CPP不能用。。。。

cpp/ToolGood.Words.sln

@@ -0,0 +1,31 @@
+
+Microsoft Visual Studio Solution File, Format Version 12.00
+# Visual Studio Version 17
+VisualStudioVersion = 17.11.35312.102
+MinimumVisualStudioVersion = 10.0.40219.1
+Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "ToolGood.Words", "ToolGood.Words\ToolGood.Words.vcxproj", "{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}"
+EndProject
+Global
+	GlobalSection(SolutionConfigurationPlatforms) = preSolution
+		Debug|x64 = Debug|x64
+		Debug|x86 = Debug|x86
+		Release|x64 = Release|x64
+		Release|x86 = Release|x86
+	EndGlobalSection
+	GlobalSection(ProjectConfigurationPlatforms) = postSolution
+		{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}.Debug|x64.ActiveCfg = Debug|x64
+		{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}.Debug|x64.Build.0 = Debug|x64
+		{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}.Debug|x86.ActiveCfg = Debug|Win32
+		{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}.Debug|x86.Build.0 = Debug|Win32
+		{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}.Release|x64.ActiveCfg = Release|x64
+		{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}.Release|x64.Build.0 = Release|x64
+		{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}.Release|x86.ActiveCfg = Release|Win32
+		{AB38D1D8-AAC3-450E-9563-29FB3339ADB3}.Release|x86.Build.0 = Release|Win32
+	EndGlobalSection
+	GlobalSection(SolutionProperties) = preSolution
+		HideSolutionNode = FALSE
+	EndGlobalSection
+	GlobalSection(ExtensibilityGlobals) = postSolution
+		SolutionGuid = {143AC62E-6044-4A30-9652-2DC31C3AEABE}
+	EndGlobalSection
+EndGlobal

