#include <iostream>
#include <string>
#include <sstream>
#include <algorithm>
#include "Structures.h"
#include "util.h"

using namespace std;

std::ostream& operator<< (std::ostream& os, const Entry& entry) {
    os << "userId: " << entry.getUserId() << ", rating: " << (int) entry.getRating() << ", movieId: " << entry.getMovieId();
    return os;
}

void addUserAverageRatingLine(string line, std::map<int,AverageRating>& ratingMap) {
    // First find the userId, then parse the rest of the line.
    string::size_type firstComma = line.find(",");
    if (firstComma != string::npos) {
        int userId = from_string<int>(line.substr(0, firstComma));
        AverageRating rating = parseMovieAverageRatingLine(line.substr(firstComma+1));
        ratingMap.insert(std::make_pair(userId, rating));
        return;
    } else {
        cout << "addUserAverageRatingLine: Didn't find first comma in line: " << line << endl;
        return;
    }
}

void addUserCorrelationLine(string line, std::map<int,double>& correlations) {
    // First find the userId, then parse the rest of the line.
    string::size_type firstComma = line.find(",");
    if (firstComma != string::npos) {
        int userId = from_string<int>(line.substr(0, firstComma));
        double correlation = from_string<double>(line.substr(firstComma+1));
        correlations.insert(std::make_pair(userId, correlation));
        return;
    } else {
        cout << "addUserCorrelationLine: Didn't find first comma in line: " << line << endl;
        return;
    }
}



AverageRating parseMovieAverageRatingLine(string line) {
    string::size_type firstComma = line.find(",");
    if (firstComma != string::npos) {
        double averageRating;
        averageRating = from_string<double>(line.substr(0, firstComma));
        string::size_type secondComma = line.find(",", firstComma + 1);
        if (secondComma != string::npos) {
            int numRatings = from_string<int>(line.substr(firstComma + 1, secondComma));
            double stdDev = from_string<double>(line.substr(secondComma + 1));
            return AverageRating(averageRating, numRatings, stdDev);

        } else {
            cout << "parseMovieAverageRatingLine: Didn't find second comma in line: " << line << endl;
            return AverageRating(0.0, 0, 0.0);
        }
    } else {
        cout << "parseMovieAverageRatingLine: Didn't find first comma in line: " << line << endl;
        return AverageRating(0.0, 0, 0.0);
    }
}

void setEntriesFromMovieAverageRatings(bool isProbe, std::vector<AverageRating>& entries) {
    entries.clear();
    // Insert something in the 0th slot.
    entries.push_back(AverageRating());
    string fileName;
    if (isProbe) {
        fileName = probeDataDir;
    } else {
        fileName = predictionDataDir;
    }
    fileName += "movieAverageRatings.txt";
    ifstream file(fileName.c_str());
    if (!file) {
        cout << "ERROR: setEntriesFromMovieAverageRatings: couldn't open " << fileName.c_str() << endl;
    }
    string line;
    getline(file, line);
    while (!file.eof()) {
        entries.push_back(parseMovieAverageRatingLine(line));
        getline(file, line);
    }
}

void appendEntriesFromUserAverageRatings(bool isProbe, std::map<int, AverageRating>& entries) {
    string fileName;
    if (isProbe) {
        fileName = probeDataDir;
    } else {
        fileName = predictionDataDir;
    }
    fileName += "userAverageRatings.txt";
    ifstream file(fileName.c_str());
    if (!file) {
        cout << "ERROR: appendEntriesFromUserAverageRatings: couldn't open " << fileName.c_str() << endl;
    }
    string line;
    getline(file, line);
    while (!file.eof()) {
        addUserAverageRatingLine(line, entries);
        getline(file, line);
    }
}

void appendDataFromMovieCorrelations(bool isProbe, int movieId, std::vector<double>& correlations) {
    string fileToOpen = "movieCorrelations/" + formatNumber(movieId/1000, 2) + "/" + formatNumber(movieId, 7) + ".txt";
    if (isProbe) {
        fileToOpen = "probe/" + fileToOpen;
    }
    correlations.clear();
    correlations.reserve(numMovies + 1);
    // Insert something in the 0th slot.
    correlations.push_back(0.00);
    ifstream file(fileToOpen.c_str());
    if (!file) {
        cout << "ERROR: appendDataFromMovieCorrelations: couldn't open " << fileToOpen.c_str() << endl;
    }
    string line;
    getline(file, line);
    while (!file.eof()) {
        correlations.push_back(from_string<double>(line));
        getline(file, line);
    }
    file.close();
    if (correlations.size() != numMovies + 1) {
        cout << "Got wrong number of entries from correlation file " << fileToOpen << "! (expected " << numMovies + 1 << ", got " << correlations.size() << ")" << endl;
    }
}

void appendDataFromMovieCorrelations(bool isProbe, int movieId, std::vector<MovieCorrelation>& correlations) {
    string fileToOpen = "movieCorrelations/" + formatNumber(movieId/1000, 2) + "/" + formatNumber(movieId, 7) + ".txt";
    if (isProbe) {
        fileToOpen = "probe/" + fileToOpen;
    }
    correlations.clear();
    correlations.reserve(numMovies);
    ifstream file(fileToOpen.c_str());
    if (!file) {
        cout << "ERROR: appendDataFromMovieCorrelations: couldn't open " << fileToOpen.c_str() << endl;
    }
    string line;
    getline(file, line);
    int mId = 1;
    while (!file.eof()) {
        correlations.push_back(MovieCorrelation(mId, from_string<double>(line)));
        ++mId;
        getline(file, line);
    }
    file.close();
    if (correlations.size() != numMovies) {
        cout << "Got wrong number of entries from correlation file " << fileToOpen << "! (expected " << numMovies << ", got " << correlations.size() << ")" << endl;
    }
}



