为基于elasticsearch和flask的web端检索系统增加新的检索功能
序言
距离上次写博客已经有20天的光阴了，期间我为之前构建的系统增加了新的检索功能，这是记录了之前检索系统的博客：
https://timtian.blog.csdn.net/article/details/124319573?spm=1001.2014.3001.5502系统展示
基于Flask和ES的检索系统的系统演示
在上面的视屏中可以看到，我们实现了多字段的检索，entity_id和entity_type是分开检索的，但是模板基本一致，都都rename（重命名功能），以及字段的增删功能。
前端字段增删功能的实现
首先我们要注意的是，按某一行的“-”号按钮，就一定要把该删除行，而不是删掉其他的行。我的想法是通过JS的DOM树的parentNode方法实现，首先先得知道该页面的html是怎样的：
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<title>search entity id</title>
<link rel="stylesheet" href="{{ url_for('static', filename='css/bootstrap.min.css') }}">
<script src="//cdn.bootcss.com/jquery/1.11.3/jquery.min.js"></script>
<script src="{{ url_for('static', filename='js/bootstrap.bundle.js') }}"></script>
<script src="{{ url_for('static', filename='js/bootstrap.bundle.min.js') }}"></script>
</head>
<body>
<nav aria-label="breadcrumb">
<ol class="breadcrumb">
<li class="breadcrumb-item"><a href="{{ url_for('index') }}">首页</a></li>
<li class="breadcrumb-item active" aria-current="page">entity_id</li>
<li class="breadcrumb-item"><a href="{{ url_for('entity_type') }}">entity_type</a></li>
</ol>
</nav>
<form method="post" enctype="multipart/form-data" action="{{ url_for('search_entity_id') }}">
<div class="jumbotron">
<h1 class="display-4">Enter Entity_id</h1>
<hr class="my-4">
<div class="input-group mb-3">
<select class="custom-select" name="luoji_1">
<option value="AND" selected>AND</option>
<option value="OR">OR</option>
<option value="NOT">NOT</option>
</select>
<div class="input-group-prepend">
<span class="input-group-text">entity_id</span>
</div>
<input name="entity_id_1" type="text" class="form-control" aria-describedby="basic-addon1">
<div class="input-group-append" id="1">
<span class="input-group-text" id="add">+</span>
</div><br/></div><div class="input-group mb-3" id="flag1">
</div>
<div class="input-group mb-3">
<div class="input-group-prepend">
<span class="input-group-text" id="basic-addon3" >Rename your file</span>
</div>
<input type="text" class="form-control" id="basic-url" name="rename">
</div>
<hr class="my-4">
<input class="btn btn-primary btn-lg" type="submit" onclick="getNumber()">
<ul>
{% for message in get_flashed_messages() %}
<div class="alert alert-warning alert-dismissible fade show" role="alert">
{{ message }}
<button type="button" class="close" data-dismiss="alert" aria-label="Close">
<span aria-hidden="true">×</span>
</button>
</div>
{% endfor %}
</ul>
</div>
</form>
</body>
</html>可以看出来， “-”号按钮的那一行是“-”的爷爷辈分，所以我们可以获取“-”号的太爷爷辈分的节点，进而删除爷爷辈分的节点。代码如下：
function jian(obj){
//alert(obj);
var parentOfGrandparent = obj.parentNode.parentNode.parentNode;
var grandparent = obj.parentNode.parentNode;
parentOfGrandparent.removeChild(grandparent);
number = document.getElementById("total_num");
//cnt--;
aim = obj.parentNode.id;
var pos = 0;
for(var i=0;i<a.length;i++){
if(a[i] == aim){
pos = i;
}
}
a.splice(pos, 1);
}你会发现多出来了一段：
aim = obj.parentNode.id;
var pos = 0;
for(var i=0;i<a.length;i++){
if(a[i] == aim){
pos = i;
}
}
a.splice(pos, 1);
这一段是为了删除a数组中的编号，前端需要和后端交互，后端就需要知道前端删掉了哪些行，又保留了那些行。我们可以将行号存放在一个数组中，记录前端存在的行号，而上述步骤就是删去被删除行的行号。通过for循环遍历存放行号的数组a，aim是删去的行号，若a[i]与之相同，便将a[i]移除。aim是通过“-”号的父节点的id属性来获取的。于是在创建一行时，就需要赋予id这个属性，以方便删除时使用。
接下来看添加行的JS代码：
var flag1 = document.getElementById("flag1");
var add = document.getElementById("add");
var cnt = 1;
var a = new Array();
a.push(cnt);
add.onclick = function(){
cnt++;
//history++;
number = document.getElementById("total_num");
//number.innerHTML = cnt + "";
a.push(cnt);
div_tmp = document.createElement("div");
div_tmp.setAttribute("class", "input-group mb-3");
flag1.appendChild(div_tmp);
select = document.createElement("select");
select.setAttribute("class", "custom-select");
select.setAttribute("name", "luoji_" + cnt);
div_tmp.appendChild(select);
option1 = document.createElement("option");
option1.setAttribute("value", "AND");
//option1.setAttribute("selected");
option1.innerHTML = "AND";
select.appendChild(option1);
option2 = document.createElement("option");
option2.setAttribute("value", "OR");
option2.innerHTML = "OR";
select.appendChild(option2);
option3 = document.createElement("option");
option3.setAttribute("value", "NOT");
option3.innerHTML = "NOT";
select.appendChild(option3);
div_tmp3 = document.createElement("div");
div_tmp3.setAttribute("class", "input-group-prepend");
div_tmp.appendChild(div_tmp3);
span_tmp = document.createElement("span");
