmp(pronoun[counter3],strings[counter2])==0) { pattern[counter2] = Pro; } } for (counter3=0; counter3<24; counter3++) { if (strcmp(adjective[counter3],strings[counter2])==0) { pattern[counter2] = Adj; } } for (counter3=0; counter3<=10; counter3++) { if (strcmp(possesive[counter3],strings[counter2])==0) { pattern[counter2] = Possesive; } } pattern[counter+1]='\0';//ends the pattern }
for (counter2=0; counter2<20; counter2++) { if (strcmp(strings[counter2],"his")==0 && pattern[counter2+1] == Noun) pattern[counter2] = Adj; }
if (strncmp(pattern,"643",3)==0 || strncmp(pattern,"43",2)==0) { //if the pattern is well formed i.e. it matches the two patterns that we have //defined so far send it to response() response(pattern); return ; } else //else return 1 to main and call search_database() function and perform keyword //search display(error_resp[rand() % 4]); }
int main(int argc, char * argv[]) { time_t t;//declare time structure for use as a seed value for randomising int counter, counter2, i; char sent[100]="\0"; //argc is a C++ defined integer that stores the number of arguments passed //to the program by the operating system //this copies argc to the global variable no_args no_args = argc; //this initialises the random number generator using system's time as seed srand((unsigned) time(&t)); //argv is a C++ defined array that stores the arguments passed to the program //the arguments are words taken from the incoming sentence the array that stores //these words then has to feed them into the sentence being created if (argv[1]) { //sent here is the beginning of the sentence structure being built strcat(sent,argv[1]); strcat(sent," "); for (counter = 2; counter < argc; counter++) { strcat(sent,argv[counter]); strcat(sent, " "); }
::counter = argc - 2; //this code cleans each word in the array and constructs the strings array //which is used by the parse function for (counter = 0; counter < argc-1; counter++) { strcpy(strings[counter], argv[counter+1]); for (i = 0; i < strlen(strings[counter]); i++) { strings[counter][i] = removed(strings[counter][i]); strings[counter][i] = tolower(strings[counter][i]); for (counter2 = strlen(strings[counter]) - 1; counter2 > 0; counter2--) { if (strings[counter][counter2] == ' ') { strings[counter][counter2] = '\0'; } } } } //if the parse function cannot parse the sentence invoke the search_database //function if (search_database(sent)!=0) parse();
} //if there are no arguments passed to the program display the empty HTML page // prompt for input else { display(" "); } return 0; } |