/* index.cpp - Matt Mahoney, mmahoney@cs.fit.edu

This program reads a file (specified on the command line, or from
standard input if no file name is given) and for each word, prints
the first 10 line numbers on which it is found at least once.  A word
is any sequence of letters (a-z).  Upper and lower case are equivalent.
The index is printed to standard output with the words in lower
case, listed alphabetically, one word per line, followed by a list
of up to 10 line numbers (separated by spaces), and ... if the word
appears on more than 10 lines.
*/

#include <iostream>
#include <fstream>
#include <string>
#include <map>
#include <vector>
#include <cctype>
using namespace std;

typedef map<string, vector<int> > Index;  // Maps words to line numbers

// make_index(in, m) reads from the open istream in until EOF and
// constructs an index into the initialy empty Index m.
// Each entry is a lowercase word mapped to a list of up to 11 unique line
// numbers in ascending order.

void make_index(istream& in, Index& m) {
  string word;  // Current word
  char c;       // Current char
  int line=1;   // Line number
  while (in.get(c)) {
    if (isalpha(c))
      word+=tolower(c);
    else if (word.size()>0) {  // End of word?  Store line number.
      vector<int>& v=m[word];
      if (v.size()<11 && (v.size()==0 || v.back()!=line))
        v.push_back(line);
      word="";
    }
    if (c=='\n')
      ++line;
  }
}

int main(int argc, char** argv) {

  // Open argv[1] or use cin if absent, and pass to make_index()
  Index m;  // Index
  if (argc>1) {
    ifstream in(argv[1]);
    if (!in) {
      cerr << "File not found: " << argv[1] << endl;
      return 1;
    }
    make_index(in, m);
  }
  else
    make_index(cin, m);

  // Print the index, replacing 11'th line number with ...
  for (Index::const_iterator p=m.begin(); p!=m.end(); ++p) {
    cout << p->first;
    for (int i=0; i<p->second.size(); ++i) {
      if (i==10)
        cout << " ...";
      else
        cout << " " << p->second[i];
    }
    cout << "\n";
  }
  return 0;
}