python使用django调用deepseek api搭建ai网站

风清扬电脑知识

2025-02-26 0 398 百度已收录

一、deepseek简介

DeepSeek是一家人工智能公司，专注于开发先进的人工智能模型和技术。以下是关于DeepSeek的一些详细介绍：

1.公司背景

DeepSeek由杭州深度求索人工智能基础技术研究有限公司开发，致力于通过创新的技术和算法，推动人工智能领域的发展。

2.技术与模型

DeepSeek-V3：这是DeepSeek开发的一个大型语言模型，具有超过600B的参数，在多项性能指标上与国际顶尖模型相当。
DeepSeek-R1：这是DeepSeek的第一代推理模型，通过大规模强化学习（RL）进行训练，展示出了在推理任务上的优异性能。
DeepSeek-R1-Distill：这些是从DeepSeek-R1中蒸馏出的小模型，具有更好的性能和效率，适合在资源受限的环境中使用。

3.应用领域

自然语言处理：DeepSeek的模型在文本生成、知识问答、推理等任务中表现出色，能够为用户提供高质量的语言交互服务。
智能助手：DeepSeek开发了AI智能助手，可用于搜索、写作、阅读、解题、翻译等多种任务，帮助用户提高效率。

4.优势与特点

成本优势：DeepSeek的模型训练成本低，调用接口成本也较低，具有较高的性价比。
中文处理能力强：对中文语法、成语、文化背景理解更深入，在中文文本生成、摘要、情感分析等任务中表现自然。
开源优势：DeepSeek-R1模型权重和技术报告开源，便于开发者二次开发和创新。

5.产品与服务

DeepSeek API：提供与OpenAI兼容的API，方便开发者将DeepSeek的模型集成到自己的应用中。
DeepSeek App：提供AI智能助手应用，可在App Store上下载，支持多种功能，如智能对话、搜索、写作等。

二、获取apikey

deepseek官方的api暂时无法充值，我使用的是阿里云的百炼平台的deepseek v1模型，阿里云百炼平台注册送百万token，可以白嫖。打开百炼控制台，开通服务，随便选择一个模型，点击右上角的“查看我的apikey”，获取apikey。

三、创建django项目，并startapp

使用django创建一个新的项目，python manage.py startapp chat新建app作为主要代码文件夹。

四、编写代码

chat\views.py

from django.shortcuts import render
from openai import OpenAI
import os
from django.conf import settings
 
 
def get_ai_response(messages):
client = OpenAI(
api_key="xxx",//填写apikey
base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
)
 
 
try:
    completion = client.chat.completions.create(
        model="qwen2.5-14b-instruct-1m",
        messages=messages
    )
    return {
        'content': completion.choices[0].message.content,
        'reasoning': getattr(completion.choices[0].message, 'reasoning_content', '')
    }
except Exception as e:
    return {
        'content': f"发生错误：{str(e)}",
        'reasoning': ''
    }
 
def chat_view(request):
if 'messages' not in request.session:
request.session['messages'] = []
 
 
if request.method == 'POST':
    user_message = request.POST.get('message', '')
    if user_message:
        request.session['messages'].append({'role': 'user', 'content': user_message})
 
        response = get_ai_response(request.session['messages'])
 
        request.session['messages'].append({
            'role': 'assistant',
            'content': response['content'],
            'reasoning': response['reasoning']
        })
 
        request.session.modified = True
 
return render(request, 'chat.html', {
    'messages': request.session['messages']
})

要将api_key=”xxx”中xxx替换为自己的apikey。

urls.py

from django.contrib import admin
from django.urls import path
from chat import views
 
urlpatterns = [
    path('admin/', admin.site.urls),
    path('', views.chat_view, name='chat'),
]

