locating matched words in a string

Question

I have a file A that has multiple paragraphs. I need to identify where I matched words from another file B. I need to tell the paragraph, line number, and word number of every word, including those matching a word in file B. I've finally gotten so far, having given up on vectors, and arrays, and string splitting. I learned (I think) stringstream. Currently, I read in the line, then split it on the "." into sentences, then read those sentences back in again, splitting on the " ". I have the line numbers counting, and the words counting and matching, but I just can't seem to get the paragraph numbers (I've realized that the p++ is actually counting the lines, and the l++ is counting words as well). Could someone please help me? edit Each paragraph is separated by "\n" and each sentence is separated by a "." I'll still have to figure out a way to ignore all other punctuation so that words match 100%, and are not thrown off by a comma, semi-colon, or other punctuation. I'm guessing that will be a regex in there somewhere.

input from file with the text would look like:

    My dog has fleas in his weak knees. This is a line.  The paragraph is ending.'\n'
    Fleas is a word to be matched.  here is another line.  The paragraph is ending.'\n'

output should look something like:

    paragraph1 line 1 word 1  My
    paragraph1 line 1 word 2  dog
    paragraph1 line 1 word 3  has
    paragraph1 line 1 word 4  MATCHED!  fleas

while (getline(fin, para)) { //get the paragraphs
    pbuffer.clear();
    pbuffer.str("."); //split on periods
    pbuffer << para;
    p++; //increase paragraph number

    while (pbuffer >> line) { //feed back into a new buffer

        lbuffer.clear();
        lbuffer.str(" "); //splitting on spaces
        lbuffer << line;
        l++; //line counter

        while (lbuffer >> word) { //feed back in
            cout << "l " << l << "   W:  " << w << "   " << word;
            fmatch.open("match.txt");
            while (fmatch >> strmatch) {  //did I find a match?
                if (strmatch.compare(word) == 0) {
                    cout << "  Matched!\n";
                }
                else {
                    cout << "\n";
                }

            }

Answer 1

Since you say that you can write each word on read, we won't bother with a collection. We'll just use istringstream and istream_iterator and counter the indices.
Assuming that fin is good, I'm going to simply write to cout you can make the appropriate adjustments to write to your file.

1^st you'll need to read in your "fmatch.txt" into a vector<string> like so:

const vector<string> strmatch{ istream_iterator<string>(fmatch), istream_iterator<string> }

Then you'll just wanna use that in a nested loop:

string paragraph;
string sentence;

for(auto p = 1; getline(fin, paragraph, '\n'); ++p) {
    istringstream sentences{ paragraph };

    for(auto s = 1; getline(sentences, sentence, '.'); ++s) {
        istringstream words{ sentence };

        for_each(istream_iterator<string>(words), istream_iterator<string>(), [&, i = 1](const auto& word) mutable { cout << 'w' << i++ << ", p" << p << ", s" << s << (find(cbegin(strmatch), cend(strmatch), word) == cend(strmatch) ? ", word, " : ", namedEntity, ") << word << endl; });
    }
}

Live Example

EDIT:

By way of explaination, I'm using a for_each to call a lambda on each word in the sentence.

Let's break apart the lambda and explain what each section does:

• [& This exposes, by reference, any variable in the scope in which the lambda was declared to the lambda for use: http://en.cppreference.com/w/cpp/language/lambda#Lambda_capture Because I'm using strmatch, p, and s in the lamda those will be captured by reference
• , i = 1] C++14 allowed us to declare a variable in the lambda capture of type auto so i is an int which will be reinitialized each time the scope in which the lambda is declared is rentered, here that's each entry into the body of the nested for-loop
• (const auto& word) This is the parameter list passed into the lambda: http://en.cppreference.com/w/cpp/language/lambda Here for_each will just be passing in strings
• mutable Because I'm modifying i, which is a owned by the lambda, I need it to be non-const so I declare the lambda mutable

In the lambda's body I'll just use find with standard insertion operators to write the values.

EDIT2:

If you're limited to C++11, you won't be able to declare a variable in the lambda capture. You can just provide that externally:

string paragraph;
string sentence;

for(auto p = 1; getline(fin, paragraph, '\n'); ++p) {
    istringstream sentences{ paragraph };

    for(auto s = 1; getline(sentences, sentence, '.'); ++s) {
        istringstream words{ sentence };
        auto i = 1;

        for_each(istream_iterator<string>(words), istream_iterator<string>(), [&](const auto& word){ cout << 'w' << i++ << ", p" << p << ", s" << s << (find(cbegin(strmatch), cend(strmatch), word) == cend(strmatch) ? ", word, " : ", namedEntity, ") << word << endl; });
    }
}

Answer 2

I did finally figure it out, but I didn't use the stream interator (sorry!) And it's certainly not as elegant @jonathanMee

I vectored the matching words and used string stream to read in the characters nesting it as I went. I then used an if statment to check for paragraphs, and delimited as I poured the data from one string to another using string stream. I incremented when I delimited the data, and the did the match. EXAMPLE:

            pholder.clear();
            pholder.str("."); //break on the delimiter
            pholder << para; //read from the paragraph into pholder
            l++;

            while (pholder >> line) {// here are all my lines now

                lholder.clear();
                lholder.str(" "); //breka on the spaces
                lholder << line; //read for it