Calculate frequency of words in text.

freqs.cpp

// requires C++11, compile with -std=c++11 flag
#include <string>
#include <unordered_map>
#include <vector>
#include <fstream>
#include <iostream>
#include <sstream>
#include <algorithm>

class Counter
{
	std::string word_;
	int value_;

public:
	Counter() {}

	Counter(const std::string &word) : word_(word), value_(1) {}

	void increase()
	{
		value_ = value_ + 1;
	}

	const std::string &word() const
	{
		return word_;
	}

	int value() const
	{
		return value_;
	}
};

int main(int argc, char const *argv[])
{
	if (argc != 3)
	{
		std::cout << "usage: freqs in.txt out.txt" << std::endl;
		return 0;
	}

	std::ifstream input(argv[1], std::ios::binary);
	if (!input)
	{
		std::cout << "error: can't read input file" << std::endl;
		return -1;
	}

	std::unordered_map<std::string, Counter> map;
	std::stringstream stream;
	while (input)
	{
		char c = input.get();
		if (c >= 'a' && c <= 'z')
		{
			stream << c;
		}
		else if (c >= 'A' && c <= 'Z')
		{
			stream << char(c + 32);
		}
		else
		{
			std::string word = stream.str();
			if (word.size() != 0)
			{
				auto p = map.find(word);
				if (p != map.end())
				{
					p->second.increase();
				}
				else
				{
					map[word] = Counter(word);
				}
				stream.str(std::string());
			}
		}
	}

	std::vector<Counter> v;
	for (auto &e : map)
	{
		v.push_back(e.second);
	}

	std::sort(v.begin(), v.end(), [](const Counter &a, const Counter &b) {
		if (a.value() > b.value()) return true;
		if (a.value() < b.value()) return false;
		return a.word() < b.word();
	});

	std::ofstream output(argv[2]);
	if (!output)
	{
		std::cout << "error: can't create output file" << std::endl;
		return -1;
	}

	for (auto &e : v)
	{
		output << e.value() << " " << e.word() << std::endl;
	}
	
	return 0;
}