/* tree1.cpp - Spring 2003 CSE2050 Homework 6 part 1
   Matt Mahoney, mmahoney@cs.fit.edu

This program takes a file name and a word as arguments, then prints
the number of times the word occurs in the file.  If the word is
omitted, then it outputs a histogram of all words that occur n times,
in decreasing order of n.  In either case, it prints the number
of different words and number of word instances.

For example:

  more < test
  This is a test.  This is another test.
  This and this and this too.

  tree1 test
  1 words occur 5 times: this
  3 words occur 2 times: is test and
  3 words occur 1 time: a another too
  test has 14 instances of 7 words.

If more than 5 words occur n times, then just the first 5 are printed,
followed by ...

  8 words occur 2 times: cat dog horse worm bird ...

With two arguments, the histogram is omitted, and only the frequency
of the given word is output.

  tree1 test THIS
  this occurs 4 times.
  test has 14 instances of 7 words.

  tree1 test foo
  foo occurs 0 times.
  test has 14 instances of 7 words.

Input is not case sensitive.  A word is considered to be any sequence
of letters (a-z).  Thus, "IT'S" and "it s" are equivalent (both are 2
words).

This code tests all of the operations required for class Tree in part 3
(but using maps for now).  It tests begin(), end(), find(), size(), and
operator[].  The program also tests all forward iterator operations:
++ (pre and post), ==, !=, *, and ->.  Iterators are declared both with
and without an initial value.
*/

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

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

  // Open input file, fail if not specified or not found
  if (argc < 2) {
    cerr << "Usage: tree1 file [word]\n";
    return 1;
  }
  ifstream f(argv[1]);
  if (!f) {
    cerr << "File not found: " << argv[1] << endl;
    return 1;
  }

  // Count words
  map<string, int> count;  // count[s] is number of occurrences of word s
  int wordcount = 0;  // Number of word instances
  string s;  // Current input word
  char c;
  while (f.get(c)) {
    if (isalpha(c))
      s+=char(tolower(c));
    else if (s != "") {
      ++wordcount;
      ++count[s];
      s="";
    }
  }

  // If there is a word, report its count
  map<string, int>::iterator p;
  if (argc>2) {
    string word = argv[2];
    for (int i=0; i<int(word.size()); ++i)  // Convert word to lower case
      word[i] = tolower(word[i]);
    p=count.find(word);
    if (p == count.end())
      cout << word << " occurs 0 times\n";
    else
      cout << word << " occurs " << p->second << " times\n";
  }

  // Otherwise, print a histogram: m words occur n times... by decreasing n
  else {
    map<int, vector<string> > hist;  // hist[-n] is words occurring n times
    for (p=count.begin(); !(p==count.end()); ++p)
      hist[-(*p).second].push_back(p->first);
    for (map<int, vector<string> >::iterator q=hist.begin();
        q!=hist.end(); q++) {
      const int m = (*q).second.size();
      const int n = -q->first;
      if (m==1)
        cout << "1 word occurs ";
      else
        cout << m << " words occur ";
      if (n==1)
        cout << "1 time:";
      else
        cout << n << " times:";
      for (int i=0; i<5 && i<m; ++i)
        cout << " " << q->second[i];
      if (m > 5)
        cout << " ...";
      cout << endl;
    }
  }

  // Print overall statistics
  cout << argv[1] << " has " << wordcount << " instances of "
       << int(count.size()) << " words\n";
  return 0;
}