1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
| import os
import re
import feedparser
import urllib.parse
import requests
from bs4 import BeautifulSoup
from flask import Flask, request, jsonify
from flask_cors import CORS
from langchain_tavily import TavilySearch
from langchain.agents import AgentExecutor, create_tool_calling_agent, tool
from langchain_core.prompts import ChatPromptTemplate
from langchain.chat_models import init_chat_model
import jieba
import jieba.analyse
from dotenv import load_dotenv
import logging
from datetime import datetime
load_dotenv(override=True)
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
)
logger = logging.getLogger(__name__)
app = Flask(__name__)
CORS(app, origins=[
"http://smallgoodgood.top",
"https://smallgoodgood.top",
"http://www.smallgoodgood.top",
"https://www.smallgoodgood.top",
"http://localhost:*"
])
search = TavilySearch(max_results=3)
jieba.initialize()
def fetch_blog_content(url):
try:
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'
}
response = requests.get(url, headers=headers, timeout=10)
response.raise_for_status()
soup = BeautifulSoup(response.content, 'html.parser')
content_selectors = [
'article',
'.post-content',
'.entry-content',
'.blog-content',
'div[class*="content"]',
'main'
]
content = None
for selector in content_selectors:
content_element = soup.select_one(selector)
if content_element:
content = content_element.get_text(strip=False)
break
if not content:
body = soup.find('body')
if body:
for element in body(['nav', 'footer', 'header', 'aside', 'script', 'style']):
element.decompose()
content = body.get_text(strip=False)
if content:
content = re.sub(r'\n\s*\n', '\n\n', content)
content = re.sub(r'^\s+', '', content, flags=re.MULTILINE)
return content[:10000]
return "无法提取文章内容"
except Exception as e:
return f"抓取文章内容时出错: {str(e)}"
@tool
def search_blog_rss(query: str):
try:
feed = feedparser.parse("https://smallgoodgood.top/rss.xml")
chinese_keywords = jieba.analyse.extract_tags(query, topK=5, withWeight=False)
english_words = set(re.findall(r'[a-zA-Z0-9]{3,}', query.lower()))
query_terms = set(chinese_keywords) | english_words
query_lower = query.lower()
results = []
for entry in feed.entries:
title = entry.title
description = entry.get('description', '')
if description:
description = re.sub(r'<[^>]+>', '', description)
search_text = f"{title} {description}".lower()
title_lower = title.lower()
score = 0
if any(term in title_lower for term in query_terms if len(term) > 1):
score += 20
elif any(term in search_text for term in query_terms if len(term) > 1):
score += 10
if query_lower in search_text:
score += 15
content_keywords = set(jieba.analyse.extract_tags(
f"{title} {description}", topK=10, withWeight=False
))
keyword_match = len(query_terms & content_keywords)
score += keyword_match * 3
if score > 0:
published = entry.get('published', entry.get('pubDate', '未知日期'))
full_content = fetch_blog_content(entry.link)
results.append({
'title': title,
'link': entry.link,
'published': published,
'summary': description[:200] + '...' if description else "暂无摘要",
'full_content': full_content,
'score': score
})
results.sort(key=lambda x: x['score'], reverse=True)
if results:
response = f"在博客中找到 {len(results)} 篇相关文章:\n\n"
for i, art in enumerate(results[:3], 1):
response += f"{i}. **{art['title']}** (相关度: {art['score']})\n"
response += f" 发布时间: {art['published']}\n"
response += f" 摘要: {art['summary']}\n"
response += f" 链接: {art['link']}\n\n"
if len(art['full_content']) > 500:
response += f" 内容预览: {art['full_content'][:500]}...\n\n"
else:
response += f" 内容: {art['full_content']}\n\n"
return response
else:
return "在博客中没有找到相关文章。您可以尝试使用其他关键词搜索。"
except Exception as e:
return f"读取博客RSS时出错: {str(e)}"
@tool
def get_blog_article(url: str):
try:
content = fetch_blog_content(url)
return f"文章内容:\n\n{content}"
except Exception as e:
return f"获取文章内容时出错: {str(e)}"
tools = [search_blog_rss, get_blog_article]
prompt = ChatPromptTemplate.from_messages([
("system", """你是一名助人为乐的助手，可以搜索smallgoodgood.top博客的内容来获取信息。
重要指令：
1. 当用户询问技术问题时，先使用search_blog_rss搜索博客中是否有相关文章
2. 找到相关文章后，使用get_blog_article获取文章的完整内容
3. 基于文章的完整内容进行详细回答，不要只依赖摘要
4. 引用具体的技术细节和方法
5. 如果文章内容不够详细，可以补充一般性的技术知识
请确保基于真实内容提供准确的回答。"""),
("human", "{input}"),
("placeholder", "{agent_scratchpad}"),
])
model = init_chat_model("deepseek-chat", model_provider="deepseek")
agent = create_tool_calling_agent(model, tools, prompt)
agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)
@app.route('/health', methods=['GET'])
def health_check():
return jsonify({
"status": "healthy",
"service": "Chat Zenith Langchain Service",
"timestamp": datetime.now().isoformat()
})
@app.route('/chat', methods=['POST', 'OPTIONS'])
def chat():
if request.method == 'OPTIONS':
return '', 204
try:
data = request.get_json()
if not data or 'messages' not in data:
return jsonify({
"error": "请求格式错误，需要包含messages字段"
}), 400
messages = data['messages']
user_message = None
for msg in reversed(messages):
if msg.get('role') == 'user':
user_message = msg.get('content', '')
break
if not user_message:
return jsonify({
"error": "没有找到用户消息"
}), 400
logger.info(f"收到用户消息: {user_message}")
result = agent_executor.invoke({"input": user_message})
response_data = {
"choices": [{
"message": {
"content": result['output'],
"role": "assistant"
}
}],
"model": "langchain-deepseek"
}
logger.info(f"生成回复: {result['output'][:100]}...")
return jsonify(response_data)
except Exception as e:
logger.error(f"处理请求时出错: {str(e)}", exc_info=True)
return jsonify({
"error": f"服务器内部错误: {str(e)}"
}), 500
@app.errorhandler(404)
def not_found(error):
return jsonify({
"error": "端点不存在"
}), 404
@app.errorhandler(500)
def internal_error(error):
logger.error(f"内部服务器错误: {str(error)}")
return jsonify({
"error": "内部服务器错误"
}), 500
if __name__ == '__main__':
logger.info("启动Chat Zenith Web服务...")
app.run(
host='0.0.0.0',
port=5000,
debug=False
)
|
