RatingParser.cpp

#include <iostream>
#include <fstream>
#include <vector>
#include <string>
#include <map>
#include <algorithm>
#include <ctime>
using namespace std;

typedef long long ll;

class RatingParser{
private:
    ifstream fin;
    ofstream fout;
    clock_t start;
    const string mcFileName = "mc.txt";
    const string ratingFileName = "rating.txt";
    const string outputFileName = "rating_weight.txt";
public:
    map<ll, ll> contentToUser;
    map<ll, int> ratingWeight;
    map<pair<ll, ll>, pair<int, int> > userToUserRating; // key: {useri, userj}, value: {rating sum, rating count}
    void run () {
        
        clockInit("mc parsing");
        fin.open(mcFileName);
        parseMC();
        fin.close();
        printDurating("mc parse");
        
        clockInit("rating parsing");
        fin.open(ratingFileName);
        parseRating();
        fin.close();
        printDurating("rating parsing");
        
        clockInit("make output");
        fout.open(outputFileName);
        makeOutput();
        fout.close();
        printDurating("make output");
    }
    
    void clockInit(string name) {
        cout << "[" << name << "] :: Start" <<  endl;
        start = clock();
    }
    void printDurating(string name) {
        cout << "["<<name << "] :: End / duration : " << (clock() - start ) / (double) CLOCKS_PER_SEC << endl;
    }
    
    void parseMC() {
        string str;
        while(!getline(fin, str).eof()){
            vector<string> row;
            int pre = 0;
            for(int i=0;i<str.size();i++){
                if(str[i] == '|') {
                    string col = str.substr(pre, i-pre);
                    
                    row.push_back(col);
                    pre = i+1;
                }
            }
            ll contentId = stoll(row[0]);
            ll userId = stoll(row[1]);
            contentToUser.insert({contentId, userId});
        }
    }
    void parseRating() {
        string str;
        while(!getline(fin, str).eof()){
            vector<string> row;
            int pre = 0;
            for(int i=0;i<str.size();i++){
                if(str[i] == '\t') {
                    string col = str.substr(pre, i-pre);
                    
                    row.push_back(col);
                    pre = i+1;
                }
            }
            ll contentId = stoll(row[0]);
            ll userId = stoll(row[1]);
            int rating = stoi(row[2]);
            
            ll authorId = contentToUser[contentId];
            
            auto iter = userToUserRating.find({userId, authorId});
            if(iter==userToUserRating.end()) {
                pair<ll, ll> key = {userId, authorId};
                pair<int, int> value = {rating, 1};
                userToUserRating.insert({key, value});
            } else {
                auto value = iter->second;
                value.first += rating;
                value.second++;
                iter->second = value;
            }
            
            // rating weight 계산을 위해서
            ratingWeight[userId]++;
        }
    }
    
    void makeOutput() {
        for(auto iter=userToUserRating.begin(); iter != userToUserRating.end(); iter++) {
            ll userI = iter->first.first;
            ll userJ = iter->first.second;
            int weight = ratingWeight[userI];
            
            fout << userI<< " " << userJ<< " ";
            // fout << iter->second.first << " " << iter->second.second << endl;
            fout << weight << endl;
        }
    }
};


int main() {
    RatingParser ratingParser;
    ratingParser.run();
    return 0;
}