SearchEngine_demo

最新推荐文章于 2024-07-18 21:35:15 发布

被狗咬的玄离

最新推荐文章于 2024-07-18 21:35:15 发布

阅读量821

点赞数 30

文章标签： go 搜索引擎 redis 爬虫

本文链接：https://blog.csdn.net/m0_70699638/article/details/139809221

版权

前言
本搜索引擎使用go语言实现，采用redis作为存储数据库，使用到colly爬虫框架，gin-gonic前端框架。
实现了包括页面抓取，建立倒排索引等功能。
注：该项目只是一个大作业，存在很多不足，仅供参考！！！！！！！

1、项目架构图

在这里插入图片描述

2、运行效果

demo

3、项目目录结构

MySearchEngine

collySpider
- collySpider.go
dict
- stop_words.utf8
models
- page.go
templates
- search.html
tools
- checkUrl.go
- createStopWords.go
- getUnVistedUrl.go
- getUrlMD5.go
- init.go
- initDB.go
- invert_index.go
- savePages.go
- saveUnVisitedUrl.go
- search.go
main.go

4、具体代码实现

该爬虫爬取域名主要为www.chinanews.com.cn、www.news.cn，可以在init.go内进行修改。

4.1 collySpider

collySpider.go

package collySpider

import (
	"log"
	"os"
	"pkg/models"
	"pkg/tools"
	"regexp"
	"strings"
	"sync"

	"github.com/gocolly/colly"
)

var mu sync.Mutex // 互斥锁

func Craw() {
	cl := colly.NewCollector(
		colly.MaxDepth(6),
		colly.AllowedDomains("www.chinanews.com.cn", "www.news.cn"),
	)

	cl.OnRequest(func(r *colly.Request) {
		r.Headers.Set("User-Agent", "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36")
	})

	cl.OnError(func(r *colly.Response, err error) {
		errorLog := log.New(os.Stderr, "Error: ", log.LstdFlags)
		errorLog.Println(err)
	})

	var lake models.Page

	cl.OnHTML("*", func(el *colly.HTMLElement) {
		href := el.Attr("href")
		href = strings.TrimSpace(href) // Remove leading and trailing whitespaces
		if href != "" {
			// fmt.Println(href)
			tools.Saveunvisitedurl(href)
		}
	})

	cl.OnHTML("body", func(el *colly.HTMLElement) {
		title := strings.TrimSpace(el.ChildText("div[class='head-line clearfix'] > h1 > span[class='title']"))
		if title == "" {
			title = strings.TrimSpace(el.ChildText("h1[class='content_left_title']"))
			if title == "" {
				title = strings.TrimSpace(el.ChildText("div[class='content_title']"))
				if title == "" {
					title = strings.TrimSpace(el.ChildText("h1[class='page_title']"))
				}
			}
		}
		lake.Title = title
		text := el.Text
		el.ForEach("script", func(_ int, s *colly.HTMLElement) {
			text = strings.ReplaceAll(text, s.Text, "")
		})
		el.ForEach("noscript", func(_ int, s *colly.HTMLElement) {
			text = strings.ReplaceAll(text, s.Text, "")
		})
		el.ForEach("style", func(_ int, h *colly.HTMLElement) {
			text = strings.ReplaceAll(text, h.Text, "")
		})
		text = strings.ReplaceAll(text, " ", "")
		text = strings.ReplaceAll(text, "\n", "")
		text = strings.ReplaceAll(text, "\t", "")
		pattern := `[0-9A-Za-zΑ-Ωα-ω!"#$%&'()*+,-./:;<=>?@[\\\]^_` + "`" + `{|}~\p{P}\p{S}\p{M}\p{C}\p{Z}]`
		re := regexp.MustCompile(pattern)
		text = re.ReplaceAllString(text, "")
		lake.Url = el.Request.URL.String()
		if text == "" {
			lake.Text = lake.Title
		} else {
			lake.Text = text
		}
		mu.Lock() // 加锁
		tools.Savepages(lake.Url, lake.Title)
		ids, _ := tools.GetDocID(lake.Url)
		tools.AddDocument(ids, lake.Text)
		mu.Unlock() // 解锁
	})

