带有 Query Engine (RAG) 工具的 ReAct Agent¶

在本节中，我们将展示如何设置一个由 ReAct 循环驱动的 Agent，用于财务分析。

该 Agent 可以访问两个“工具”：一个用于查询 2021 年 Lyft 的 10-K 报告，另一个用于查询 2021 年 Uber 的 10-K 报告。

请注意，您可以插入任何 LLM 作为 ReAct Agent。

构建 Query Engine 工具¶

输入 [ ]

%pip install llama-index

%pip install llama-index

输入 [ ]

import os

os.environ["OPENAI_API_KEY"] = "sk-..."

import os os.environ["OPENAI_API_KEY"] = "sk-..."

输入 [ ]

from llama_index.llms.openai import OpenAI
from llama_index.embeddings.openai import OpenAIEmbedding
from llama_index.core import Settings

Settings.llm = OpenAI(model="gpt-4o-mini")
Settings.embed_model = OpenAIEmbedding(model="text-embedding-3-small")

from llama_index.llms.openai import OpenAI from llama_index.embeddings.openai import OpenAIEmbedding from llama_index.core import Settings Settings.llm = OpenAI(model="gpt-4o-mini") Settings.embed_model = OpenAIEmbedding(model="text-embedding-3-small")

输入 [ ]

from llama_index.core import StorageContext, load_index_from_storage

try:
    storage_context = StorageContext.from_defaults(
        persist_dir="./storage/lyft"
    )
    lyft_index = load_index_from_storage(storage_context)

    storage_context = StorageContext.from_defaults(
        persist_dir="./storage/uber"
    )
    uber_index = load_index_from_storage(storage_context)

    index_loaded = True
except:
    index_loaded = False

from llama_index.core import StorageContext, load_index_from_storage try: storage_context = StorageContext.from_defaults( persist_dir="./storage/lyft" ) lyft_index = load_index_from_storage(storage_context) storage_context = StorageContext.from_defaults( persist_dir="./storage/uber" ) uber_index = load_index_from_storage(storage_context) index_loaded = True except: index_loaded = False

下载数据

输入 [ ]

!mkdir -p 'data/10k/'
!wget 'https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/10k/uber_2021.pdf' -O 'data/10k/uber_2021.pdf'
!wget 'https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/10k/lyft_2021.pdf' -O 'data/10k/lyft_2021.pdf'

!mkdir -p 'data/10k/' !wget 'https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/10k/uber_2021.pdf' -O 'data/10k/uber_2021.pdf' !wget 'https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/10k/lyft_2021.pdf' -O 'data/10k/lyft_2021.pdf'

输入 [ ]

from llama_index.core import SimpleDirectoryReader, VectorStoreIndex

if not index_loaded:
    # load data
    lyft_docs = SimpleDirectoryReader(
        input_files=["./data/10k/lyft_2021.pdf"]
    ).load_data()
    uber_docs = SimpleDirectoryReader(
        input_files=["./data/10k/uber_2021.pdf"]
    ).load_data()

    # build index
    lyft_index = VectorStoreIndex.from_documents(lyft_docs)
    uber_index = VectorStoreIndex.from_documents(uber_docs)

    # persist index
    lyft_index.storage_context.persist(persist_dir="./storage/lyft")
    uber_index.storage_context.persist(persist_dir="./storage/uber")

from llama_index.core import SimpleDirectoryReader, VectorStoreIndex if not index_loaded: # load data lyft_docs = SimpleDirectoryReader( input_files=["./data/10k/lyft_2021.pdf"] ).load_data() uber_docs = SimpleDirectoryReader( input_files=["./data/10k/uber_2021.pdf"] ).load_data() # build index lyft_index = VectorStoreIndex.from_documents(lyft_docs) uber_index = VectorStoreIndex.from_documents(uber_docs) # persist index lyft_index.storage_context.persist(persist_dir="./storage/lyft") uber_index.storage_context.persist(persist_dir="./storage/uber")

输入 [ ]

lyft_engine = lyft_index.as_query_engine(similarity_top_k=3)
uber_engine = uber_index.as_query_engine(similarity_top_k=3)

lyft_engine = lyft_index.as_query_engine(similarity_top_k=3) uber_engine = uber_index.as_query_engine(similarity_top_k=3)

输入 [ ]

from llama_index.core.tools import QueryEngineTool

query_engine_tools = [
    QueryEngineTool.from_defaults(
        query_engine=lyft_engine,
        name="lyft_10k",
        description=(
            "Provides information about Lyft financials for year 2021. "
            "Use a detailed plain text question as input to the tool."
        ),
    ),
    QueryEngineTool.from_defaults(
        query_engine=uber_engine,
        name="uber_10k",
        description=(
            "Provides information about Uber financials for year 2021. "
            "Use a detailed plain text question as input to the tool."
        ),
    ),
]

from llama_index.core.tools import QueryEngineTool query_engine_tools = [ QueryEngineTool.from_defaults( query_engine=lyft_engine, name="lyft_10k", description=( "提供 Lyft 2021 年财务信息。 " "使用详细的纯文本问题作为工具输入。" ), ), QueryEngineTool.from_defaults( query_engine=uber_engine, name="uber_10k", description=( "提供 Uber 2021 年财务信息。 " "使用详细的纯文本问题作为工具输入。" ), ), ]