span_tmp.setAttribute("class", "input-group-text");
span_tmp.innerHTML = "entity_id";
div_tmp3.appendChild(span_tmp);
input_tmp = document.createElement("input");
input_tmp.setAttribute("class", "form-control");
input_tmp.setAttribute("type", "text");
input_tmp.setAttribute("name", "entity_id_" + cnt);
div_tmp.appendChild(input_tmp);
div_tmp2 = document.createElement("div");
div_tmp2.setAttribute("class", "input-group-append");
div_tmp2.setAttribute("id", cnt + "");
div_tmp.appendChild(div_tmp2);
span_tmp3 = document.createElement("span");
span_tmp3.setAttribute("class", "input-group-text");
span_tmp3.innerHTML = "-";
span_tmp3.setAttribute("onclick", "jian(this)");
div_tmp2.appendChild(span_tmp3);
span_tmp3 = document.createElement("span");
}同样创建行的功能也是同过DOM树实现的。这个其实没什么好讲，就是通过原有的第一行的HTML来构建新的行。重点是里面的属性，属性的标号通过cnt来计数，每增加一行，cnt就自增，以告知当前是新增的第几行，假设a数组增加了五行，那a数组的内容是这样的{1，2，3，4，5}。然后假设删掉第三行，那a数组的内容是这样的{1，2，4，5}。此时在增加一行是这样的{1，2，4，5， 6}，而不是这样的{1，2，4，5， 3}。总之，已经删掉的行是不会再重复了。然后后端就可以通过这个a数组，获得当前前端页面存在的每个行的id值，进而获得各个选项的内容。你会发现这些选项均与id相关，例如：entity_id_5,luoji_5。luoji_5指的是行号为5的逻辑符是NOT还是AND抑或是OR。而entity_id_5则是行号为5的行的entity_id的内容。
还有一个js代码是发送ajax，用来发送数组a的内容给后端，代码如下，不再赘述。
function getNumber(){
$.ajax({
type: 'POST',
url: '/testGet',
dataType: 'json',
data: {
//'number': document.getElementById("total_num").innerHTML,
'array': a,
},
success: function(res){
console.log(res)
},
error: function(){
console.log('error')
}
})
}entity_type页面实现类似，这里不再赘述。
系统的后端实现
代码如下：
@app.route('/search_entity_id/', methods=["POST", "GET"])
def search_entity_id():
if request.method == "POST":
entity_id = []
luoji= []
for num in lis:
entity_id.append(request.form["entity_id_" + num])
luoji.append(request.form["luoji_" + num])
rename = request.form["rename"]
search_by_entity_id(entity_id, luoji, rename)
return render_template('search_entity_id.html')
elif request.method == "GET":
return render_template('search_entity_id.html')
def search_by_entity_id(entity_id, luoji, rename):
es = Elasticsearch(hosts=['http://52.14.194.191:9200'], request_timeout=60)
body1 = {
"query": {
"bool": {
"must": [],
"should": [],
"must_not": []
}
},
"track_total_hits": "true"
}
for i in range(len(luoji)):
if luoji[i] == 'AND':
flash(luoji[i])
body1["query"]["bool"]["must"].append(
{
"match": {
"entities.entity_id": entity_id[i]
}
})
elif luoji[i] == 'OR':
flash(luoji[i])
body1["query"]["bool"]["should"].append(
{
"match": {
"entities.entity_id": entity_id[i]
}
})
elif luoji[i] == 'NOT':
flash(luoji[i])
body1["query"]["bool"]["must_not"].append(
{
"match": {
"entities.entity_id": entity_id[i]
}
})
res = es.search(index="pubmed-paper-index-2", body=body1)
total_value = res['hits']['total']['value']
for i in range(total_value // 10):
body = {
"query": {
"bool": {
"must": [],
"should": [],
"must_not": []
}
},
"track_total_hits": "true",
"from": i,
"size": 10
}
for i in range(len(luoji)):
if luoji[i] == "AND":
body["query"]["bool"]["must"].append(
{
"match": {
"entities.entity_id": entity_id[i]
}
})
elif luoji[i] == "OR":
body["query"]["bool"]["should"].append(
{
"match": {
"entities.entity_id": entity_id[i]
}
})
elif luoji[i] == "NOT":
body["query"]["bool"]["must_not"].append(
{
"match": {
"entities.entity_id": entity_id[i]
}
})
results = es.search(index="pubmed-paper-index-2", body=body)
result_list = results['hits']['hits']
try:
for result in result_list:
pmid = result['_source']['pmid']
title = result['_source']['title']
abstract = result['_source']['abstract']
entity_list = result['_source']['entities']
for entity in entity_list:
tmp_entity_string = entity['entity_string']
tmp_entity_type = entity['entity_type']
tmp_entity_id = entity['entity_id']
tmp_entity_span = str(entity['span'])
if rename is not None:
f = open(f"query/{rename}.csv", 'a', newline='')
writer = csv.writer(f)
writer.writerow([
pmid,
title,
abstract,
tmp_entity_string,
tmp_entity_type,
tmp_entity_id,
tmp_entity_span,
])
f.close()
else:
f = open(f"query/data.csv", 'a', newline='')
writer = csv.writer(f)
writer.writerow([
pmid,
title,
abstract,
tmp_entity_string,
tmp_entity_type,
tmp_entity_id,
tmp_entity_span,
])
f.close()
except Exception:
pass
if rename is not None:
os.system(f"aws s3 cp query/{rename}.csv s3://meta-adhoc/nlp/es/ --recursive")
else:
os.system(f"aws s3 cp query/data.csv s3://meta-adhoc/nlp/es/ --recursive")
flash('Over!!!')它的逻辑是按照AND、NOT和OR的逻辑分别转化为es的must、must_not和should语法，然后进行检索，检索结果存放放在ec2对应的位置即可。不是很复杂。
以上就是这么多
END
浙公网安备 33010602011771号