cpp/ToolGood.Words/BaseSearchEx.cpp

@@ -0,0 +1,228 @@
+#pragma
+
+#include "IntDictionary.cpp"
+#include <map>
+#include <vector>
+#include <string>
+#include "TrieNode.h"
+using std::map;
+using std::vector;
+using std::string;
+
+class BaseSearchEx
+{
+protected:
+	unsigned short* _dict;
+	int* _first;
+
+	IntDictionary* _nextIndex;
+	int* _end;
+	int* _resultIndex;
+	int* _keywordLengths;
+
+protected:
+	virtual void SetKeywords2(string _keywords[])
+	{
+		TrieNode root;
+		map<int, vector<TrieNode>> allNodeLayers;
+		int kindex = 0;
+
+		for (size_t i = 0; i < _keywords->size(); i++) /// _keywords->size() 这个值有问题
+		{
+			string p = _keywords[i];
+
+			TrieNode nd = root;
+			for (size_t j = 0; j < p.length(); j++)   //  这个 p.length() 对中文有问题
+			{
+				nd = nd.Add((char)p.c_str()[j]);   // 返回是byte 不是char
+				if (nd.Layer == 0) {
+					nd.Layer = j + 1;
+					auto find = allNodeLayers.find(nd.Layer);
+					if (find == allNodeLayers.end())
+					{
+						vector<TrieNode> trieNodes;
+						trieNodes.push_back(nd);
+						allNodeLayers[nd.Layer] = trieNodes;
+					}
+					else
+					{
+						find->second.push_back(nd);
+					}
+				}
+			}
+		}
+		vector<TrieNode> allNode;
+		allNode.push_back(root);
+		for (size_t i = 0; i < allNodeLayers.size(); i++)
+		{
+			auto item = allNodeLayers.at(i);
+			for (size_t j = 0; j < item.size(); j++)
+			{
+				allNode.push_back(item[j]);
+			}
+		}
+		allNodeLayers.clear();
+
+		for (size_t i = 0; i < allNode.size(); i++)
+		{
+			TrieNode nd = allNode[i];
+			nd.Index = i;
+			TrieNode r = *(nd.Parent->Failure);
+			char c = nd.Char;
+			while (&r != NULL && (r.m_values.size() == 0 || r.m_values.find(c) == r.m_values.end())) r = *r.Failure;
+			if (&r == NULL)
+			{
+				nd.Failure = &(root);
+			}
+			else
+			{
+				nd.Failure = &(r.m_values[c]);
+				if (nd.Failure->Results.size() > 0) {
+					for (size_t i = 0; i < nd.Failure->Results.size(); i++)
+					{
+						nd.SetResults(nd.Failure->Results[i]);
+					}
+				}
+			}
+		}
+		root.Failure = &(root);
+
+		string stringBuilder = "";
+		for (int i = 1; i < allNode.size(); i++) {
+			stringBuilder += (allNode[i].Char);
+		}
+		int length = CreateDict(stringBuilder);
+		stringBuilder.clear();
+
+		int first[0x10000];
+		if (allNode[0].m_values.size() > 0) {
+			for (size_t i = 0; i < allNode[0].m_values.size(); i++)
+			{
+				auto it = allNode[0].m_values.at(i);
+				char key = (char)_dict[it.Char];
+				first[key] = it.Index;
+			}
+		}
+		_first = first;
+
+		vector<int> resultIndex2;
+		vector<bool> isEndStart;
+		int len = allNode.size();
+		IntDictionary* nextIndex2 = new IntDictionary[len];
+
+		for (int i = allNode.size() - 1; i >= 0; i--) {
+			map<unsigned short, int> dict;
+			vector<int> result;
+			TrieNode oldNode = allNode[i];
+
+			if (oldNode.m_values.size() > 0) {
+				for (size_t i = 0; i < oldNode.m_values.size(); i++)
+				{
+					char key = (char)oldNode.m_values.at(i).Char;
+					int index = oldNode.m_values.at(i).Index;
+					dict[key] = index;
+				}
+			}
+			if (oldNode.Results.size() > 0) {
+				for (size_t i = 0; i < oldNode.Results.size(); i++)
+				{
+					result.push_back(oldNode.Results[i]);
+				}
+			}
+
+			oldNode = *oldNode.Failure;
+			while (oldNode.Index != root.Index) {
+				if (oldNode.m_values.size() > 0) {
+					for (size_t i = 0; i < oldNode.m_values.size(); i++)
+					{
+						char key = (char)oldNode.m_values.at(i).Char;
+						int index = oldNode.m_values.at(i).Index;
+						if (dict.find(key) == dict.end())
+						{
+							dict[key] = index;
+						}
+					}
+				}
+				if (oldNode.Results.size() > 0) {
+					for (size_t i = 0; i < oldNode.Results.size(); i++)
+					{
+						int idx = oldNode.Results[i];
+						bool find = false;
+						for (size_t j = 0; j < result.size(); j++)
+						{
+							if (result[i] == idx)
+							{
+								find = true;
+								break;
+							}
+						}
+						if (find == false)
+						{
+							result.push_back(oldNode.Results[i]);
+						}
+					}
+				}
+				oldNode = *(oldNode.Failure);
+			}
+			nextIndex2[i] = *(new IntDictionary(dict));
+
+			if (result.size() > 0) {
+				for (int j = result.size() - 1; j >= 0; j--) {
+					resultIndex2.push_back(result[j]);
+					isEndStart.push_back(false);
+				}
+				isEndStart[isEndStart.size() - 1] = true;
+			}
+			else {
+				resultIndex2.push_back(-1);
+				isEndStart.push_back(true);
+			}
+		}
+
+		allNode.clear();
+		_nextIndex = nextIndex2;
+
+		vector<int> resultIndex;
+		vector<int> end;
+
+		for (int i = isEndStart.size() - 1; i >= 0; i--) {
+			if (isEndStart[i]) {
+				end.push_back(resultIndex.size());
+			}
+			if (resultIndex2[i] > -1) {
+				resultIndex.push_back(resultIndex2[i]);
+			}
+		}
+		end.push_back(resultIndex.size());
+
+		_resultIndex = new int[resultIndex.size()];
+		for (size_t i = 0; i < resultIndex.size(); i++)
+		{
+			_resultIndex[i] = resultIndex[i];
+		}
+		_end = new int[end.size()];
+		for (size_t i = 0; i < end.size(); i++)
+		{
+			_end[i] = end[i];
+		}
+	}
+
+
+private:
+	int CreateDict(string keywords) {
+		_dict = new unsigned short[0x10000];
+		map<char, unsigned int> dictionary;
+		int index = 1;
+		for (size_t i = 0; i < keywords.size(); i++)
+		{
+			char item = keywords.at(i);
+			if (_dict[item] == 0)
+			{
+				_dict[item] = index;
+				index++;
+			}
+		}
+		return index - 1;
+	}
+
+};

