Word Counting

word_count.cpp

#include <iostream>
#include <fstream>
#include <string>
#include <unordered_map>
#include <algorithm>
#include <chrono>

using std::cout;
using std::cerr;
using std::cin;
using std::endl;
using std::string;
using Clock = std::chrono::steady_clock;

class WordCounter
{
  public:
    WordCounter() : total_count(0), dif_count(0), string_count(0) {}
    void count_word(const string& s);
    void count_string(const string& s);
    unsigned int total_count, dif_count, string_count;
    std::unordered_map<string,int> word_map;
};

// Count single word
void WordCounter::count_word(const string& s)
{
  ++total_count;
  if (++word_map[s] == 1) { ++dif_count; }
}

// Count all words in string
void WordCounter::count_string(const string& s)
{
  ++string_count;
  auto cur = s.begin(), end = s.end(), last = end;
  string word;
  while (cur != end) {
    cur = find_if(cur, end, [](char ch){ return ch >= 'a' && ch <= 'z'; });
    if (cur == end) { break; }
    last = find_if_not(cur, end, [](char ch){ return ch >= 'a' && ch <= 'z'; });
    word = string(cur, last);
    count_word(word);
    cur = last;
  }
}

// Main
int main(int argc, char* argv[])
{
  if (argc < 2) {
    cout << "usage:\n"
            "  word_count infile.txt [outfile.txt]" << endl;
    return 1;
  }

  WordCounter wc;
  Clock::time_point clock_start, clock_end;
  double elapsed_time;

  // Open source file
  {
    std::ifstream infile(argv[1]);
    if (!infile) {
      cerr << "Input file open error!" << endl;
      return 1;
    }
  
    // Read and process file
    clock_start = Clock::now();
    string s;
    while (getline(infile, s)) {
      std::transform(s.begin(), s.end(), s.begin(), tolower);
      wc.count_string(s);
    }
    clock_end = Clock::now();
    elapsed_time = std::chrono::duration_cast<std::chrono::duration<double>>(clock_end - clock_start).count();
  }

  // Word counts
  if (argc >= 3) {
    std::ofstream outfile(argv[2]);
    if (!outfile) {
      cerr << "Output file create error!" << endl;
    } else {
      for (auto& word : wc.word_map) {
        outfile << word.first << " = " << word.second << "\n";
        if (!outfile) {
          cerr << "Output file write error!" << endl;
          break;
        }
      }
      outfile << "\nNumber of lines = " << wc.string_count << "\n";
      outfile << "Number of words total = " << wc.total_count << "\n";
      outfile << "Number of different words = " << wc.dif_count << "\n";
      outfile << "Elapsed time: " << elapsed_time << " seconds\n";
    }
  }

  // Statistics
  cout << "Number of lines = " << wc.string_count << "\n";
  cout << "Number of words total = " << wc.total_count << "\n";
  cout << "Number of different words = " << wc.dif_count << "\n";
  cout << "Elapsed time: " << elapsed_time << " seconds" << endl;

  return 0;
}