Revision of requirements.txt

openhands 已修改 1730886503. 還原成這個修訂版本

2 files changed, 162 insertions

advanced_langchain_demo.py(檔案已創建)

		@@ -0,0 +1,152 @@
1	+	"""
2	+	Advanced LangChain Demo
3	+
4	+	This script demonstrates advanced LangChain features including:
5	+	1. Agents with multiple tools
6	+	2. Memory and conversation management
7	+	3. Custom tools and chains
8	+	4. Structured output parsing
9	+	5. Sequential chains
10	+	"""
11	+
12	+	import os
13	+	from typing import List, Dict, Union
14	+	import re
15	+	from datetime import datetime
16	+
17	+	from rich.console import Console
18	+	from rich.panel import Panel
19	+	from rich.table import Table
20	+	from rich import print as rprint
21	+
22	+	from langchain.agents import load_tools, initialize_agent, Tool, AgentExecutor, AgentType
23	+	from langchain.memory import ConversationBufferMemory
24	+	from langchain.chat_models import ChatOpenAI
25	+	from langchain.chains import LLMChain, SimpleSequentialChain
26	+	from langchain.prompts import PromptTemplate, ChatPromptTemplate, HumanMessagePromptTemplate
27	+	from langchain.schema import HumanMessage, AgentAction, AgentFinish
28	+	from langchain.tools import DuckDuckGoSearchRun
29	+	from langchain.callbacks import get_openai_callback
30	+
31	+	# Initialize Rich console for pretty output
32	+	console = Console()
33	+
34	+	def create_news_analyzer_agent():
35	+	"""
36	+	Creates an agent that can search for news and analyze them.
37	+	Demonstrates tool integration and agent creation.
38	+	"""
39	+	# Initialize the language model
40	+	llm = ChatOpenAI(temperature=0, model="gpt-3.5-turbo")
41	+
42	+	# Initialize tools
43	+	search = DuckDuckGoSearchRun()
44	+
45	+	tools = [
46	+	Tool(
47	+	name="Search",
48	+	func=search.run,
49	+	description="Useful for searching recent news and information"
50	+	)
51	+	]
52	+
53	+	# Initialize memory
54	+	memory = ConversationBufferMemory(
55	+	memory_key="chat_history",
56	+	return_messages=True
57	+	)
58	+
59	+	# Create the agent
60	+	agent = initialize_agent(
61	+	tools,
62	+	llm,
63	+	agent=AgentType.CHAT_CONVERSATIONAL_REACT_DESCRIPTION,
64	+	memory=memory,
65	+	verbose=True
66	+	)
67	+
68	+	return agent
69	+
70	+	def create_analysis_chain():
71	+	"""
72	+	Creates a sequential chain for analyzing topics.
73	+	Demonstrates chain composition and prompt management.
74	+	"""
75	+	llm = ChatOpenAI(temperature=0, model="gpt-3.5-turbo")
76	+
77	+	# First chain: Generate main points
78	+	main_points_prompt = PromptTemplate(
79	+	input_variables=["topic"],
80	+	template="What are the 3 main points to understand about {topic}?"
81	+	)
82	+	main_points_chain = LLMChain(llm=llm, prompt=main_points_prompt)
83	+
84	+	# Second chain: Generate implications
85	+	implications_prompt = PromptTemplate(
86	+	input_variables=["points"],
87	+	template="Given these points: {points}
88	+	What are the potential future implications?"
89	+	)
90	+	implications_chain = LLMChain(llm=llm, prompt=implications_prompt)
91	+
92	+	# Combine chains
93	+	overall_chain = SimpleSequentialChain(
94	+	chains=[main_points_chain, implications_chain],
95	+	verbose=True
96	+	)
97	+
98	+	return overall_chain
99	+
100	+	def display_results(title: str, content: str):
101	+	"""Pretty print results using Rich."""
102	+	console.print()
103	+	console.print(Panel.fit(
104	+	f"🔍 {title}",
105	+	style="bold blue"
106	+	))
107	+	console.print()
108	+	console.print(content)
109	+	console.print()
110	+
111	+	def main():
112	+	"""
113	+	Main function demonstrating various LangChain features.
114	+	"""
115	+	console.print(Panel.fit(
116	+	"🚀 Advanced LangChain Features Demo",
117	+	subtitle="Demonstrating Agents, Chains, and Tools",
118	+	style="bold green"
119	+	))
120	+
121	+	# Initialize components
122	+	news_agent = create_news_analyzer_agent()
123	+	analysis_chain = create_analysis_chain()
124	+
125	+	# Track token usage
126	+	with get_openai_callback() as cb:
127	+	# 1. Use agent to search and analyze news
128	+	console.print("
129	+	[bold]🔍 Using Agent to Search and Analyze News[/bold]")
130	+	news_query = "What are the latest developments in AI regulation?"
131	+	agent_result = news_agent.run(news_query)
132	+	display_results("Agent Analysis Result", agent_result)
133	+
134	+	# 2. Use sequential chain for deeper analysis
135	+	console.print("
136	+	[bold]📊 Using Sequential Chain for Analysis[/bold]")
137	+	chain_result = analysis_chain.run("AI regulation and its impact on technology companies")
138	+	display_results("Sequential Chain Analysis", chain_result)
139	+
140	+	# Display token usage
141	+	console.print(Panel.fit(
142	+	f"""[bold]📊 Token Usage Statistics[/bold]
143	+
144	+	Total Tokens: {cb.total_tokens}
145	+	Prompt Tokens: {cb.prompt_tokens}
146	+	Completion Tokens: {cb.completion_tokens}
147	+	Total Cost (USD): ${cb.total_cost:.4f}""",
148	+	style="bold yellow"
149	+	))
150	+
151	+	if __name__ == "__main__":
152	+	main()

requirements.txt(檔案已創建)

		@@ -0,0 +1,10 @@
1	+	langchain==0.0.340
2	+	openai==1.3.7
3	+	python-dotenv==1.0.0
4	+	requests==2.31.0
5	+	rich==13.7.0
6	+	beautifulsoup4==4.12.2
7	+	wikipedia==1.4.0
8	+	google-search-results==2.4.2
9	+	duckduckgo-search==3.9.9
10	+	tiktoken==0.5.1