	// // 创建定时器,在一分钟后执行后面的代码
	// timer := time.NewTimer(1 * time.Minute)
	// go func() {
	// 	<-timer.C
	// 	const concurrentRequests = 64

	// 	ch := make(chan string, concurrentRequests)
	// 	wg := sync.WaitGroup{}
	// 	for i := 0; i < concurrentRequests; i++ {
	// 		wg.Add(1)
	// 		go func() {
	// 			defer wg.Done()
	// 			for url := range ch {
	// 				err := cl.Visit(url)
	// 				if err != nil {
	// 					errorLog := log.New(os.Stderr, "Error: ", log.LstdFlags)
	// 					errorLog.Println(err)
	// 				}
	// 			}
	// 		}()
	// 	}

	// 	for {
	// 		url, err := tools.Getunvistedurl()
	// 		if err != nil {
	// 			if err.Error() == "redis: nil" {
	// 				time.Sleep(20 * time.Second) // 等待1秒后重试
	// 				continue
	// 			}
	// 			log.Println(err)
	// 			return
	// 		}
	// 		ch <- url
	// 	}
	// }()

	// 在定时器启动前执行前一分钟的代码
	raw_url, _ := tools.Getunvistedurl()
	err := cl.Visit(raw_url)
	if err != nil {
		errorLog := log.New(os.Stderr, "Error: ", log.LstdFlags)
		errorLog.Println(err)
	}
}

注：多线程似乎还存在部分问题，docID存在锁不住的问题，如需使用请检查改进后使用。

4.2 dict

该目录下存放停词库文件,请自行查找加入目录.

4.3 models

page.go

package models

type Page struct {
	ID    int    `gorm:"primaryKey"`
	Url   string `gorm:"default:null"`
	Title string `gorm:"default:null"`
	Text  string `gorm:"default:null"`
}

4.4 templates

search.html

<!DOCTYPE html>
<html>

<head>
    <title>🍙搜索</title>
    <style>
        html, body {
            height: 100%;
            margin: 0;
            padding: 0;
            font-family: Arial, sans-serif;
            overflow-x: hidden; /* Prevent horizontal scrollbar */
        }

        body {
            overflow-y: auto; /* Allow vertical scrolling */
            scroll-behavior: smooth; /* Enable smooth scrolling */
        }

        #particles-js {
            position: fixed;
            width: 100%;
            height: 100%;
            background-color: #2b2e4a;
            z-index: -1;
        }

        #app {
            position: relative;
            z-index: 1;
            max-width: 800px;
            margin: 50px auto;
            padding: 20px;
            background: rgba(255, 255, 255, 0.9);
            box-shadow: 0 0 10px rgba(0, 0, 0, 0.1);
            border-radius: 10px;
        }

        h1 {
            text-align: center;
            color: #4CAF50;
            font-size: 2.5em;
            margin-bottom: 20px;
        }

        input[type="text"] {
            width: calc(100% - 22px);
            padding: 10px;
            margin-bottom: 20px;
            border: 1px solid #ddd;
            border-radius: 4px;
            font-size: 1em;
        }

        button {
            display: block;
            width: 100%;
            padding: 10px;
            background-color: #4CAF50;
            color: white;
            border: none;
            border-radius: 4px;
            cursor: pointer;
            font-size: 1em;
        }

        button:hover {
            background-color: #45a049;
        }

        ul {
            list-style-type: none;
            padding: 0;
        }

        li {
            background: #f9f9f9;
            margin: 10px 0;
            padding: 20px;
            border: 1px solid #ddd;
            border-radius: 4px;
        }

        h3 {
            margin: 0 0 10px;
        }

        a {
            color: #4CAF50;
            text-decoration: none;
        }

        a:hover {
            text-decoration: underline;
        }
    </style>
</head>

<body>
    <div id="particles-js"></div>
    <div id="app">
        <h1>🍙搜索</h1>
        <input type="text" id="query" placeholder="原神启动！！！！！！！！！！" onkeyup="checkEnter(event)">
        <button id="searchBtn" onclick="search()">启动</button>
        <ul id="results">
            <!-- Search results will be inserted here -->
        </ul>
    </div>

