flask入门（四）前后端数据传输

文章目录

1、flask后端接收来自前端的数据

1）如果前端提交的方法为POST

后端接收时的代码：

xx=request.form.get('xx');
xx=request.form['xx']

2）如果前段提交的方法是GET

xx=request.args.get(xx)

2、flask后端向前端传数据

传送单个数据：

return render_template('html文件',xx='xx')

传送多个数据：先把数据写进字典，字典整体进行传输

return render_template('html文件',xx='字典变量')

3、案例

目录结构：
在这里插入图片描述
index.py文件：

# --*-- coding:utf-8 --*--
# @Author : 一只楚楚猫
# @File : index.py
# @Software : PyCharm

from flask import *
from sentence_transformers import SentenceTransformer
import torch.nn as nn
import torch
import torch.nn.functional as F

model = SentenceTransformer(r'E:楚楚猫codepython1design1creativity1distanceall-MiniLM-L6-v2')

app = Flask(__name__)

result = dict()
result["results"] = ""


@app.route('/', methods=('GET', 'POST'))
def index():
    global result
    if request.method == 'POST':
        step1 = request.form.get("step1")
        step2 = request.form.get("step2")
        step3 = request.form.get("step3")
        step4 = request.form.get("step4")

        # 用户输入的内容
        sentences = [step1, step2, step3, step4]

        results = list()

        # 384维
        embeddings = torch.FloatTensor(model.encode(sentences))

        # p=2就是计算欧氏距离，p=1就是曼哈顿距离
        euclidean_distance = nn.PairwiseDistance(p=2)

        for i in range(0, embeddings.size()[0]):
            for j in range(i + 1, embeddings.size()[0]):
                cosine_similarity = round(F.cosine_similarity(embeddings[i], embeddings[j], dim=0).item(), 4)
                distance = round(euclidean_distance(embeddings[i], embeddings[j]).item(), 4)

                results.append(
                    f"step{i + 1} & step{j + 1}的相关性：{cosine_similarity}       step{i + 1} & step{j + 1}的距离：{distance}")

                print(
                    f"step{i + 1} & step{j + 1}之间的相关性：{cosine_similarity}step{i + 1} & step{j + 1}之间的距离：{distance}")

        result["results"] = results

        return render_template('hello.html', result=result)

    return render_template('hello.html', result=result)


if __name__ == '__main__':
    app.run(port=11252)

hello.html文件：

<!DOCTYPE html>
<html lang="en">
<meta charset="utf-8">
<head>
    <meta name="viewport" content="width=device-width, initial-scale=1.0, maximum-scale=1.0, user-scalable=no"/>
    <meta http-equiv="content-type" content="text/html;charset=utf-8">
    <meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1">
    <title>欢迎来到我的世界</title>

    <link href="../static/css/style.css" rel="stylesheet" type="text/css">
</head>

<div id="container">
    <div id="output">
        <div class="containerT">
            <h1>Yimo</h1>
            <form class="form" id="entry_form" method="post">
                <h2><input type="text" name="step1" style="height:30px"></h2><br>
                <h2><input type="text" name="step2" style="height:30px"></h2><br>
                <h2><input type="text" name="step3" style="height:30px"></h2><br>
                <h2><input type="text" name="step4" style="height:30px"></h2><br>
                <center>
                    <button><h3>Click me!(๑•̀ㅂ•́)و✧</h3></button>
                </center>

                {{result.results}}
            </form>
        </div>
    </div>
</div>


</body>
</html>

上面涉及到的sentence_transformers模块来自论文：《Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks》，使用孪生网络（Siamse Network）将两个文本映射到特征空间得到特征向量（384维）

参考文献

1、flask 前端（html）与后端（python）相互传数据：https://www.cnblogs.com/zzai/p/html_dataStream.html
2、如何从Python发布数据并进入Flask前端：https://www.cnpython.com/qa/1513860