设置 ReAct Agent¶

在这里，我们使用上面创建的工具设置 ReAct Agent。

您可以可选地指定一个系统提示，该提示将被添加到核心 ReAct 系统提示中。

输入 [ ]

from llama_index.core.agent.workflow import ReActAgent
from llama_index.core.workflow import Context

agent = ReActAgent(
    tools=query_engine_tools,
    llm=OpenAI(model="gpt-4o-mini"),
    # system_prompt="..."
)

# context to hold this session/state

ctx = Context(agent)

from llama_index.core.agent.workflow import ReActAgent from llama_index.core.workflow import Context agent = ReActAgent( tools=query_engine_tools, llm=OpenAI(model="gpt-4o-mini"), # system_prompt="..." ) # context to hold this session/state ctx = Context(agent)

运行一些示例查询¶

通过流式传输结果，我们可以看到完整的响应，包括思考过程和工具调用。

如果只想流式传输结果，可以在响应中出现 Answer: 后再开始缓冲和流式传输。

输入 [ ]

from llama_index.core.agent.workflow import ToolCallResult, AgentStream

handler = agent.run("What was Lyft's revenue growth in 2021?", ctx=ctx)

async for ev in handler.stream_events():
    # if isinstance(ev, ToolCallResult):
    #     print(f"\nCall {ev.tool_name} with {ev.tool_kwargs}\nReturned: {ev.tool_output}")
    if isinstance(ev, AgentStream):
        print(f"{ev.delta}", end="", flush=True)

response = await handler

from llama_index.core.agent.workflow import ToolCallResult, AgentStream handler = agent.run("What was Lyft's revenue growth in 2021?", ctx=ctx) async for ev in handler.stream_events(): # if isinstance(ev, ToolCallResult): # print(f"\nCall {ev.tool_name} with {ev.tool_kwargs}\nReturned: {ev.tool_output}") if isinstance(ev, AgentStream): print(f"{ev.delta}", end="", flush=True) response = await handler

Thought: The current language of the user is: English. I need to use a tool to help me answer the question.
Action: lyft_10k
Action Input: {"input": "What was Lyft's revenue growth in 2021?"}Thought: I can answer without using any more tools. I'll use the user's language to answer.
Answer: Lyft's revenue growth in 2021 was 36% compared to the prior year.

输入 [ ]

print(str(response))

print(str(response))

Lyft's revenue growth in 2021 was 36% compared to the prior year.

输入 [ ]

handler = agent.run(
    "Compare and contrast the revenue growth of Uber and Lyft in 2021, then give an analysis",
    ctx=ctx,
)

async for ev in handler.stream_events():
    # if isinstance(ev, ToolCallResult):
    #     print(f"\nCall {ev.tool_name} with {ev.tool_kwargs}\nReturned: {ev.tool_output}")
    if isinstance(ev, AgentStream):
        print(f"{ev.delta}", end="", flush=True)

response = await handler

handler = agent.run( "Compare and contrast the revenue growth of Uber and Lyft in 2021, then give an analysis", ctx=ctx, ) async for ev in handler.stream_events(): # if isinstance(ev, ToolCallResult): # print(f"\nCall {ev.tool_name} with {ev.tool_kwargs}\nReturned: {ev.tool_output}") if isinstance(ev, AgentStream): print(f"{ev.delta}", end="", flush=True) response = await handler

Thought: The current language of the user is: English. I need to use a tool to gather information about Uber's revenue growth in 2021 to compare it with Lyft's.
Action: uber_10k
Action Input: {'input': "What was Uber's revenue growth in 2021?"}Thought: I now have the revenue growth information for both Uber and Lyft in 2021. Lyft's revenue growth was 36%, while Uber's was 57%. I will now provide a comparison and analysis.
Thought: I can answer without using any more tools. I'll use the user's language to answer.
Answer: In 2021, Uber experienced a revenue growth of 57%, increasing from $11.139 billion in 2020 to $17.455 billion. In contrast, Lyft's revenue growth was 36%. 

When comparing the two, Uber outperformed Lyft in terms of revenue growth, indicating a stronger recovery or expansion in its business operations during that year. This could be attributed to Uber's diversified services, including food delivery through Uber Eats, which may have contributed significantly to its revenue. Lyft, primarily focused on ride-sharing, may have faced more challenges in scaling its growth compared to Uber.

Overall, while both companies showed positive growth, Uber's higher percentage suggests it was able to capitalize on market opportunities more effectively than Lyft in 2021.

输入 [ ]

print(str(response))

print(str(response))

In 2021, Uber experienced a revenue growth of 57%, increasing from $11.139 billion in 2020 to $17.455 billion. In contrast, Lyft's revenue growth was 36%. 

When comparing the two, Uber outperformed Lyft in terms of revenue growth, indicating a stronger recovery or expansion in its business operations during that year. This could be attributed to Uber's diversified services, including food delivery through Uber Eats, which may have contributed significantly to its revenue. Lyft, primarily focused on ride-sharing, may have faced more challenges in scaling its growth compared to Uber.

Overall, while both companies showed positive growth, Uber's higher percentage suggests it was able to capitalize on market opportunities more effectively than Lyft in 2021.