    <script src="https://cdn.jsdelivr.net/particles.js/2.0.0/particles.min.js"></script>
    <script>
        document.addEventListener("DOMContentLoaded", function() {
            particlesJS("particles-js", {
                particles: {
                    number: {
                        value: 80,
                        density: {
                            enable: true,
                            value_area: 800
                        }
                    },
                    color: {
                        value: "#ffffff"
                    },
                    shape: {
                        type: "circle",
                        stroke: {
                            width: 0,
                            color: "#000000"
                        },
                        polygon: {
                            nb_sides: 5
                        }
                    },
                    opacity: {
                        value: 0.5,
                        random: false,
                        anim: {
                            enable: false,
                            speed: 1,
                            opacity_min: 0.1,
                            sync: false
                        }
                    },
                    size: {
                        value: 3,
                        random: true,
                        anim: {
                            enable: false,
                            speed: 40,
                            size_min: 0.1,
                            sync: false
                        }
                    },
                    line_linked: {
                        enable: true,
                        distance: 150,
                        color: "#ffffff",
                        opacity: 0.4,
                        width: 1
                    },
                    move: {
                        enable: true,
                        speed: 6,
                        direction: "none",
                        random: false,
                        straight: false,
                        out_mode: "out",
                        bounce: false,
                        attract: {
                            enable: false,
                            rotateX: 600,
                            rotateY: 1200
                        }
                    }
                },
                interactivity: {
                    detect_on: "canvas",
                    events: {
                        onhover: {
                            enable: true,
                            mode: "repulse" // Interaction mode on hover
                        },
                        onclick: {
                            enable: true,
                            mode: "push"
                        },
                        resize: true
                    },
                    modes: {
                        grab: {
                            distance: 200,
                            line_linked: {
                                opacity: 1
                            }
                        },
                        bubble: {
                            distance: 400,
                            size: 40,
                            duration: 2,
                            opacity: 8,
                            speed: 3
                        },
                        repulse: {
                            distance: 100, // Distance of repulsion effect
                            duration: 0.4
                        },
                        push: {
                            particles_nb: 4
                        },
                        remove: {
                            particles_nb: 2
                        }
                    }
                },
                retina_detect: true
            });
        });

        function checkEnter(event) {
            if (event.key === 'Enter') {
                document.getElementById('searchBtn').click();
            }
        }

        function search() {
            const query = document.getElementById('query').value;
            fetch('/search', {
                method: 'POST',
                headers: {
                    'Content-Type': 'application/json'
                },
                body: JSON.stringify({ query })
            })
            .then(response => response.json())
            .then(data => {
                const resultsElement = document.getElementById('results');
                resultsElement.innerHTML = data.map(result => `
                    <li>
                        <h3>${result.Title}</h3>
                        <p>Relevance Score: ${result.Score}</p>
                        <p>URL: <a href="${result.URL}" target="_blank">${result.URL}</a></p>
                    </li>
                `).join('');

                resultsElement.scrollIntoView({ behavior: 'smooth' });
            })
            .catch(error => console.error('Error:', error));
        }
    </script>
</body>

</html>

4.5 tools

checkUrl.go

package tools

import (
	"fmt"
	"net/url"
)

// IsValidURL 检查URL是否有效
func IsValidURL(rawURL string) bool {
	u, err := url.ParseRequestURI(rawURL)
	if err != nil {
		return false
	}
	if u.Scheme != "http" && u.Scheme != "https" {
		return false
	}
	return true
}

// Checkvistedurl 检查URL是否已访问过
func Checkvistedurl(urlMD5 string) (flag bool) {
	flag = false

	visitedQueueName := "visited_urls_queue"

