media_engine_streamlit_app.py
8.39 KB
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
"""
Streamlit Web界面
为Deep Search Agent提供友好的Web界面
"""
import os
import sys
import streamlit as st
from datetime import datetime
import json
# 添加src目录到Python路径
sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..'))
from MediaEngine import DeepSearchAgent, Config
from config import DEEPSEEK_API_KEY, BOCHA_Web_Search_API_KEY, GEMINI_API_KEY
def main():
"""主函数"""
st.set_page_config(
page_title="Deep Search Agent",
page_icon="🔍",
layout="wide"
)
st.title("Deep Search Agent")
st.markdown("基于DeepSeek的无框架深度搜索AI代理")
# 侧边栏配置
with st.sidebar:
st.header("配置")
# 高级配置
st.subheader("高级配置")
max_reflections = st.slider("反思次数", 1, 5, 2)
max_content_length = st.number_input("最大内容长度", 1000, 50000, 20000)
# 模型选择
llm_provider = st.selectbox("LLM提供商", ["deepseek", "openai", "gemini"])
openai_key = "" # 初始化变量
if llm_provider == "deepseek":
model_name = st.selectbox("DeepSeek模型", ["deepseek-chat"])
elif llm_provider == "openai":
model_name = st.selectbox("OpenAI模型", ["gpt-4o-mini", "gpt-4o"])
openai_key = st.text_input("OpenAI API Key", type="password", value="")
else: # gemini
model_name = st.selectbox("Gemini模型", ["gemini-2.5-pro"])
# 主界面
col1, col2 = st.columns([2, 1])
with col1:
st.header("研究查询")
query = st.text_area(
"请输入您要研究的问题",
placeholder="例如：2025年人工智能发展趋势",
height=100
)
# 预设查询示例
st.subheader("示例查询")
example_queries = [
"2025年人工智能发展趋势",
"深度学习在医疗领域的应用",
"区块链技术的最新发展",
"可持续能源技术趋势",
"量子计算的发展现状"
]
selected_example = st.selectbox("选择示例查询", ["自定义"] + example_queries)
if selected_example != "自定义":
query = selected_example
with col2:
st.header("状态信息")
if 'agent' in st.session_state and hasattr(st.session_state.agent, 'state'):
progress = st.session_state.agent.get_progress_summary()
st.metric("总段落数", progress['total_paragraphs'])
st.metric("已完成", progress['completed_paragraphs'])
st.progress(progress['progress_percentage'] / 100)
else:
st.info("尚未开始研究")
# 执行按钮
col1, col2, col3 = st.columns([1, 1, 1])
with col2:
start_research = st.button("开始研究", type="primary", use_container_width=True)
# 验证配置
if start_research:
if not query.strip():
st.error("请输入研究查询")
return
if llm_provider == "openai" and not openai_key:
st.error("请提供OpenAI API Key")
return
# 自动使用配置文件中的API密钥
deepseek_key = DEEPSEEK_API_KEY
gemini_key = GEMINI_API_KEY # 使用config.py中的Gemini API密钥
bocha_key = BOCHA_Web_Search_API_KEY
# 创建配置
config = Config(
deepseek_api_key=deepseek_key if llm_provider == "deepseek" else None,
openai_api_key=openai_key if llm_provider == "openai" else None,
gemini_api_key=gemini_key if llm_provider == "gemini" else None,
bocha_api_key=bocha_key,
default_llm_provider=llm_provider,
deepseek_model=model_name if llm_provider == "deepseek" else "deepseek-chat",
openai_model=model_name if llm_provider == "openai" else "gpt-4o-mini",
gemini_model=model_name if llm_provider == "gemini" else "gemini-2.5-pro",
max_reflections=max_reflections,
max_content_length=max_content_length,
output_dir="media_engine_streamlit_reports"
)
# 执行研究
execute_research(query, config)
def execute_research(query: str, config: Config):
"""执行研究"""
try:
# 创建进度条
progress_bar = st.progress(0)
status_text = st.empty()
# 初始化Agent
status_text.text("正在初始化Agent...")
agent = DeepSearchAgent(config)
st.session_state.agent = agent
progress_bar.progress(10)
# 生成报告结构
status_text.text("正在生成报告结构...")
agent._generate_report_structure(query)
progress_bar.progress(20)
# 处理段落
total_paragraphs = len(agent.state.paragraphs)
for i in range(total_paragraphs):
status_text.text(f"正在处理段落 {i+1}/{total_paragraphs}: {agent.state.paragraphs[i].title}")
# 初始搜索和总结
agent._initial_search_and_summary(i)
progress_value = 20 + (i + 0.5) / total_paragraphs * 60
progress_bar.progress(int(progress_value))
# 反思循环
agent._reflection_loop(i)
agent.state.paragraphs[i].research.mark_completed()
progress_value = 20 + (i + 1) / total_paragraphs * 60
progress_bar.progress(int(progress_value))
# 生成最终报告
status_text.text("正在生成最终报告...")
final_report = agent._generate_final_report()
progress_bar.progress(90)
# 保存报告
status_text.text("正在保存报告...")
agent._save_report(final_report)
progress_bar.progress(100)
status_text.text("研究完成！")
# 显示结果
display_results(agent, final_report)
except Exception as e:
st.error(f"研究过程中发生错误: {str(e)}")
def display_results(agent: DeepSearchAgent, final_report: str):
"""显示研究结果"""
st.header("研究结果")
# 结果标签页
tab1, tab2, tab3 = st.tabs(["最终报告", "详细信息", "下载"])
with tab1:
st.markdown(final_report)
with tab2:
# 段落详情
st.subheader("段落详情")
for i, paragraph in enumerate(agent.state.paragraphs):
with st.expander(f"段落 {i+1}: {paragraph.title}"):
st.write("**预期内容:**", paragraph.content)
st.write("**最终内容:**", paragraph.research.latest_summary[:300] + "..."
if len(paragraph.research.latest_summary) > 300
else paragraph.research.latest_summary)
st.write("**搜索次数:**", paragraph.research.get_search_count())
st.write("**反思次数:**", paragraph.research.reflection_iteration)
# 搜索历史
st.subheader("搜索历史")
all_searches = []
for paragraph in agent.state.paragraphs:
all_searches.extend(paragraph.research.search_history)
if all_searches:
for i, search in enumerate(all_searches):
with st.expander(f"搜索 {i+1}: {search.query}"):
st.write("**URL:**", search.url)
st.write("**标题:**", search.title)
st.write("**内容预览:**", search.content[:200] + "..." if len(search.content) > 200 else search.content)
if search.score:
st.write("**相关度评分:**", search.score)
with tab3:
# 下载选项
st.subheader("下载报告")
# Markdown下载
st.download_button(
label="下载Markdown报告",
data=final_report,
file_name=f"deep_search_report_{datetime.now().strftime('%Y%m%d_%H%M%S')}.md",
mime="text/markdown"
)
# JSON状态下载
state_json = agent.state.to_json()
st.download_button(
label="下载状态文件",
data=state_json,
file_name=f"deep_search_state_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json",
mime="application/json"
)
if __name__ == "__main__":
main()