Histogram.h

Go to the documentation of this file.

 
#pragma once
 
#include <algorithm>
#include <cmath>
#include <fstream>
#include <iosfwd>
#include <iterator>
#include <utility>
#include <vector>
 
#include "Logging.h"
 
namespace BaseLib
{
template <typename T>
class Histogram
{
public:
    using Data =
        typename std::vector<double>;  
 
public:
    template <typename InputIterator>
    Histogram(InputIterator first, InputIterator last, const int nr_bins = 16,
              const bool computeHistogram = true)
        : data_(first, last), nr_bins_(nr_bins), dirty_(true)
    {
        init(computeHistogram);
    }
 
    explicit Histogram(std::vector<T> data, const unsigned int nr_bins = 16,
                       const bool computeHistogram = true)
        : data_(std::move(data)), nr_bins_(nr_bins), dirty_(true)
    {
        init(computeHistogram);
    }
 
    void update()
    {
        if (!dirty_)
        {
            return;
        }
 
        bin_width_ = (max_ - min_) / nr_bins_;
 
        auto it = data_.begin();
        for (unsigned int bin = 0; bin < nr_bins_; bin++)
        {
            auto itEnd = std::upper_bound(it, data_.end(),
                                          min_ + (bin + 1) * bin_width_);
            histogram_[bin] = std::distance(it, itEnd);
            it = itEnd;
        }
        dirty_ = false;
    }
 
    void setMinimum(const T& minimum)
    {
        min_ = minimum;
        dirty_ = true;
    }
    void setMaximum(const T& maximum)
    {
        max_ = maximum;
        dirty_ = true;
    }
 
    const Data& getSortedData() const { return data_; }
    const std::vector<std::size_t>& getBinCounts() const { return histogram_; }
    const unsigned int& getNumberOfBins() const { return nr_bins_; }
    const T& getMinimum() const { return min_; }
    const T& getMaximum() const { return max_; }
    const T& getBinWidth() const { return bin_width_; }
 
    void prettyPrint(std::ostream& os, const unsigned int line_width = 16) const
    {
        const std::size_t count_max =
            *std::max_element(histogram_.begin(), histogram_.end());
        for (unsigned int bin = 0; bin < nr_bins_; ++bin)
        {
            os << "[" << min_ + bin * bin_width_ << ", "
               << min_ + (bin + 1) * bin_width_ << ")\t";
            os << histogram_[bin] << "\t";
 
            const int n_stars =
                std::ceil(line_width * ((double)histogram_[bin] / count_max));
            for (int star = 0; star < n_stars; star++)
            {
                os << "*";
            }
            os << "\n";
        }
    }
 
    int write(std::string const& file_name, std::string const& data_set_name,
              std::string const& param_name) const
    {
        if (file_name.empty())
        {
            ERR("No file name specified.");
            return 1;
        }
 
        std::ofstream out(file_name);
        if (!out)
        {
            ERR("Error writing histogram: Could not open file.");
            return 1;
        }
 
        out << "# Histogram for parameter " << param_name << " of data set "
            << data_set_name << "\n";
        std::size_t const n_bins = this->getNumberOfBins();
        std::vector<std::size_t> const& bin_cnts(this->getBinCounts());
        double const min(this->getMinimum());
        double const bin_width(this->getBinWidth());
 
        for (std::size_t k(0); k < n_bins; k++)
        {
            out << min + k * bin_width << " " << bin_cnts[k] << "\n";
        }
        out.close();
        return 0;
    }
 
protected:
    void init(const bool computeHistogram = true)
    {
        std::sort(data_.begin(), data_.end());
        histogram_.resize(nr_bins_);
        min_ = data_.front();
        max_ = data_.back();
        bin_width_ = (max_ - min_) / nr_bins_;
 
        dirty_ = true;
        if (computeHistogram)
        {
            update();
        }
    }
 
    Data data_;
    const unsigned int nr_bins_;
    std::vector<std::size_t> histogram_;
    T min_, max_;  
    T bin_width_;
 
private:
    bool dirty_;  
};
 
template <typename T>
std::ostream& operator<<(std::ostream& os, const Histogram<T>& h)
{
    os << h.getNumberOfBins() << " " << h.getMinimum() << " " << h.getMaximum()
       << " ";
    std::copy(h.getBinCounts().begin(), h.getBinCounts().end(),
              std::ostream_iterator<T>(os, " "));
    return os << std::endl;
}
}  // namespace BaseLib