	// 从visited队列中检索一系列元素
	elements, err := rdb.LRange(ctx, visitedQueueName, 0, -1).Result()
	if err != nil {
		fmt.Println("Error while retrieving elements from Redis:", err)
		return
	}

	// 检查URL是否存在于检索到的元素中
	for _, element := range elements {
		if Getmd5(element) == urlMD5 {
			// fmt.Println("URL已在visited队列中.")
			flag = true
			break
		}
	}

	// fmt.Println("URL不在visited队列中.")
	return flag
}

createStopWords.go

package tools

import (
	"bufio"
	"os"
	"strings"
)

// Createstopwords 从文件中创建停用词列表
func Createstopwords() []string {
	file, _ := os.Open("/home/nanzi/goCode/src/MySearchEngine/dict/stop_words.utf8")

	defer file.Close()

	scanner := bufio.NewScanner(file)
	var stopWords []string
	for scanner.Scan() {
		line := strings.TrimSpace(scanner.Text())
		stopWords = append(stopWords, line)
	}

	return stopWords
}

getUnVistedUrl.go

package tools

import (
	"fmt"
)

// Getunvistedurl 从未访问URL队列中获取一个URL
func Getunvistedurl() (string, error) {
	// Retrieve a URL from the "unvisited_urls_queue"
	url, err := rdb.LPop(ctx, "unvisited_urls_queue").Result()
	if err != nil {
		return "", fmt.Errorf("failed to get URL from the unvisited URL queue: %w", err)
	}

	// Check if the URL has been visited
	if !Checkvistedurl(Getmd5(url)) {
		// Add the URL to the "visited_urls_queue"
		err = rdb.LPush(ctx, "visited_urls_queue", url).Err()
		if err != nil {
			return "", fmt.Errorf("failed to add URL to the visited URL queue: %w", err)
		}
		return url, nil
	}

	// 如果该URL已被访问过，则递归调用该函数获取新的URL
	return Getunvistedurl()
}

// Getunvistedurlsize 获取未访问URL队列的大小
func Getunvistedurlsize() int {
	queueLen, _ := rdb.LLen(ctx, "unvisited_urls_queue").Result()
	return int(queueLen)
}

getUrlMD5.go

package tools

import (
	"crypto/md5"
	"fmt"
	"io"
)

// Getmd5 计算URL的MD5哈希值
func Getmd5(url string) (urlMD5 string) {
	hasher := md5.New()
	io.WriteString(hasher, url)
	result := fmt.Sprintf("%x", hasher.Sum(nil))

	// fmt.Println("URL的MD5哈希值:", result)
	return result
}

init.go

package tools

func Init() {
	url_list := []string{"https://www.chinanews.com.cn/", "http://www.news.cn/"} 

	unvisitedQueueName := "unvisited_urls_queue"
	_ = rdb.LPush(ctx, unvisitedQueueName, url_list).Err()
}

initDB.go

package tools

import (
	"context"

	"github.com/go-redis/redis/v8"
)

var rdb *redis.Client

var ctx = context.Background()

func init() {
	rdb = redis.NewClient(&redis.Options{
		Addr:     "localhost:6379",
		Password: "", // no password set
		DB:       1,  // use default DB
	})
}

invert_index.go

package tools

import (
	"strings"
	"sync"

	"github.com/yanyiwu/gojieba"
)

func in(s string, list []string) bool {
	for _, e := range list {
		if s == e {
			return true
		}
	}
	return false

}

var mu sync.Mutex // 互斥锁

// InvertedIndex 用于存储单词的倒排索引
var InvertedIndex = make(map[string][]string)

var jiebaInstance *gojieba.Jieba
var jiebaOnce sync.Once

func getJiebaInstance() *gojieba.Jieba {
	jiebaOnce.Do(func() {
		jiebaInstance = gojieba.NewJieba()
	})
	return jiebaInstance
}

// AddDocument 函数用于向索引中添加文档
func AddDocument(id_cnt string, content string) {
	x := getJiebaInstance()

