VE280 SU2020 reference parsing

#VE280 SU2020 reference parsing

VE280 SU2020 reference parsing

1.txt

Weng, Paul; Fairness in Reinforcement Learning (2019); AI for Social Good Workshop in IJCAI
Wu, Qitian and Zheng, Hengrui and Gao, Xiaofeng and He, Peng and Weng, Pual and Gao, Han and Chen, Guihai; Dual Graph Attention Networks for Deep Latent Representation of Multifaceted Social Effects in Recommender Systems (2019); The Web Conference

cout

Fairness in Reinforcement Learning - P. Weng - AI for Social Good Workshop in IJCAI 2019
Dual Graph Attention Networks for Deep Latent Representation of Multifaceted Social Effects in Recommender Systems - Q. Wu; H. Zheng; X. Gao; P. He; P. Weng; H. Gao; G. Chen - The Web Conference 2019

main.cpp

#include <iostream>
#include <fstream>
#include <string>
#include <cstring>
#include <sstream>
#include <vector>

using namespace std;

class Author {
private:
    string firstName, lastName;
public:
    explicit Author(const string &name) {
        auto pos = name.find(',');
        if (pos != string::npos) {
            firstName = name.substr(0, pos);
            lastName = name.substr(pos + 2);
        }
    }

    friend ostream &operator<<(ostream &os, const Author &author);
};


class Reference {
private:
    vector<Author> authors;
    string title;
    string year;
    string conference;
public:
    explicit Reference(const string &line) {
        istringstream iss(line);
        string segment;

        // parse authors
        getline(iss, segment, ';');
        size_t start = 0, end = 0;
        while (end != string::npos) {
            end = segment.find(" and ", start);
            auto name = segment.substr(start, end - start);
            authors.emplace_back(name);
            start = end + 5;
        }

        // parse title and year
        getline(iss, segment, ';');
        start = segment.find_first_not_of(' ');
        end = segment.find_last_of(" (");
        title = segment.substr(start, end - start - 1);
        if (end != string::npos) {
            start = end + 1;
            end = segment.find_last_of(')');
            year = segment.substr(start, end - start);
        }

        // parse conference
        getline(iss, segment, ';');
        start = segment.find_first_not_of(' ');
        conference = segment.substr(start);
    }

    friend ostream &operator<<(ostream &os, const Reference &ref);
};

ostream &operator<<(ostream &os, const Reference &ref) {
    os << ref.title << " - ";
    for (size_t i = 0; i < ref.authors.size(); i++) {
        os << ref.authors[i];
        if (i != ref.authors.size() - 1) {
            os << "; ";
        }
    }
    os << " - " << ref.conference << " " << ref.year;
    return os;
}

ostream &operator<<(ostream &os, const Author &author) {
    if (!author.lastName.empty()) {
        os << author.lastName[0] << ". " << author.firstName;
    }
    return os;
}

int main() {
    ifstream fin("1.txt");
    string line;
    while (getline(fin, line)) {
        cout << Reference(line) << endl;
    }
    return 0;
}