void appendDataFromUserCorrelations(bool isProbe, int userId, std::map<int,double>& correlations) {
    string fileToOpen = "userCorrelations/" + formatNumber(userId%100, 2) + "/" + formatNumber(userId, 7) + ".txt";
    if (isProbe) {
        fileToOpen = "probe/" + fileToOpen;
    }
    correlations.clear();
    ifstream file(fileToOpen.c_str());
    string line;
    if (!file) {
        cout << "ERROR: appendDataFromUserCorrelations: couldn't open " << fileToOpen.c_str() << endl;
        exit(1);
    }
    getline(file, line);
    while (!file.eof()) {
        addUserCorrelationLine(line, correlations);
        getline(file, line);
    }
    file.close();
}



void appendEntriesFromProbeFile(std::map<int, vector<Entry> >& entries) {
    ifstream file(probeFileName.c_str());
    if (!file) {
        cout << "ERROR: appendEntriesFromProbeFile: couldn't open " << probeFileName.c_str() << endl;
    }
    string line;

    int movieId = -1;
    bool changedId;
    getline(file, line);
    while (!file.eof()) {
        Entry e = parseChallengeLine(line, movieId, changedId, false);
        if (!changedId) {
            entries[movieId].push_back(e);
        }
        getline(file, line);
    }
}

Entry parseChallengeLine(string line, int& movieId, bool& changedId, bool hasDates) {
    if (line[line.size() - 1] == ':') {
        // This is a movie line.
        string restOfLine = line.substr(0, line.size() - 1);
        movieId = from_string<int>(restOfLine);
        changedId = true;
        return Entry(0,0,0);
    }
    changedId = false;

    if (hasDates) {
        string::size_type firstComma = line.find(",");
        if (firstComma != string::npos) {
            int userId;
            userId = from_string<int>(line.substr(0, firstComma));
            return Entry(userId, 0, movieId);
        } else {
            cout << "Didn't find first comma in line: " << line << endl;
            return Entry(0,0,0);
        }
    } else {
        int userId = from_string<int>(line);
        if (userId == 2466194 && movieId == 9939) {
            cout << "Got bad entry from line: " << line;
        }
        return Entry(from_string<int>(line), 0, movieId);
    }
}

void filterEntries(vector<Entry>& entries, map<int, vector<Entry> >& probeEntriesMap) {
    vector<Entry>::const_iterator probeIt;
    vector<Entry>::iterator it;
    int movieId = entries[0].getMovieId();
    vector<Entry>& probeEntries = probeEntriesMap[movieId];
    for (probeIt = probeEntries.begin(); probeIt != probeEntries.end(); ++probeIt) {
        it = std::find_if(entries.begin(), entries.end(), ResultComparator(*probeIt));
        if (it != entries.end()) {
            entries.erase(it);
        } else {
            cout << "Couldn't find entry for movieId " << movieId << "! " << *probeIt << endl;
            abort();
        }
    }
}

Entry parseRatingLine(string line, int& movieId) {
    string::size_type firstComma = line.find(",");
    if (firstComma != string::npos) {
        string::size_type secondComma = line.find(",", firstComma + 1);
        if (secondComma != string::npos) {
            int userId = from_string<int>(line.substr(0, firstComma));
            char rating = from_string<char>(line.substr(firstComma+1, secondComma));
            return Entry(userId, rating, movieId);
        } else {
            cout << "parseRatingLine: Didn't find second comma in line: " << line << endl;
            return Entry(0,0,0);
        }
    } else {
        cout << "parseRatingLine: Didn't find first comma in line: " << line << endl;
        return Entry(0,0,0);
    }
}

void appendEntriesFromMovieFile(int movieId, std::vector<Entry>& entries, bool isBinary) {
    string fileName = makeTrainingFileName(movieId);
    {
        ifstream file(fileName.c_str());
        if (!file) {
            cout << "ERROR: appendEntriesFromMovieFile: couldn't open " << fileName.c_str() << endl;
            cout << "movieId was " << movieId << endl;
            exit(1);
        }
        if (isBinary) {
            union {
                char charRep[4];
                int intRep;
            } charIntUnion;
            int userId;
            char rating;
            file.read(charIntUnion.charRep, 4);
            while (!file.eof()) {
                userId = charIntUnion.intRep;
                file.read(&rating, 1);
                entries.push_back(Entry(userId, rating, movieId));
                file.read(charIntUnion.charRep, 4);
            }
        } else {
            string line;
            // Read the "header" line.
            getline(file, line);

            getline(file, line);
            while (!file.eof()) {
                entries.push_back(parseRatingLine(line, movieId));
                getline(file, line);
            }
        }
    } 
}

string makeTrainingFileName(int num) {
    return binaryEntriesDir + formatNumber(num/1000, 2) + "/" + trainingFilePrefix + formatNumber(num, 7)  + trainingFileSuffix;
}

double calculateStdDev(const std::vector<Entry>& entries, double average) {
    int numEntries = entries.size();
    double sumSquareDist = 0.0;
    vector<Entry>::const_iterator pos;
    for (pos = entries.begin(); pos != entries.end(); ++pos) {
        sumSquareDist += (pos->getRating() - average) * (pos->getRating() - average);
    }
    return sqrt(sumSquareDist/((double) numEntries)); 
}