Génération améliorée par récupération (RAG) avec Milvus et LangChain

Ce guide montre comment construire un système de génération améliorée par récupération (RAG) en utilisant LangChain et Milvus.

Le système RAG combine un système de recherche avec un modèle génératif pour générer un nouveau texte basé sur une invite donnée. Le système récupère d'abord les documents pertinents d'un corpus à l'aide de Milvus, puis utilise un modèle génératif pour générer un nouveau texte basé sur les documents récupérés.

LangChain est un cadre de développement d'applications alimentées par de grands modèles de langage (LLM). Milvus est la base de données vectorielles open-source la plus avancée au monde, conçue pour alimenter la recherche de similarité d'intégration et les applications d'intelligence artificielle.

Conditions préalables

Avant d'exécuter ce bloc-notes, assurez-vous que les dépendances suivantes sont installées :

$ pip install --upgrade --quiet  langchain langchain-core langchain-community langchain-text-splitters langchain-milvus langchain-openai bs4

Nous utiliserons les modèles d'OpenAI. Vous devez préparer la clé api OPENAI_API_KEY comme variable d'environnement.

import os

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

Préparer les données

Nous utilisons le Langchain WebBaseLoader pour charger des documents à partir de sources web et les découper en morceaux en utilisant le RecursiveCharacterTextSplitter.

import bs4
from langchain_community.document_loaders import WebBaseLoader
from langchain_text_splitters import RecursiveCharacterTextSplitter

# Create a WebBaseLoader instance to load documents from web sources
loader = WebBaseLoader(
    web_paths=(
        "https://lilianweng.github.io/posts/2023-06-23-agent/",
        "https://lilianweng.github.io/posts/2023-03-15-prompt-engineering/",
    ),
    bs_kwargs=dict(
        parse_only=bs4.SoupStrainer(
            class_=("post-content", "post-title", "post-header")
        )
    ),
)
# Load documents from web sources using the loader
documents = loader.load()
# Initialize a RecursiveCharacterTextSplitter for splitting text into chunks
text_splitter = RecursiveCharacterTextSplitter(chunk_size=2000, chunk_overlap=200)

# Split the documents into chunks using the text_splitter
docs = text_splitter.split_documents(documents)

# Let's take a look at the first document
docs[1]

Document(page_content='Fig. 1. Overview of a LLM-powered autonomous agent system.\nComponent One: Planning#\nA complicated task usually involves many steps. An agent needs to know what they are and plan ahead.\nTask Decomposition#\nChain of thought (CoT; Wei et al. 2022) has become a standard prompting technique for enhancing model performance on complex tasks. The model is instructed to “think step by step” to utilize more test-time computation to decompose hard tasks into smaller and simpler steps. CoT transforms big tasks into multiple manageable tasks and shed lights into an interpretation of the model’s thinking process.\nTree of Thoughts (Yao et al. 2023) extends CoT by exploring multiple reasoning possibilities at each step. It first decomposes the problem into multiple thought steps and generates multiple thoughts per step, creating a tree structure. The search process can be BFS (breadth-first search) or DFS (depth-first search) with each state evaluated by a classifier (via a prompt) or majority vote.\nTask decomposition can be done (1) by LLM with simple prompting like "Steps for XYZ.\\n1.", "What are the subgoals for achieving XYZ?", (2) by using task-specific instructions; e.g. "Write a story outline." for writing a novel, or (3) with human inputs.\nAnother quite distinct approach, LLM+P (Liu et al. 2023), involves relying on an external classical planner to do long-horizon planning. This approach utilizes the Planning Domain Definition Language (PDDL) as an intermediate interface to describe the planning problem. In this process, LLM (1) translates the problem into “Problem PDDL”, then (2) requests a classical planner to generate a PDDL plan based on an existing “Domain PDDL”, and finally (3) translates the PDDL plan back into natural language. Essentially, the planning step is outsourced to an external tool, assuming the availability of domain-specific PDDL and a suitable planner which is common in certain robotic setups but not in many other domains.\nSelf-Reflection#', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/'})

Comme nous pouvons le voir, le document est déjà divisé en morceaux. Et le contenu des données concerne l'agent d'intelligence artificielle.