	// 将内容拆分为单词并转换为小写
	words := x.CutForSearch(strings.ToLower(content), true)
	stop_words := Createstopwords()
	// 更新每个单词的倒排索引
	for _, word := range words {
		if in(word, stop_words) {
			continue
		}
		mu.Lock() // 加锁
		docID := id_cnt
		err := rdb.RPush(ctx, word, docID).Err()
		mu.Unlock() // 解锁
		if err != nil {
			panic(err)
		}
	}
}

savePages.go

package tools

import (
	"fmt"
	"net/url"
)

// IsValidURL 检查URL是否有效
func IsValidURL(rawURL string) bool {
	u, err := url.ParseRequestURI(rawURL)
	if err != nil {
		return false
	}
	if u.Scheme != "http" && u.Scheme != "https" {
		return false
	}
	return true
}

// Checkvistedurl 检查URL是否已访问过
func Checkvistedurl(urlMD5 string) (flag bool) {
	flag = false

	visitedQueueName := "visited_urls_queue"

	// 从visited队列中检索一系列元素
	elements, err := rdb.LRange(ctx, visitedQueueName, 0, -1).Result()
	if err != nil {
		fmt.Println("Error while retrieving elements from Redis:", err)
		return
	}

	// 检查URL是否存在于检索到的元素中
	for _, element := range elements {
		if Getmd5(element) == urlMD5 {
			// fmt.Println("URL已在visited队列中.")
			flag = true
			break
		}
	}

	// fmt.Println("URL不在visited队列中.")
	return flag
}

saveUnVisitedUrl.go

package tools

import (
	"fmt"
)

// Saveunvisitedurl 保存未访问的URL到Redis队列
func Saveunvisitedurl(url string) {
	urlMD5 := Getmd5(url)
	unvisitedQueueName := "unvisited_urls_queue"

	if IsValidURL(url) {
		if !Checkvistedurl(urlMD5) {
			err := rdb.LPush(ctx, unvisitedQueueName, url).Err()
			if err != nil {
				fmt.Println("Failed to add URL to Redis queue:", err)
				return
			}
		}
	}
	// fmt.Println("URL已添加到Redis队列:", urlMD5)
}

search.go

package tools

import (
	"math"
	"sort"
	"strings"

	"github.com/yanyiwu/gojieba"
)

type Document struct {
	ID    string
	URL   string
	Title string
	Score float64
}

func BestMatchSearch(query string) []Document {
	x := gojieba.NewJieba()

	// 将内容拆分为单词并转换为小写
	words := x.CutForSearch(strings.ToLower(query), true)

	// 获取每个单词对应的文档列表及其位置信息
	docPositions := make(map[string]map[string][]int)
	for _, word := range words {
		docIDs, err := rdb.LRange(ctx, word, 0, -1).Result()
		if err != nil {
			panic(err)
		}
		for _, docID := range docIDs {
			if _, ok := docPositions[word]; !ok {
				docPositions[word] = make(map[string][]int)
			}
			docPositions[word][docID] = append(docPositions[word][docID], len(docPositions[word][docID]))
		}
	}

	// 计算每个单词的IDF值和文档长度总和
	idfValues := make(map[string]float64)
	totalDocuments, err := rdb.DBSize(ctx).Result()
	if err != nil {
		panic(err)
	}
	totalDocLength := 0
	for _, word := range words {
		docIDs, err := rdb.LRange(ctx, word, 0, -1).Result()
		if err != nil {
			panic(err)
		}
		docCount := len(docIDs)
		idf := math.Log(float64(totalDocuments+1) / (1 + float64(docCount)))
		idfValues[word] = idf
		for _, docID := range docIDs {
			totalDocLength += len(docPositions[word][docID])
		}
	}
	avgdl := float64(totalDocLength) / float64(totalDocuments)

