• Django和elasticsearch搜索引擎网站后端功能实现


    一、输入框智能提示(es提供了接口 )
    修改type
    需要在mapping中设置一个字段 suggest:{“type”:“completion”}
    所以要修改我们定义的type:
    在type中新增一个字段:suggest,由于es-dsl源码有一些问题,所以这样定义是会报错的,要自己定义一个CustomAnalyzer,再声明一个自定义的对象,ik_analyzer,再把对象赋给type中的suggest :

    ...
    from elasticsearch_dsl.analysis import CustomAnalyzer as _CustomAnalyzer

    class CustomAnalyzer(_CustomAnalyzer):

    def get_analysis_definition(self):
    # 这里什么都不做,只是为了避免报错的问题
    return ()


    # 声明一个自定义的对象,传递ik_max_word并且做大小写转换
    ik_analyzer = CustomAnalyzer('ik_max_word', filter=['lowercase'])


    class DuowanType(DocType):
    ...
    # 定义suggest是为了完成自动补全功能。
    # 由于es-dsl源码有一些问题,所以这样定义是会报错的,要自己定义一个CustomAnalyzer
    suggest = Completion(analyzer=ik_analyzer)

    生成suggest值
    在save_to_es里面生成搜索建议
    要通过生成suggest的接口来生成自己的结构。
    在items类中定义一个全局函数gen_suggests,传递index和info_tuple用于weight信息,新建一个set用于去重,一个suggest数组用于保存返回的内容。遍历info_tuple,如果text字符串不为空,则调用es的analyze接口来分析字符串,再整理好需要返回的结构

    def gen_suggests(index, info_tuple): # 用tuple 就可以传递多个weight信息 并且还可以按顺序
    # 根据字符串生成搜索建议数组
    uesd_words = set() # 用于去重
    suggests = [] # 用于返回
    for text, weight in info_tuple:
    if text: # 排除空字符串
    # 调用es的analyze接口来分析字符串
    words = es.indices.analyze(index=index, analyzer='ik_max_word', params={'filter': ['lowercase']}, body=text)
    anylyzed_words = set(r["token"] for r in words["tokens"] if len(r["token"]) > 1) # 用来过滤单个字
    new_words = anylyzed_words - uesd_words # 去重
    else:
    new_words = set()
    if new_words:
    suggests.append({'input': list(new_words), 'weight': weight})
    return suggests

    然后在save_to_es中调用这个函数:

    info_tuple = ((duowan.title, 10), (duowan.author, 7))
    duowan.suggest = (gen_suggests(DuowanType._doc_type.index, info_tuple))

    搭建django搜索网站
    创建新的虚拟环境
    进入虚拟环境并安装django包 pip install -i https://pypi.douban.com/simple/ django
    然后用pycharm新建一个django的项目,直接运行,可以在log中看到服务器地址。
    再新建一个static目录,把css,html,js文件粘贴进去,把html文件粘贴到templates目录下。
    在urls文件中新增一个url

    from django.contrib import admin
    from django.urls import path
    from django.conf.urls import url
    from django.views.generic import TemplateView

    urlpatterns = [
    path('admin/', admin.site.urls),
    url(r'^$', TemplateView.as_view(template_name='index.html'), name='index'),
    ]

    在settings中添加一行设置;

    # 这里也可以用tuple 用tuple的话路径后面要加逗号
    STATICFILES_DIRS = [
    os.path.join(BASE_DIR, 'static') # 可以传递多个
    ]

    在index.html中把<link href="css/style.css" rel="stylesheet" type="text/css" />之类的导入css和js的语句改为

    {% load staticfiles %}
    <head>
    ...
    <link href="{% static 'css/style.css'%}" rel="stylesheet" type="text/css" />
    ...</head>

    这样就可以把settings中的static_url join到‘ ’ 内容前面,这样就可以找得到html文件了。

    搜索建议
    在虚拟环境中安装同版本的es-dsl
    f模糊搜索
    fuzzy:

    GET duowan/video/_search
    {
    "query": {
    "fuzzy": {
    "title": {
    "value": "军团骑士",
    "fuzziness": 2,
    "prefix_length": 3
    }
    }
    },
    "_source": ["title"]
    }

    fuzziness:编辑距离
    prefix_length:前面的不参与变换的词的长度
    “_source”: [“title”]:指明字段

    suggest:

    POST duowan/video/_search
    {
    "suggest": {
    "my-suggest": {
    "text":"PVQ",
    "completion": {
    "field": "suggest",
    "fuzzy": {
    "fuzziness":1
    }
    }
    }
    },
    "_source": ["title"]
    }

    my-suggest可以自定义,field不能变,

    在index.html文件中嵌入了js脚本,绑定了input事件,当里面的内容发生变化时,向服务器发送请求,参数包括input内容,和type类型

    $(function(){
    $('.searchInput').bind(' input propertychange ',function(){
    var searchText = $(this).val();
    var tmpHtml = ""
    $.ajax({
    cache: false,
    type: 'get', //get方法获取
    dataType:'json',
    url:suggest_url+"?s="+searchText+"&s_type="+$(".searchItem.current").attr('data-type'),
    async: true,
    success: function(data) {
    for (var i=0;i<data.length;i++){
    tmpHtml += '<li><a href="'+search_url+'?q='+data[i]+'">'+data[i]+'</a></li>'
    }
    $(".dataList").html("")
    $(".dataList").append(tmpHtml);
    if (data.length == 0){
    $('.dataList').hide()
    }else {
    $('.dataList').show()
    }
    }
    });
    } );
    })

    在urls中新增:

    url(r'^suggest/$', TemplateView.as_view(template_name='index.html'), name='index')
    1
    然后把爬虫文件中es-type中的内容复制到django项目的models中
    再编辑views文件:

    import json
    from django.shortcuts import render
    from django.views.generic.base import View
    from search.models import DuowanType
    from django.http import HttpResponse


    # Create your views here.
    # 继承 View
    class SearchSuggest(View):
    def get(self, request):
    key_words = request.GET.get('s', '') # 用request获取传过来的参数s 默认值为空
    re_dates = [] # 用来保存搜索建议返回来的title
    if key_words:
    s = DuowanType.search()
    # 写好查询语句
    s = s.suggest('my_suggest', key_words, completion={
    "field": "suggest",
    "fuzzy": {
    "fuzziness": 2
    },
    "size": 10
    })
    # 执行并获取结果
    suggestions = s.execute_suggest()
    for match in suggestions.my_suggest[0].options:
    source = match._source
    re_dates.append(source['title'])
    # 用HttpResponse来返回结果,把数组转成json返回
    return HttpResponse(json.dumps(re_dates), content_type='application/json')

    把urls中的

    url(r'^suggest/$', TemplateView.as_view(template_name='index.html'), name='index')
    改为

    url(r'^suggest/$', SearchSuggest.as_view(), name='suggest')

    记得是SearchSuggest.as_view(),不是SearchSuggest.as_view,否则会报错如下:
    TypeError: as_view() takes 1 positional argument but 2 were given

    二 搜索功能
    urls中:

    from search.views import SearchSuggest, SearchView
    ...
    url(r'^search/$', SearchView.as_view(), name='search')

    在views中添加一个 SearchView(View):
    接收传过来的查询关键词参数和页码参数,
    创建一个client连接es服务器,使用client.search可以执行原始的语句,使用client.search来执行查询语句,在接收返回来的值,把返回来的结果取出来存放到list中,最后用render返回给页面,
    查询时间:记录client.search运行前后的时间,再做减法

    from elasticsearch import Elasticsearch
    from datetime import datetime
    client = Elasticsearch(hosts='127.0.0.1')
    .......
    class SearchView(View):
    def get(self, request):
    key_words = request.GET.get('q', '')
    pagesize = 10
    page = request.GET.get('p', '')
    try:
    page = int(page)
    except:
    page = 1

