First version of weight calculation, not storing as JSON yet

.gitignore

 # custom
 data/
 .idea/
+.vscode/
 # Byte-compiled / optimized / DLL files
 __pycache__/

requirements.md

+# Požadavky
+- kompilátor podporující minimálně C++ 11
+- vytvořit adresář ```lib``` v kořenovém adresáři projektu pro následující knihovny
+- knihovna pro práci s SQLite:
+  - ke stažení [ZDE](https://github.com/SRombauts/SQLiteCpp/releases)
+  - rozbalit do adresáře ```lib``` a složku přejmenovat na ```SQLiteCpp```
+- knihovna pro práci s JSON soubory:
+  - ke stažení [ZDE](https://github.com/nlohmann/json/releases)
+  - stačí stáhnout hlavičkový soubor ```json.hpp``` a umístit ho do adresáře ```lib```
+- pro sestavení lze využít skript ```src/weight_calculation/build.sh```
+- binárka se poté nachází v ```src/weight_calculation/build/main```, je nutné ji spustit z adresáře ```build```

src/weight_calculation/CMakeLists.txt

+# Minimum CMake version, project name and project version
+cmake_minimum_required(VERSION 3.1)
+project(VectorModel VERSION 1.0)
+
+# C++ 11 compiler is required
+set(CMAKE_CXX_STANDARD 11)
+set(CMAKE_CXX_STANDARD_REQUIRED ON)
+
+# Add subdirectory SQLiteCpp with all necessary files
+add_subdirectory(./../../lib/SQLiteCpp ./SQLite)
+
+add_executable(main main.cpp)
+
+target_link_libraries(main SQLiteCpp)

src/weight_calculation/build.sh

+#!/bin/bash
+
+# Exit on first error
+set -e
+
+# Create dir build
+mkdir -p build
+cd build
+
+cmake -DCMAKE_BUILD_TYPE=Debug ..
+
+# Build
+cmake --build .

src/weight_calculation/main.cpp

+#include <iostream>
+#include <iomanip>
+#include <fstream>
+#include <string>
+
+#include <SQLiteCpp/SQLiteCpp.h>
+#include "../../lib/json.hpp"
+
+using json = nlohmann::json;
+
+bool calculateWeight(SQLite::Database & db, std::ofstream & ostream, const json & maxOccurrences, const std::string & term)
+{
+    uint32_t occurrences;
+    double weight;
+    maxOccurrences[term].get_to(occurrences);
+
+    try
+    {
+        SQLite::Statement query(db, "SELECT * FROM TermDocumentOccurrence JOIN Term ON TermDocumentOccurrence.Term_id = Term.id WHERE Term.value = :term");
+        query.bind(":term", term);
+
+        ostream << term << ":";
+        while(query.executeStep())
+        {
+            weight = query.getColumn("count").getInt() / (occurrences*1.0);
+            ostream << " " << query.getColumn("Document_id") << " " << std::setprecision(20) << weight;
+        }
+        ostream << std::endl;
+    }
+
+    catch(const std::exception& e)
+    {
+        std::cout << "SQLite exception: " << e.what() << std::endl;
+        return false;
+    }
+
+    return true;
+}
+
+bool process(std::ofstream & ostream, const json & maxOccurrences)
+{
+    try
+    {
+        SQLite::Database db("./../../../data/persistance/db", SQLite::OPEN_READWRITE|SQLite::OPEN_CREATE);
+        SQLite::Statement query(db, "SELECT value FROM Term");
+
+        while (query.executeStep())
+            if(!calculateWeight(db, ostream, maxOccurrences, query.getColumn("value")))
+                return false;
+    }
+
+    catch(const std::exception& e)
+    {
+        std::cout << "SQLite exception: " << e.what() << std::endl;
+        return false;
+    }
+
+    return true;
+}
+
+int main (void)
+{
+    std::ifstream istream("./../../../data/persistance/most_frequent_words.json");
+    std::ofstream ostream("./../../../data/persistance/invertedList.json");
+
+    if (istream.fail() || ostream.fail())
+    {
+        std::cout << "Cannot open/find file 'most_frequent_words.json' or cannot create file 'invertedList.json'" << std::endl;
+        return EXIT_FAILURE;
+    }
+
+    json maxOccurrences;
+    istream >> maxOccurrences;
+    istream.close();
+
+    if (istream.fail())
+    {
+        std::cout << "Something went wrong with file 'most_frequent_words.json'" << std::endl;
+        return EXIT_FAILURE;
+    }
+
+    if (!process(ostream, maxOccurrences))
+        return EXIT_FAILURE;
+
+    ostream.close();
+    if (ostream.fail())
+    {
+        std::cout << "Something went wrong during writing to output file 'invertedList.json'" << std::endl;
+        return EXIT_FAILURE;
+    }
+
+    return EXIT_SUCCESS;
+}