NER.h

Go to the documentation of this file.

#pragma once
 
#include <algorithm>
#include <tuple>
 
namespace krotos
{
    class NER
    {
      public:
        std::unordered_map<String, StringArray> findEntity(String text,
                                                           const std::unordered_map<String, StringArray>& dictionary,
                                                           float similarityThreshold = 0.9f);
 
      private:
        StringArray ngrams(const StringArray& tokens, int n = 1);
 
        int levenshteinDistance(const String& str1, const String& str2);
 
        float stringSimilarity(const String& str1, const String& str2);
 
        std::tuple<float, String, String> getFuzzySimilarity(String text,
                                                             const std::unordered_map<String, StringArray>& dictionary,
                                                             float similarityThreshold);
 
        // Struct to hold named entity info
        struct Entity
        {
            String name;
            String category;
            float score;
            int startIndex;
            int endIndex;
            Entity(String name = String(), String category = String(), float score = 0.0f, int startIndex = 0,
                   int endIndex = 0)
                : name(name), category(category), score(score), startIndex(startIndex), endIndex(endIndex)
            {
            }
        };
 
        std::vector<Entity> removeOverlapping(std::vector<Entity> entities);
    };
 
} // namespace krotos