Construire une chaîne RAG avec le Milvus Vector Store

Nous allons initialiser un magasin de vecteurs Milvus avec les documents, puis charger les documents dans le magasin de vecteurs Milvus et construire un index sous le capot.

from langchain_milvus import Milvus, Zilliz
from langchain_openai import OpenAIEmbeddings

embeddings = OpenAIEmbeddings()

vectorstore = Milvus.from_documents(  # or Zilliz.from_documents
    documents=docs,
    embedding=embeddings,
    connection_args={
        "uri": "./milvus_demo.db",
    },
    drop_old=True,  # Drop the old Milvus collection if it exists
)

Recherchez les documents dans le magasin vectoriel Milvus à l'aide d'une question test. Examinons le premier document.

query = "What is self-reflection of an AI Agent?"
vectorstore.similarity_search(query, k=1)

[Document(page_content='Self-Reflection#\nSelf-reflection is a vital aspect that allows autonomous agents to improve iteratively by refining past action decisions and correcting previous mistakes. It plays a crucial role in real-world tasks where trial and error are inevitable.\nReAct (Yao et al. 2023) integrates reasoning and acting within LLM by extending the action space to be a combination of task-specific discrete actions and the language space. The former enables LLM to interact with the environment (e.g. use Wikipedia search API), while the latter prompting LLM to generate reasoning traces in natural language.\nThe ReAct prompt template incorporates explicit steps for LLM to think, roughly formatted as:\nThought: ...\nAction: ...\nObservation: ...\n... (Repeated many times)', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'pk': 449281835035555859})]

from langchain_core.runnables import RunnablePassthrough
from langchain_core.prompts import PromptTemplate
from langchain_core.output_parsers import StrOutputParser
from langchain_openai import ChatOpenAI

# Initialize the OpenAI language model for response generation
llm = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=0)

# Define the prompt template for generating AI responses
PROMPT_TEMPLATE = """
Human: You are an AI assistant, and provides answers to questions by using fact based and statistical information when possible.
Use the following pieces of information to provide a concise answer to the question enclosed in <question> tags.
If you don't know the answer, just say that you don't know, don't try to make up an answer.
<context>
{context}
</context>

<question>
{question}
</question>

The response should be specific and use statistics or numbers when possible.

Assistant:"""

# Create a PromptTemplate instance with the defined template and input variables
prompt = PromptTemplate(
    template=PROMPT_TEMPLATE, input_variables=["context", "question"]
)
# Convert the vector store to a retriever
retriever = vectorstore.as_retriever()


# Define a function to format the retrieved documents
def format_docs(docs):
    return "\n\n".join(doc.page_content for doc in docs)

Utilisez le LCEL (LangChain Expression Language) pour construire une chaîne RAG.

# Define the RAG (Retrieval-Augmented Generation) chain for AI response generation
rag_chain = (
    {"context": retriever | format_docs, "question": RunnablePassthrough()}
    | prompt
    | llm
    | StrOutputParser()
)

# rag_chain.get_graph().print_ascii()

# Invoke the RAG chain with a specific question and retrieve the response
res = rag_chain.invoke(query)
res

"Self-reflection of an AI agent involves the process of synthesizing memories into higher-level inferences over time to guide the agent's future behavior. It serves as a mechanism to create higher-level summaries of past events. One approach to self-reflection involves prompting the language model with the 100 most recent observations and asking it to generate the 3 most salient high-level questions based on those observations. This process helps the AI agent optimize believability in the current moment and over time."

Félicitations ! Vous avez construit une chaîne RAG de base à l'aide de Milvus et de LangChain.

Filtrage des métadonnées

Nous pouvons utiliser les règles de filtrage scalaire de Milvus pour filtrer les documents en fonction des métadonnées. Nous avons chargé les documents à partir de deux sources différentes et nous pouvons filtrer les documents en fonction des métadonnées source.

vectorstore.similarity_search(
    "What is CoT?",
    k=1,
    expr="source == 'https://lilianweng.github.io/posts/2023-06-23-agent/'",
)

# The same as:
# vectorstore.as_retriever(search_kwargs=dict(
#     k=1,
#     expr="source == 'https://lilianweng.github.io/posts/2023-06-23-agent/'",
# )).invoke("What is CoT?")

