simple string, search in input

**lode** · 08-02-2005

Hello everybody,

I have made an activeworlds bot which have to react on chat strings.
It has to filter badwords out. The activeworlds software prevents to do that realtime, but reacting on badwords that are said is possible.

If I use this if statement:

Code:

if (aw_string(AW_CHAT_MESSAGE)=="badword")

The problem is this only works when someone 'only' says "badword" and not when saying "hello badword".

I got a suggestion from a friend:

Code:

char* msg = aw_string

(AW_CHAT_MESSAGE);
if (strcmp ("badword", msg) == 0) {

But here also the bot only reacts when I say "badword" and not when saying "hello badword".

I must admit i'm not very good in C++, more like just beginning. I've worked myself trough some examples and the tutorials and faq on this website, but couldn't find an answer for this.
If any one knows a place where this is explained then I would love to hear!

**achacha** · 08-02-2005

strstr() finds a string in a string.

**Tonto** · 08-02-2005

Unless we are really bound by char*'s here, I would recommend std::string for C++ programming. And if we made this conversion, we could use the find to search for badword substrings. But with these kind of things if I said...um...like, "bass" it would target that. So what you could do is tokenize the string using the stringstream technqiue described in our very own FAQ entry on the subject

**elad** · 08-02-2005

There are always tradeoffs in doing such a search. Looking for substrings that you want to indentify within the original string is the basic principle (achacha gives one example of how this can be done), but what if the "badword" is embedded in another word. Would you want to identify "assiduous" if the first three letters met screening criteria? Still, parsing strings for substrings is a time honored and often used algorhithm with many alternative strategies based on individual needs, so continue on your journey; just beware you may find potholes along the way.

**lode** · 08-02-2005

Thanks for your answers!

I now use this piece of code:

Code:

   char* msg = aw_string

   (AW_CHAT_MESSAGE);
   if (!strstr (msg,"badword") == 0) {

   sprintf (message, "you shouldn't say that", aw_string (AW_AVATAR_NAME), (WORLD));

Which works perfectly, it reacts on "badword", "hi badword" ánd "hibadword". Just what I need

But now another question comes up. Some people will report words that they normaly use which will conflict with my "badword" search. So it should be possible to also exclude words out. Can this be done by after this check, checking for words that are ok and then skip the warning?

Now I'm thinking about this... maybe working with lists of words is more efficient, like in a badwords (and goodwords) lists, just in a .txt file.
I've looked at the ifstream tutorial on the website, that doesn't seem very difficult

I could read and write to files.

But how do I get to check if the message (said in the chat) contains any word from that file? I know a bit of the array command from php, can I put the words in the file in some kind of array and then check the chat message for that array?

**~~ILoveVectors~~** · 08-02-2005

id use a string and find with a vector holding all possiblies to
search for, and have the vertor filled by a file contain the words
to be edited out. but making a file to contain all possible words
will take more effort then right this code.
if your interested in this but dont have an idea on what the code
woudl look like, i can/will provide you with a sample.

**Dweia** · 08-02-2005

using c++ strings and find

Code:

if(message.find(badword, 0) != string::npos)
{
    //stuff to do here
}

That will find in a word. If you didn't want to find in a word, only if the word's by itself, could try just adding on a space to the start and end and searching for that

Code:

badword = " "+badword+" ";
if(msg.find...

checking for exceptions can also be done

Code:

if(msg.find(badword, 0) != string::npos && msg.find(exception) == string::npos)

I'll show you an example using a file of bad words later on today maybe if I get time and noone else has, but it's easy enough to do.

**Dweia** · 08-02-2005

Ok, from a file, it's easy enough. You'd just have your file of bad words, in a lovely little list separated by new lines or spaces, then just read them into an array.

Code:

ifstream in_file;
in_file.open("badwords.txt");
string bad_words[500];
int ncnt = 0;
while(in_file.good())
{
    in_file >> bad_words[ncnt];
    ncnt++;
}

Simple

. Then from there you can just check against them in a loop.

**lode** · 08-03-2005

Thanks Dweia! This works ok now.
Only I haven't got much of an idea about how to check this in a loop. The tutorials don't provide me the information I'm looking for (or I'm not looking in a good way). Can someone help me further with any example of how to check the said chat message on the array of badwords?

**Dweia** · 08-03-2005

Code:

while(whatever)
{
    if(msg.find(bad_words[ncnt], 0) != string::npos)
    {
        //your stuff here
    }
    ncnt++;
}

**elad** · 08-04-2005

Other examples:

Goal: search a vector of strings to see if current string is in vector of strings and if so do something, otherwise do nothing.

