parse_book.cc

#include <iostream>
#include <fstream>
#include <vector>
#include <sstream>

using namespace std;

// TODO: Explain how this entire jumble of code works clearly and precisely.
vector<string> ParseWords(const string &line) {
  vector<string> tokens;
  stringstream stream(line);
  string token;

  // Custom predicate to make sure we handle unicode characters properly.
  std::function<bool(char)> is_alpha_char = [](char character) {
    return (character >= 'a' && character <= 'z') || (character >= 'A' && character <= 'Z');
  };

  while (getline(stream, token, ' ')) {
    // Verify that the string is not empty, and that it contains at least
    // one alphabetical character (lowercase or uppercase), before adding to list.
    if (!token.empty() && any_of(begin(token), end(token), is_alpha_char)) {
      token.erase(remove_if(begin(token), end(token), not1(is_alpha_char)), end(token));
      // We check once more to make the string isn't empty, after trimming special characters.
      if (!token.empty()) tokens.push_back(token);
    } // else, it's an empty string DoNothing();
  }
  return tokens;
}

// TODO: Write a comment here about what this code block is doing, and how it works.
int main() {
  string line;
  int line_number = 1;
  ifstream infile("../book.txt");
  if (infile.is_open()) {
    while (getline(infile, line)) {
      vector<string> words = ParseWords(line);
      cout << line_number << ": " << words.size() << endl;
      line_number++;
    }
    infile.close();
  } else cout << "An error occurred when trying to load the specified input file." << endl;
  return 0;
}