Спринт 2. Средний рейтинг. Задание 2.

#include <algorithm>
#include <cmath>
#include <iostream>
#include <map>
#include <set>
#include <string>
#include <utility>
#include <vector>

using namespace std;

const int MAX_RESULT_DOCUMENT_COUNT = 5;

string ReadLine() {
    string s;
    getline(cin, s);
    return s;
}

int ReadLineWithNumber() {
    int result;
    cin >> result;
    ReadLine();
    return result;
}

vector<string> SplitIntoWords(const string& text) {
    vector<string> words;
    string word;
    for (const char c : text) {
        if (c == ' ') {
            if (!word.empty()) {
                words.push_back(word);
                word.clear();
            }
        } else {
            word += c;
        }
    }
    if (!word.empty()) {
        words.push_back(word);
    }

    return words;
}



struct Document {
    int id;
    double relevance;
    int rating;
};

class SearchServer {
public:
    void SetStopWords(const string& text) {
        for (const string& word : SplitIntoWords(text)) {
            stop_words_.insert(word);
        }
    }

    void AddDocument(int document_id, const string& document, const vector<int>& ratings) {
        ++document_count_;
        const vector<string> words = SplitIntoWordsNoStop(document);
        const double inv_word_count = 1.0 / words.size();
        for (const string& word : words) {
            word_to_document_freqs_[word][document_id] += inv_word_count;
            document_ratings_[document_id] = ComputeAverageRating(ratings);
        }
    }

    vector<Document> FindTopDocuments(const string& raw_query) const {
        const Query query = ParseQuery(raw_query);
        auto matched_documents = FindAllDocuments(query);

        sort(matched_documents.begin(), matched_documents.end(),
             [](const Document& lhs, const Document& rhs) {
                 return lhs.relevance > rhs.relevance;
             });
        if (matched_documents.size() > MAX_RESULT_DOCUMENT_COUNT) {
            matched_documents.resize(MAX_RESULT_DOCUMENT_COUNT);
        }
        return matched_documents;
    }

private:
    int document_count_ = 0;
    set<string> stop_words_;
    map<string, map<int, double>> word_to_document_freqs_;
    map<int, int> document_ratings_;

    bool IsStopWord(const string& word) const {
        return stop_words_.count(word) > 0;
    }

    vector<string> SplitIntoWordsNoStop(const string& text) const {
        vector<string> words;
        for (const string& word : SplitIntoWords(text)) {
            if (!IsStopWord(word)) {
                words.push_back(word);
            }
        }
        return words;
    }

    struct QueryWord {
        string data;
        bool is_minus;
        bool is_stop;
    };

    QueryWord ParseQueryWord(string text) const {
        bool is_minus = false;
        // Word shouldn't be empty
        if (text[0] == '-') {
            is_minus = true;
            text = text.substr(1);
        }
        return {text, is_minus, IsStopWord(text)};
    }

    struct Query {
        set<string> plus_words;
        set<string> minus_words;
    };

    Query ParseQuery(const string& text) const {
        Query query;
        for (const string& word : SplitIntoWords(text)) {
            const QueryWord query_word = ParseQueryWord(word);
            if (!query_word.is_stop) {
                if (query_word.is_minus) {
                    query.minus_words.insert(query_word.data);
                } else {
                    query.plus_words.insert(query_word.data);
                }
            }
        }
        return query;
    }

    // Existence required
    double ComputeWordInverseDocumentFreq(const string& word) const {
        return log(document_count_ * 1.0 / word_to_document_freqs_.at(word).size());
    }

    vector<Document> FindAllDocuments(const Query& query) const {
        map<int, double> document_to_relevance;
        for (const string& word : query.plus_words) {
            if (word_to_document_freqs_.count(word) == 0) {
                continue;
            }
            const double inverse_document_freq = ComputeWordInverseDocumentFreq(word);
            for (const auto [document_id, term_freq] : word_to_document_freqs_.at(word)) {
                document_to_relevance[document_id] += term_freq * inverse_document_freq;
            }
        }

        for (const string& word : query.minus_words) {
            if (word_to_document_freqs_.count(word) == 0) {
                continue;
            }
            for (const auto [document_id, _] : word_to_document_freqs_.at(word)) {
                document_to_relevance.erase(document_id);
            }
        }

        vector<Document> matched_documents;
        for (const auto [document_id, relevance] : document_to_relevance) {
            matched_documents.push_back({document_id, relevance, document_ratings_.at(document_id)});
        }
        return matched_documents;
    }

    static int ComputeAverageRating(const vector<int>& ratings) {
        int sum_rates = 0;
        int ratings_size = ratings.size();
        for(const auto& rate : ratings){
            sum_rates+=rate;
        }
        return sum_rates/ratings_size;
    }

};

SearchServer CreateSearchServer() {
    SearchServer search_server;
    search_server.SetStopWords(ReadLine());

    const int document_count = ReadLineWithNumber();
    for (int document_id = 0; document_id < document_count; ++document_id) {
        const string document = ReadLine();
        int amount_of_ratings;
        cin >> amount_of_ratings;
        vector<int> ratings(amount_of_ratings, 0);
        for (int& rate : ratings) {
            cin >> rate;
        }
        search_server.AddDocument(document_id, document, ratings);
        ReadLine();
    }

    return search_server;
}

int main() {
    const SearchServer search_server = CreateSearchServer();

    const string query = ReadLine();
    for (auto [document_id, relevance, rating] : search_server.FindTopDocuments(query)) {
        cout << "{ document_id = "s << document_id << ", "s
             << "relevance = "s << relevance << ", rating = "s <<rating<<" }"<< endl;
    }
}