Code:

//USING C++ with STL stuff
std::vector<string> badwords;
//fill badwords with individual words that will trigger an action here
std::ifstream fin("myFile.txt");
std::string temp;
std::vector<string>::iterator start = badwords.begin();
std::vector<string>::iterator stop = badwords.end();
while(fin >> temp)
{
    for( ; start != stop; ++start)
    {
       if(temp == *start)
          cout << "Naughty, naughty." << std::endl;
   }
}

Code:

//USING C++ without STL 
char badwords[100][20];
//fill badwords with desired vocabulary here
ifstream fin("myFile.txt");
char temp[20];
int i = 0;
while(fin >> temp)
{
   for(i = 0; i < 100; ++i)
  {
     if(strncmp(temp, badwords[i]) == 0)
        cout << "Naughty, naughty" << endl;
  }
}

**Rashakil Fol** · 08-04-2005

Originally Posted by lode

Thanks for your answers!

I now use this piece of code:

Code:

   char* msg = aw_string

   (AW_CHAT_MESSAGE);
   if (!strstr (msg,"badword") == 0) {

   sprintf (message, "you shouldn't say that", aw_string (AW_AVATAR_NAME), (WORLD));

Which works perfectly, it reacts on "badword", "hi badword" ánd "hibadword". Just what I need

Does it filter "Badword"
What about "baDword"?
What if they write "baDw0rd"?
What about "B A D W O R D"?

And there are things like fvck and f\/ck. And f\_/ck.

I recommend at least converting characters to lowercase before comparing them. This should at least stop the idiots who curse because it's thrilling or something, without disrupting the "geniunely profane."

**lode** · 08-06-2005

Originally Posted by elad

Other examples:
...

I tried many different ways of making this while, but didn't manage to get it working, accept for this last one you provided. I now got this part:

Code:

char badwords[20][20];
ifstream file_bwchk ("badwords.txt");
char temp[20];
strcpy (temp, chattext);
int i = 0;
cout << temp << endl;
while(file_bwchk >> temp) {
 for(i = 0; i < 20; ++i) {
   if(strncmp(temp, badwords[i]) == 0) {
     cout << "Naughty, naughty" << endl;
   }
 }
}

This one gives me this error:
/usr/include/string.h:100: error: too few arguments to function `int
strncmp(const char*, const char*, unsigned int)'
Just adding an int there won't help.. I guess I have to use that last int somewhere, but I can't exactly figure out where. When not using that int anymore than in the required line the console returns "Naughty, naughty" all the time, also when saying "test" or something like that.
Do you know how I have to use that extra int?

Originally Posted by Rashakil Fol

I recommend at least converting characters to lowercase before comparing them. This should at least stop the idiots who curse because it's thrilling or something, without disrupting the "geniunely profane."

You're correct.. I didn't checked that

I searched on this site, other c++ sites, google and more, but provided examples didn't get me further than making text like "TEST" translating to "tEST" or "tEsT"...

This is where I make the variable:

Code:

char chattext[256];
char chatname[256];
strcpy (chattext, aw_string (AW_CHAT_MESSAGE));
strcpy (chatname, aw_string (AW_AVATAR_NAME));

And with these two methods (the only two I got working without errors):
["tEST"]

Code:

if (strcmp (chattext,"TEST") == 0) {

printf ("Before conversion: %s\n", chattext);
*chattext = tolower(*chattext);
printf ("After conversion: %s\n", chattext);
}

"tEsT"

Code:

if (strcmp (chattext,"TEST") == 0) {
printf ("Before conversion: %s\n", chattext);
for (char *iter = chattext; *iter != '\0'; ++iter) {
    *iter = tolower(*iter);
    ++iter;
}
printf ("After conversion: %s\n", chattext);
}

But they but don't fulle lowercase the whole word.. Is there something wrong with the translation process, or do I use the wrong type of variable?

**Rashakil Fol** · 08-06-2005

Originally Posted by lode

Code:

for (char *iter = chattext; *iter != '\0'; ++iter) {
    *iter = tolower(*iter);
    ++iter;
}

You're incrementing iter twice each time through the loop. This should only happen once.

**lode** · 08-06-2005

Originally Posted by Rashakil Fol

You're incrementing iter twice each time through the loop. This should only happen once.

Ah thanks! I was a bit confused by the *iter != '\0'; part (i didn't know what '\0' meant) so I didn't read the rest quite well :$

Thread: simple string, search in input

Thread Tools

Search Thread

Display

simple string, search in input

Similar Threads

char Handling, probably typical newbie stuff

Checking array for string

Message class Need help befor 12am tonight

Another overloading "<<" problem

........ed off at functions