Sprint_4

#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 {
    Document() = default;

    Document(int id, double relevance, int rating)
        : id(id)
        , relevance(relevance)
        , rating(rating) {
    }

    int id = 0;
    double relevance = 0.0;
    int rating = 0;
};

template <typename StringContainer>
set<string> MakeUniqueNonEmptyStrings(const StringContainer& strings) {
    set<string> non_empty_strings;
    for (const string& str : strings) {
        if (!str.empty()) {
            non_empty_strings.insert(str);
        }
    }
    return non_empty_strings;
}

enum class DocumentStatus {
    ACTUAL,
    IRRELEVANT,
    BANNED,
    REMOVED,
};
class StopWords {
public:

    StopWords() = default;

    StopWords(const string& text) {

        for (const string& word : SplitIntoWords(text)) {
            stop_words.insert(word);
        }
    }

    template <typename Container>
    StopWords(const Container& container) {

        for (const auto& element : container) {
            if (!element.empty()) {
                stop_words.insert(element);
            }
        }
    }

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

private:
    set<string> stop_words = { ""s };
};

class SearchServer {
public:
    inline static constexpr int INVALID_DOCUMENT_ID = -1;

    template <typename StringContainer>
    explicit SearchServer(const StringContainer& stop_words)
        : stop_words_(MakeUniqueNonEmptyStrings(stop_words)) {
        /*bool check_valid = none_of(stop_words.begin(), stop_words.end(), [](char c) {
            return c >= '\0' && c < ' ';
            });
        if (check_valid) throw invalid_argument("Ban-word contains forbidden characher");*/
        for (const auto& word : stop_words) {
            if (IsValidWord(word)) throw invalid_argument("blabla");
        }
    }

    explicit SearchServer(const string& stop_words_text)
        : SearchServer(
            SplitIntoWords(stop_words_text))  // Invoke delegating constructor from string container
    {
        //if (IsValidWord(stop_words_text)) throw invalid_argument("blabla");
    }

    void AddDocument(int document_id, const string& document, DocumentStatus status,
        const vector<int>& ratings) {
        if (document_id < 0 or documents_.count(document_id) > 0 or IsValidWord(document) == false) {
            throw invalid_argument("invalid action");
        }

        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;
        }
        documents_.emplace(document_id, DocumentData{ ComputeAverageRating(ratings), status });
        count_documents_.push_back(document_id);
    }

    template <typename DocumentPredicate>
    vector<Document> FindTopDocuments(const string& raw_query,
        DocumentPredicate document_predicate) const {

        Query query;
        if (!ParseQuery(raw_query, query)) {
            throw invalid_argument("invalid action");
        }

        auto matched_documents = FindAllDocuments(query, document_predicate);

        sort(matched_documents.begin(), matched_documents.end(),
            [](const Document& lhs, const Document& rhs) {
                if (abs(lhs.relevance - rhs.relevance) < 1e-6) {
                    return lhs.rating > rhs.rating;
                }
                else {
                    return lhs.relevance > rhs.relevance;
                }
            });
        if (matched_documents.size() > MAX_RESULT_DOCUMENT_COUNT) {
            matched_documents.resize(MAX_RESULT_DOCUMENT_COUNT);
        }
        /*if (matched_documents.empty()) {
            return false;
        }*/
        vector<Document> result = matched_documents;
        return result;
    }

    vector<Document> FindTopDocuments(const string& raw_query, DocumentStatus status) const {
        return FindTopDocuments(
            raw_query, [status](int document_id, DocumentStatus document_status, int rating) {
                return document_status == status;
            });
    }

    vector<Document> FindTopDocuments(const string& raw_query) const {
        return FindTopDocuments(raw_query, DocumentStatus::ACTUAL);
    }

    int GetDocumentCount() const {
        return documents_.size();
    }

    int GetDocumentId(int index) const {
        if (index < 0 || index >= static_cast<int>(count_documents_.size())) {
            throw out_of_range("Out of range (0; amount of documents)");
        }
        return count_documents_[index];
    }

    tuple<vector<string>, DocumentStatus> MatchDocument(const string& raw_query, int document_id) const {
        Query query;
        tuple<vector<string>, DocumentStatus> result;
        if (!ParseQuery(raw_query, query)) {
            //return false;
            //return nullopt;
            throw invalid_argument("invalid action");
        }
        vector<string> match_words;

        for (const auto& plus : query.plus_words) {
            if (!word_to_document_freqs_.count(plus) == 0 && word_to_document_freqs_.at(plus).count(document_id)) {
                match_words.push_back(plus);
            }
        }
        for (const auto& minus : query.minus_words) {
            if ((!word_to_document_freqs_.count(minus) == 0 && word_to_document_freqs_.at(minus).count(document_id))) {
                match_words.clear();
                break;
            }
        }

        result = { match_words, documents_.at(document_id).status };
        return result;
    }

    static bool IsValidWord(const string& word) {
        // A valid word must not contain special characters
        return none_of(word.begin(), word.end(), [](char c) {
            return c >= '\0' && c < ' ';
            });
    }

private:
    struct DocumentData {
        int rating;
        DocumentStatus status;
    };

    StopWords stop_words_;
    vector<int> count_documents_;
    map<string, map<int, double>> word_to_document_freqs_;
    map<int, DocumentData> documents_;


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

    static int ComputeAverageRating(const vector<int>& ratings) {
        if (ratings.empty()) {
            return 0;
        }
        int rating_sum = 0;
        for (const int rating : ratings) {
            rating_sum += rating;
        }
        return rating_sum / static_cast<int>(ratings.size());
    }

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

    [[nodiscard]] bool ParseQueryWord(string text, QueryWord& queryWord) const {
        bool is_minus = false;
        // Word shouldn't be empty
        if (text[0] == '-' && static_cast<int>(text.size()) == 1) {
            return false;
        }
        if (text[0] == '-' && text[1] == '-') {
            return false;
        }
        if (text[static_cast<int>(text.size()) - 1] == '-') {
            return false;
        }
        if (text[0] == '-') {
            is_minus = true;
            text = text.substr(1);
            if (!IsValidWord(text)) {
                return  false;
            }
        }
        queryWord = { text, is_minus, stop_words_.IsStopWord(text) };
        return true;
    }

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

    [[nodiscard]] bool ParseQuery(const string& text, Query& query) const {

        auto split_words = SplitIntoWords(text);
        for (const string& word : split_words) {
            if (!IsValidWord(word)) {
                return false;
            }
            QueryWord query_word;
            if (!ParseQueryWord(word, query_word)) {
                return false;
            }
            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 true;
    }

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

    template <typename DocumentPredicate>
    vector<Document> FindAllDocuments(const Query& query,
        DocumentPredicate document_predicate) 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)) {
                const auto& document_data = documents_.at(document_id);
                if (document_predicate(document_id, document_data.status, document_data.rating)) {
                    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, documents_.at(document_id).rating });
        }
        return matched_documents;
    }
};

void PrintDocument(const Document& document) {
    cout << "{ "s
        << "document_id = "s << document.id << ", "s
        << "relevance = "s << document.relevance << ", "s
        << "rating = "s << document.rating << " }"s << endl;
}
int main() {
}