Build a Retrieval Augmented Generation (RAG) App: Part 1
One of the most powerful applications enabled by LLMs is sophisticated question-answering (Q&A) chatbots. These are applications that can answer questions about specific source information. These applications use a technique known as Retrieval Augmented Generation, or RAG.
This is a multi-part tutorial:
- Part 1 (this guide) introduces RAG and walks through a minimal implementation.
- Part 2 extends the implementation to accommodate conversation-style interactions and multi-step retrieval processes.
This tutorial will show how to build a simple Q&A application over a text data source. Along the way weβll go over a typical Q&A architecture and highlight additional resources for more advanced Q&A techniques. Weβll also see how LangSmith can help us trace and understand our application. LangSmith will become increasingly helpful as our application grows in complexity.
If you're already familiar with basic retrieval, you might also be interested in this high-level overview of different retrieval techinques.
Note: Here we focus on Q&A for unstructured data. If you are interested for RAG over structured data, check out our tutorial on doing question/answering over SQL data.
Overviewβ
A typical RAG application has two main components:
Indexing: a pipeline for ingesting data from a source and indexing it. This usually happens offline.
Retrieval and generation: the actual RAG chain, which takes the user query at run time and retrieves the relevant data from the index, then passes that to the model.
Note: the indexing portion of this tutorial will largely follow the semantic search tutorial.
The most common full sequence from raw data to answer looks like:
Indexingβ
- Load: First we need to load our data. This is done with Document Loaders.
- Split: Text splitters break large
Documents
into smaller chunks. This is useful both for indexing data and passing it into a model, as large chunks are harder to search over and won't fit in a model's finite context window. - Store: We need somewhere to store and index our splits, so that they can be searched over later. This is often done using a VectorStore and Embeddings model.
Retrieval and generationβ
- Retrieve: Given a user input, relevant splits are retrieved from storage using a Retriever.
- Generate: A ChatModel / LLM produces an answer using a prompt that includes both the question with the retrieved data
Once we've indexed our data, we will use LangGraph as our orchestration framework to implement the retrieval and generation steps.
Setupβ
Jupyter Notebookβ
This and other tutorials are perhaps most conveniently run in a Jupyter notebooks. Going through guides in an interactive environment is a great way to better understand them. See here for instructions on how to install.
Installationβ
This tutorial requires these langchain dependencies:
- Pip
- Conda
%pip install --quiet --upgrade langchain-text-splitters langchain-community langgraph
conda install langchain-text-splitters langchain-community langgraph -c conda-forge
For more details, see our Installation guide.
LangSmithβ
Many of the applications you build with LangChain will contain multiple steps with multiple invocations of LLM calls. As these applications get more complex, it becomes crucial to be able to inspect what exactly is going on inside your chain or agent. The best way to do this is with LangSmith.
After you sign up at the link above, make sure to set your environment variables to start logging traces:
export LANGCHAIN_TRACING_V2="true"
export LANGCHAIN_API_KEY="..."
Or, if in a notebook, you can set them with:
import getpass
import os
os.environ["LANGCHAIN_TRACING_V2"] = "true"
os.environ["LANGCHAIN_API_KEY"] = getpass.getpass()
Componentsβ
We will need to select three components from LangChain's suite of integrations.
pip install -qU langchain-openai
import getpass
import os
if not os.environ.get("OPENAI_API_KEY"):
os.environ["OPENAI_API_KEY"] = getpass.getpass("Enter API key for OpenAI: ")
from langchain_openai import ChatOpenAI
llm = ChatOpenAI(model="gpt-4o-mini")
pip install -qU langchain-openai
import getpass
import os
if not os.environ.get("OPENAI_API_KEY"):
os.environ["OPENAI_API_KEY"] = getpass.getpass("Enter API key for OpenAI: ")
from langchain_openai import OpenAIEmbeddings
embeddings = OpenAIEmbeddings(model="text-embedding-3-large")
pip install -qU langchain-core
from langchain_core.vectorstores import InMemoryVectorStore
vector_store = InMemoryVectorStore(embeddings)
Previewβ
In this guide weβll build an app that answers questions about the website's content. The specific website we will use is the LLM Powered Autonomous Agents blog post by Lilian Weng, which allows us to ask questions about the contents of the post.
We can create a simple indexing pipeline and RAG chain to do this in ~50 lines of code.