cpp/ToolGood.Words/IntDictionary.cpp

@@ -0,0 +1,100 @@
+#pragma once
+
+#include <map>
+#include <vector>
+#include <string>
+#include <algorithm>
+using std::map;
+using std::vector;
+using std::string;
+using namespace std;
+
+
+typedef pair<unsigned short, int> PAIR;
+//int cmp2(const PAIR& x, const PAIR& y) {
+//	return x.first < y.first;
+//}
+
+class IntDictionary
+{
+public:
+	unsigned short* _keys;
+	int* _values;
+	int last;
+
+public:
+
+	IntDictionary()
+	{
+
+	}
+	IntDictionary(unsigned short* keys, int* values, int len)
+	{
+		_keys = keys;
+		_values = values;
+		last = len - 1;
+	}
+	IntDictionary(map<unsigned short, int> dict)
+	{
+		int len = dict.size();
+		_keys = new unsigned short[len];
+		_values = new int[len];
+		last = len - 1;
+
+		vector<PAIR> vec(dict.begin(), dict.end());
+		sort(vec.begin(), vec.end());
+
+		for (size_t i = 0; i < vec.size(); i++)
+		{
+			_keys[i] = vec[i].first;
+			_values[i] = vec[i].second;
+		}		
+		last = len - 1;
+	}
+
+
+	bool TryGetValue(unsigned short key, int& value)
+	{
+		if (last == -1) {
+			value = 0;
+			return false;
+		}
+		if (_keys[0] == key) {
+			value = _values[0];
+			return true;
+		}
+		else if (last == 0 || _keys[0] > key) {
+			value = 0;
+			return false;
+		}
+
+		if (_keys[last] == key) {
+			value = _values[last];
+			return true;
+		}
+		else if (_keys[last] < key) {
+			value = 0;
+			return false;
+		}
+
+		int left = 1;
+		int right = last - 1;
+		while (left <= right) {
+			int mid = (left + right) >> 1;
+			int d = _keys[mid] - key;
+
+			if (d == 0) {
+				value = _values[mid];
+				return true;
+			}
+			else if (d > 0) {
+				right = mid - 1;
+			}
+			else {
+				left = mid + 1;
+			}
+		}
+		value = 0;
+		return false;
+	}
+};

cpp/ToolGood.Words/ToolGood.Words.cpp

@@ -0,0 +1,30 @@
+// ToolGood.Words.cpp : 此文件包含 "main" 函数。程序执行将在此处开始并结束。
+//
+
+#include <iostream>
+#include "WordsSearchEx.cpp"
+
+int main()
+{
+	WordsSearchEx search;
+	auto ks = new string[2];
+	ks[0] = "中国";
+	ks[1] = "国人";
+	ks[2] = "zg人";
+
+	search.SetKeywords(ks);
+	auto r = search.FindFirst("我是中国人");
+
+	std::cout << "Hello World!\n";
+}
+
+// 运行程序: Ctrl + F5 或调试 >“开始执行(不调试)”菜单
+// 调试程序: F5 或调试 >“开始调试”菜单
+
+// 入门使用技巧: 
+//   1. 使用解决方案资源管理器窗口添加/管理文件
+//   2. 使用团队资源管理器窗口连接到源代码管理
+//   3. 使用输出窗口查看生成输出和其他消息
+//   4. 使用错误列表窗口查看错误
+//   5. 转到“项目”>“添加新项”以创建新的代码文件，或转到“项目”>“添加现有项”以将现有代码文件添加到项目
+//   6. 将来，若要再次打开此项目，请转到“文件”>“打开”>“项目”并选择 .sln 文件