    # client.search允许像最原始的写法一样写
    body = {
    "query": {
    "multi_match": {
    "query": key_words,
    "fields": ["title", "author"]
    }
    },
    "from": (page-1)*pagesize,
    "size": pagesize,
    # 高亮 返回来的值会把高亮的内容放到highlight字段中
    "highlight": {
    # 可以指明想要加进去的html tag tag里面可以写想知道的值
    "pre_tags": ["<span class='keyword'>"],
    "post_tags": ["</span>"],
    "fields": {
    "title": {},
    "content": {}
    }
    }
    }
    start_time = datetime.now()
    response = client.search(
    index="duowan",
    body=body
    )
    end_time = datetime.now()
    last_seconds = (end_time-start_time).total_seconds()
    # 不管分不分页都有的总数量
    total_nums = response['hits']['total']
    if (page % 10) > 0:
    page_nums = int(total_nums/10)+1
    else:
    page_nums = int(page/10)
    # 构造一些值 传到数组 在返回给html
    hit_list = []
    for hit in response['hits']['hits']:
    hit_dict = {}
    if 'title' in hit['highlight']:
    hit_dict['title'] = hit['highlight']['title'][0]
    else:
    # 截取长度 hit_dict['title'] = hit['_source']['title'][:100
    hit_dict['title'] = hit['_source']['title']
    hit_dict['len'] = hit['_source']['len']
    hit_dict['tag'] = hit['_source']['tag']
    hit_dict['update_time'] = hit['_source']['update_time']
    hit_dict['author'] = hit['_source']['author']
    hit_dict['playnum_text'] = hit['_source']['playnum_text']
    hit_dict['url'] = hit['_source']['url']
    hit_list.append(hit_dict)
    return render(request, 'result.html', {'page': page,
    'total_nums': total_nums,
    'all_hits': hit_list,
    'key_words': key_words,
    'page_nums': page_nums,
    'last_seconds': last_seconds})

    在页面中:
    找到item的div,用{% for hit in all_hits %} <div>...</div> {% endfor %}来使用for循环,遍历传过来的查询结果list all_hits.在页面中填充值

    {% for hit in all_hits %}
    <div class="resultItem">
    <div class="itemHead">
    <a href="{{ hit.url }}" target="_blank" class="title">{{ hit.title }}</a>
    <span class="divsion">-</span>
    <span class="fileType">
    <span class="label">分类:</span>
    <span class="value">{{ hit.tag }}</span>
    </span>
    <span class="dependValue">
    <span class="label">播放次数:</span>
    <span class="value">{{ hit.playnum_text }}</span>
    </span>
    </div>
    <div class="itemBody">

    </div>
    <div class="itemFoot">
    <span class="info">
    <label>网站:</label>
    <span class="value">伯乐在线</span>
    </span>
    <span class="info">
    <label>发布时间:</label>
    <span class="value">{{ hit.update_time }}</span>
    </span>
    </div>
    </div>
    {% endfor %}

    用js实现搜索记录:
    点击搜索按钮的时候触发add_search()方法,获取关键词,再用KillRepeat()给搜索记录去重,去重后把数组存储到浏览器localStorage,然后再把搜索内容显示出来

    //点击搜索的时候触发
    function add_search(){
    var val = $(".searchInput").val();
    if (val.length>=2){
    //点击搜索按钮时,去重
    KillRepeat(val);
    //去重后把数组存储到浏览器localStorage
    localStorage.search = searchArr;
    //然后再把搜索内容显示出来
    MapSearchArr();
    }

    window.location.href=search_url+'?q='+val+"&s_type="+$(".searchItem.current").attr('data-type')

    }

    function MapSearchArr(){
    var tmpHtml = "";
    var arrLen = 0
    if (searchArr.length >= 5){
    arrLen = 5
    }else {
    arrLen = searchArr.length
    }
    //把数组内容拼接成html内容
    for (var i=0;i<arrLen;i++){
    tmpHtml += '<a href="'+search_url+'?q='+searchArr[i]+'">'+searchArr[i]+'</a>'
    }
    $(".mysearch .all-search").html(tmpHtml);
    }
    //去重 把以前搜索过的记录删除,并把本次搜索词放在前面
    function KillRepeat(val){
    var kill = 0;
    for (var i=0;i<searchArr.length;i++){
    //判断这个词是否存在历史搜索记录里
    if(val===searchArr[i]){
    kill ++;
    }
    }
    if(kill<1){//不存在
    //放到队列头部
    searchArr.unshift(val);
    }else {//存在
    //把原来的值删除
    removeByValue(searchArr, val)
    searchArr.unshift(val)
    }
    }

    原文:https://blog.csdn.net/qq_40916110/article/details/87855502

  • 相关阅读:
    Linux/shell: remove adjacent similar patterns
    Calculate difference between consecutive data points in a column from a file
    awk
    自定义Cordova插件(基础篇)
    npm init 命令生成package.json文件
    自定义Cordova插件详解
    Android 回退键监听
    Cordova结合Vue学习Camera
    解决悬浮的<header>、<footer>遮挡内容的处理技巧
    npm 是干什么的
  • 原文地址:https://www.cnblogs.com/hanzeng1993/p/11280518.html
Copyright © 2020-2023  润新知