前端代码

<!DOCTYPE html>
<html>
<head>
    <title>AI对话助手（Markdown支持版）</title>
    <!-- Markdown 渲染依赖 -->
    <link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/github-markdown-css@5.2.0/github-markdown.min.css" rel="external nofollow" >
    <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/styles/github-dark.min.css" rel="external nofollow" >
    <style>
        :root {
            --user-color: #1a73e8;
            --assistant-color: #0b8043;
        }
        body {
            max-width: 800px;
            margin: 0 auto;
            padding: 20px;
            font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif;
            background-color: #f8f9fa;
        }
        .chat-container {
            background: white;
            border-radius: 12px;
            box-shadow: 0 2px 8px rgba(0,0,0,0.1);
            padding: 20px;
            margin-bottom: 20px;
        }
        .message {
            margin: 15px 0;
            padding: 12px 16px;
            border-radius: 8px;
        }
        .user-message {
            background-color: #e8f0fe;
            border: 1px solid var(--user-color);
            margin-left: 30px;
        }
        .assistant-message {
            background-color: #e6f4ea;
            border: 1px solid var(--assistant-color);
            margin-right: 30px;
        }
        .role-label {
            font-weight: 500;
            margin-bottom: 8px;
            display: flex;
            align-items: center;
            gap: 8px;
        }
        .role-label::before {
            content: '';
            display: inline-block;
            width: 12px;
            height: 12px;
            border-radius: 50%;
        }
        .user-message .role-label::before {
            background: var(--user-color);
        }
        .assistant-message .role-label::before {
            background: var(--assistant-color);
        }
        form {
            display: flex;
            gap: 10px;
            margin-top: 20px;
        }
        input[type="text"] {
            flex: 1;
            padding: 12px;
            border: 1px solid #ddd;
            border-radius: 8px;
            font-size: 16px;
        }
        button {
            padding: 12px 24px;
            background-color: var(--user-color);
            color: white;
            border: none;
            border-radius: 8px;
            cursor: pointer;
            transition: background 0.2s;
        }
        button:hover {
            background-color: #1557b0;
        }
        .markdown-body pre {
            padding: 16px;
            border-radius: 8px;
            overflow-x: auto;
        }
        .reasoning-box {
            margin-top: 12px;
            padding: 12px;
            background: #fff8e5;
            border-left: 4px solid #ffd700;
            border-radius: 4px;
        }
    </style>
</head>
<body>
    <div class="chat-container">
        <h1 style="color: var(--user-color); text-align: center;">AI对话助手</h1>
 
        <div class="messages">
            {% for message in messages %}
                <div class="message {% if message.role == 'user' %}user-message{% else %}assistant-message{% endif %}">
                    <div class="role-label">
                        {% if message.role == 'user' %}
                            👤 用户
                        {% else %}
                            🤖 助手
                        {% endif %}
                    </div>
                    <!-- Markdown 内容容器 -->
                    <div class="markdown-body"
                         data-markdown="{{ message.content|escape }}"
                         data-raw="{{ message.content|escape }}">
                        {{ message.content|safe }}
                    </div>
 
                    {% if message.reasoning %}
                    <div class="reasoning-box">
                        <div class="reasoning-label">💡 思考过程</div>
                        <div class="markdown-body"
                             data-markdown="{{ message.reasoning|escape }}"
                             data-raw="{{ message.reasoning|escape }}">
                            {{ message.reasoning|safe }}
                        </div>
                    </div>
                    {% endif %}
                </div>
            {% endfor %}
        </div>
 
        <form method="post">
            {% csrf_token %}
            <input type="text" name="message" placeholder="请输入您的问题..." required autofocus>
            <button type="submit">发送</button>
        </form>
    </div>
 
    <!-- 依赖库 -->
    <script src="https://cdn.jsdelivr.net/npm/marked@12.0.0/lib/marked.min.js"></script>
    <script src="https://cdn.jsdelivr.net/npm/dompurify@3.0.5/dist/purify.min.js"></script>
    <script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.7.0/highlight.min.js"></script>
 
    <script>
        // 初始化配置
        marked.setOptions({
            breaks: true,
            highlight: (code, lang) => {
                const language = hljs.getLanguage(lang) ? lang : 'plaintext';
                return hljs.highlight(code, { language }).value;
            }
        });
 
        // Markdown 渲染函数
        const renderMarkdown = () => {
            document.querySelectorAll('.markdown-body').forEach(container => {
                try {
                    // 优先使用 data-markdown 属性
                    const raw = container.dataset.markdown || container.dataset.raw;
                    const clean = DOMPurify.sanitize(raw, {
                        ALLOWED_TAGS: ['b', 'i', 'em', 'strong', 'code', 'pre', 'blockquote',
                                     'ul', 'ol', 'li', 'h1', 'h2', 'h3', 'a', 'p', 'br', 'hr'],
                        ALLOWED_ATTR: ['href', 'target']
                    });
                    container.innerHTML = marked.parse(clean);
                } catch (e) {
                    console.error('Markdown渲染失败:', e);
                    container.innerHTML = container.dataset.raw;
                }
            });
 
            // 触发代码高亮
            hljs.highlightAll();
        };
 
        // 确保依赖加载完成后执行
        const checkDependencies = () => {
            if (window.marked && window.DOMPurify && window.hljs) {
                renderMarkdown();
            } else {
                setTimeout(checkDependencies, 100);
            }
        };
 
        // 启动渲染流程
        document.addEventListener('DOMContentLoaded', checkDependencies);
    </script>
</body>
</html>