[Document(page_content='Fig. 1. Overview of a LLM-powered autonomous agent system.\nComponent One: Planning#\nA complicated task usually involves many steps. An agent needs to know what they are and plan ahead.\nTask Decomposition#\nChain of thought (CoT; Wei et al. 2022) has become a standard prompting technique for enhancing model performance on complex tasks. The model is instructed to “think step by step” to utilize more test-time computation to decompose hard tasks into smaller and simpler steps. CoT transforms big tasks into multiple manageable tasks and shed lights into an interpretation of the model’s thinking process.\nTree of Thoughts (Yao et al. 2023) extends CoT by exploring multiple reasoning possibilities at each step. It first decomposes the problem into multiple thought steps and generates multiple thoughts per step, creating a tree structure. The search process can be BFS (breadth-first search) or DFS (depth-first search) with each state evaluated by a classifier (via a prompt) or majority vote.\nTask decomposition can be done (1) by LLM with simple prompting like "Steps for XYZ.\\n1.", "What are the subgoals for achieving XYZ?", (2) by using task-specific instructions; e.g. "Write a story outline." for writing a novel, or (3) with human inputs.\nAnother quite distinct approach, LLM+P (Liu et al. 2023), involves relying on an external classical planner to do long-horizon planning. This approach utilizes the Planning Domain Definition Language (PDDL) as an intermediate interface to describe the planning problem. In this process, LLM (1) translates the problem into “Problem PDDL”, then (2) requests a classical planner to generate a PDDL plan based on an existing “Domain PDDL”, and finally (3) translates the PDDL plan back into natural language. Essentially, the planning step is outsourced to an external tool, assuming the availability of domain-specific PDDL and a suitable planner which is common in certain robotic setups but not in many other domains.\nSelf-Reflection#', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'pk': 449281835035555858})]

Si nous voulons modifier dynamiquement les paramètres de recherche sans reconstruire la chaîne, nous pouvons configurer les internes de la chaîne d'exécution. Définissons un nouveau récupérateur avec cette configuration dynamique et utilisons-le pour construire une nouvelle chaîne RAG.

from langchain_core.runnables import ConfigurableField

# Define a new retriever with a configurable field for search_kwargs
retriever2 = vectorstore.as_retriever().configurable_fields(
    search_kwargs=ConfigurableField(
        id="retriever_search_kwargs",
    )
)

# Invoke the retriever with a specific search_kwargs which filter the documents by source
retriever2.with_config(
    configurable={
        "retriever_search_kwargs": dict(
            expr="source == 'https://lilianweng.github.io/posts/2023-06-23-agent/'",
            k=1,
        )
    }
).invoke(query)

[Document(page_content='Self-Reflection#\nSelf-reflection is a vital aspect that allows autonomous agents to improve iteratively by refining past action decisions and correcting previous mistakes. It plays a crucial role in real-world tasks where trial and error are inevitable.\nReAct (Yao et al. 2023) integrates reasoning and acting within LLM by extending the action space to be a combination of task-specific discrete actions and the language space. The former enables LLM to interact with the environment (e.g. use Wikipedia search API), while the latter prompting LLM to generate reasoning traces in natural language.\nThe ReAct prompt template incorporates explicit steps for LLM to think, roughly formatted as:\nThought: ...\nAction: ...\nObservation: ...\n... (Repeated many times)', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'pk': 449281835035555859})]

# Define a new RAG chain with this dynamically configurable retriever
rag_chain2 = (
    {"context": retriever2 | format_docs, "question": RunnablePassthrough()}
    | prompt
    | llm
    | StrOutputParser()
)

Essayons cette chaîne RAG configurable dynamiquement avec différentes conditions de filtrage.

# Invoke this RAG chain with a specific question and config
rag_chain2.with_config(
    configurable={
        "retriever_search_kwargs": dict(
            expr="source == 'https://lilianweng.github.io/posts/2023-06-23-agent/'",
        )
    }
).invoke(query)

"Self-reflection of an AI agent involves the process of synthesizing memories into higher-level inferences over time to guide the agent's future behavior. It serves as a mechanism to create higher-level summaries of past events. One approach to self-reflection involves prompting the language model with the 100 most recent observations and asking it to generate the 3 most salient high-level questions based on those observations. This process helps the AI agent optimize believability in the current moment and over time."

Lorsque nous modifions la condition de recherche pour filtrer les documents par la deuxième source, comme le contenu de ce blog n'a rien à voir avec la question posée, nous obtenons une réponse sans aucune information pertinente.

rag_chain2.with_config(
    configurable={
        "retriever_search_kwargs": dict(
            expr="source == 'https://lilianweng.github.io/posts/2023-03-15-prompt-engineering/'",
        )
    }
).invoke(query)

"I'm sorry, but based on the provided context, there is no specific information or statistical data available regarding the self-reflection of an AI agent."

---

Ce tutoriel se concentre sur l'utilisation de base de l'intégration de Milvus LangChain et sur l'approche RAG simple. Pour des techniques RAG plus avancées, veuillez vous référer au bootcamp RAG avancé.