	// 计算每个文档的相关度
	scores := make(map[string]Document)
	for _, word := range words {
		docIDs, err := rdb.LRange(ctx, word, 0, -1).Result()
		if err != nil {
			panic(err)
		}
		for _, docID := range docIDs {
			positions := docPositions[word][docID]
			bm25Score := calculateBM25Score(word, positions, avgdl)
			tfidfScore := bm25Score * idfValues[word]
			if _, ok := scores[docID]; !ok {
				url, title := getURLAndTitle(docID)
				if !isExcludedURL(url) && title != "" { // 检查标题是否为空
					doc := Document{
						ID:    docID,
						URL:   url,
						Title: title,
						Score: 0,
					}
					scores[docID] = doc
				}
			}
			doc := scores[docID]
			doc.Score += tfidfScore
			scores[docID] = doc
		}
	}

	// 根据相关度排序结果
	var sortedResults []Document
	for _, doc := range scores {
		sortedResults = append(sortedResults, doc)
	}

	// 根据相关度降序排序
	sort.Slice(sortedResults, func(i, j int) bool {
		return sortedResults[i].Score > sortedResults[j].Score
	})

	// 过滤没有内容的文档
	var filteredResults []Document
	for _, doc := range sortedResults {
		if doc.Title != "" && !containsDocument(filteredResults, doc) { // 检查文档是否已经存在于切片中
			filteredResults = append(filteredResults, doc)
		}
	}

	// 返回前30个有内容的结果
	return filteredResults[:min(30, len(filteredResults))]
}

func calculateBM25Score(_ string, positions []int, avgdl float64) float64 {
	// BM25算法参数
	k1 := 1.25
	b := 0.75

	// 计算文档长度
	docLen := float64(len(positions))

	// 计算词频
	frequency := float64(len(positions))

	// 计算BM25分数
	score := math.Log(2) * (frequency * (k1 + 1)) / (frequency + k1*(1-b+b*docLen/avgdl))

	return score
}

func getURLAndTitle(docID string) (string, string) {
	values, err := rdb.LRange(ctx, docID, 0, 1).Result()
	if err != nil {
		panic(err)
	}

	if len(values) < 2 {
		return "", ""
	}

	return values[0], values[1]
}

func min(a, b int) int {
	if a < b {
		return a
	}
	return b
}

func isExcludedURL(url string) bool {
	return url == "http://www.news.cn/" || url == "https://www.news.cn/"
}

func containsDocument(docs []Document, doc Document) bool {
	for _, d := range docs {
		if d.ID == doc.ID {
			return true
		}
	}
	return false
}

4.6 主函数

main.go

注：
开始运行时将注释内容解开，等页面抓取足够后再注释起来.

package main

import (
	"net/http"
	"pkg/tools"

	"github.com/gin-gonic/gin"
)

func main() {
	//初始化
	// tools.Init()

	// // Get the current time
	// startTime := time.Now()

	// for tools.Getunvistedurlsize() > 0 {
	// 	collySpider.Craw()

	// 	// Check if 4 hours have elapsed
	// 	if time.Since(startTime) >= 2*time.Hour {
	// 		break
	// 	}
	// }

	// 初始化一个新的 Gin 路由器实例
	router := gin.Default()

	// 从 "templates" 目录加载 HTML 模板
	router.LoadHTMLGlob("templates/*")

	// 定义一个 GET 路由来提供搜索页面
	router.GET("/", func(c *gin.Context) {
		// 当访问根 URL 时，渲染 "search.html" 模板
		c.HTML(http.StatusOK, "search.html", nil)
	})

	// 定义一个 POST 路由来处理搜索请求
	router.POST("/search", func(c *gin.Context) {
		// 定义一个结构体来绑定 JSON 请求负载
		var json struct {
			Query string `json:"query"`
		}

		// 绑定 JSON 负载到结构体并检查错误
		if err := c.ShouldBindJSON(&json); err != nil {
			// 如果有错误，响应一个 Bad Request 状态和错误信息
			c.JSON(http.StatusBadRequest, gin.H{"error": err.Error()})
			return
		}

		// 使用请求负载中的查询执行搜索
		results := tools.BestMatchSearch(json.Query)

		// 以 JSON 格式响应搜索结果
		c.JSON(http.StatusOK, results)
	})

	// 在端口 50086 上启动服务器
	router.Run(":50086")
}