duplication_finder_diff.hpp

Go to the documentation of this file.

 
#pragma once
 
#include <arkanjo/utils/utils.hpp>
#include <string>
#include <tuple>
#include <vector>
#include <filesystem>
using namespace std;
namespace fs = std::filesystem;
 
class DuplicationFinderDiff {
  private:
    string SAVING_MESSAGE = "Saving results..."; 
 
    fs::path base_path;  
    double similarity; 
 
    vector<string> find_files(const fs::path& folder_path);
 
    bool is_empty_line(string line);
 
    vector<string> remove_blank_lines(vector<string> content);
 
    bool is_equal_files(vector<string> content1, vector<string> content2);
 
    double find_similarity(string path1, string path2);
 
    vector<tuple<double, string, string>> find_similar_pairs(vector<string>& file_paths);
 
    void save_duplications(vector<tuple<double, string, string>>& file_duplication_pairs);
 
  public:
    DuplicationFinderDiff(const fs::path& base_path_, double similarity_);
 
    void execute();
};