import bs4
from langchain import hub
from langchain_community.document_loaders import WebBaseLoader
from langchain_core.documents import Document
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langgraph.graph import START, StateGraph
from typing_extensions import List, TypedDict
# Load and chunk contents of the blog
loader = WebBaseLoader(
web_paths=("https://lilianweng.github.io/posts/2023-06-23-agent/",),
bs_kwargs=dict(
parse_only=bs4.SoupStrainer(
class_=("post-content", "post-title", "post-header")
)
),
)
docs = loader.load()
text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
all_splits = text_splitter.split_documents(docs)
# Index chunks
_ = vector_store.add_documents(documents=all_splits)
# Define prompt for question-answering
prompt = hub.pull("rlm/rag-prompt")
# Define state for application
class State(TypedDict):
question: str
context: List[Document]
answer: str
# Define application steps
def retrieve(state: State):
retrieved_docs = vector_store.similarity_search(state["question"])
return {"context": retrieved_docs}
def generate(state: State):
docs_content = "\n\n".join(doc.page_content for doc in state["context"])
messages = prompt.invoke({"question": state["question"], "context": docs_content})
response = llm.invoke(messages)
return {"answer": response.content}
# Compile application and test
graph_builder = StateGraph(State).add_sequence([retrieve, generate])
graph_builder.add_edge(START, "retrieve")
graph = graph_builder.compile()
response = graph.invoke({"question": "What is Task Decomposition?"})
print(response["answer"])
Task Decomposition is the process of breaking down a complicated task into smaller, manageable steps to facilitate easier execution and understanding. Techniques like Chain of Thought (CoT) and Tree of Thoughts (ToT) guide models to think step-by-step, allowing them to explore multiple reasoning possibilities. This method enhances performance on complex tasks and provides insight into the model's thinking process.
Check out the LangSmith trace.
Detailed walkthroughβ
Letβs go through the above code step-by-step to really understand whatβs going on.
1. Indexingβ
This section is an abbreviated version of the content in the semantic search tutorial. If you're comfortable with document loaders, embeddings, and vector stores, feel free to skip to the next section on retrieval and generation.
Loading documentsβ
We need to first load the blog post contents. We can use DocumentLoaders for this, which are objects that load in data from a source and return a list of Document objects.
In this case weβll use the
WebBaseLoader,
which uses urllib
to load HTML from web URLs and BeautifulSoup
to
parse it to text. We can customize the HTML -> text parsing by passing
in parameters into the BeautifulSoup
parser via bs_kwargs
(see
BeautifulSoup
docs).
In this case only HTML tags with class βpost-contentβ, βpost-titleβ, or
βpost-headerβ are relevant, so weβll remove all others.
import bs4
from langchain_community.document_loaders import WebBaseLoader
# Only keep post title, headers, and content from the full HTML.
bs4_strainer = bs4.SoupStrainer(class_=("post-title", "post-header", "post-content"))
loader = WebBaseLoader(
web_paths=("https://lilianweng.github.io/posts/2023-06-23-agent/",),
bs_kwargs={"parse_only": bs4_strainer},
)
docs = loader.load()
assert len(docs) == 1
print(f"Total characters: {len(docs[0].page_content)}")
Total characters: 43131
print(docs[0].page_content[:500])
LLM Powered Autonomous Agents
Date: June 23, 2023 | Estimated Reading Time: 31 min | Author: Lilian Weng
Building agents with LLM (large language model) as its core controller is a cool concept. Several proof-of-concepts demos, such as AutoGPT, GPT-Engineer and BabyAGI, serve as inspiring examples. The potentiality of LLM extends beyond generating well-written copies, stories, essays and programs; it can be framed as a powerful general problem solver.
Agent System Overview#
In
Go deeperβ
DocumentLoader
: Object that loads data from a source as list of Documents
.
- Docs:
Detailed documentation on how to use
DocumentLoaders
. - Integrations: 160+ integrations to choose from.
- Interface: API reference for the base interface.
Splitting documentsβ
Our loaded document is over 42k characters which is too long to fit into the context window of many models. Even for those models that could fit the full post in their context window, models can struggle to find information in very long inputs.
To handle this weβll split the Document
into chunks for embedding and
vector storage. This should help us retrieve only the most relevant parts
of the blog post at run time.
As in the semantic search tutorial, we use a RecursiveCharacterTextSplitter, which will recursively split the document using common separators like new lines until each chunk is the appropriate size. This is the recommended text splitter for generic text use cases.
from langchain_text_splitters import RecursiveCharacterTextSplitter
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=1000, # chunk size (characters)
chunk_overlap=200, # chunk overlap (characters)
add_start_index=True, # track index in original document
)
all_splits = text_splitter.split_documents(docs)
print(f"Split blog post into {len(all_splits)} sub-documents.")
Split blog post into 66 sub-documents.
Go deeperβ
TextSplitter
: Object that splits a list of Document
s into smaller
chunks. Subclass of DocumentTransformer
s.
- Learn more about splitting text using different methods by reading the how-to docs
- Code (py or js)
- Scientific papers
- Interface: API reference for the base interface.
DocumentTransformer
: Object that performs a transformation on a list
of Document
objects.
- Docs: Detailed documentation on how to use
DocumentTransformers
- Integrations
- Interface: API reference for the base interface.
Storing documentsβ
Now we need to index our 66 text chunks so that we can search over them at runtime. Following the semantic search tutorial, our approach is to embed the contents of each document split and insert these embeddings into a vector store. Given an input query, we can then use vector search to retrieve relevant documents.
We can embed and store all of our document splits in a single command using the vector store and embeddings model selected at the start of the tutorial.
document_ids = vector_store.add_documents(documents=all_splits)
print(document_ids[:3])
['07c18af6-ad58-479a-bfb1-d508033f9c64', '9000bf8e-1993-446f-8d4d-f4e507ba4b8f', 'ba3b5d14-bed9-4f5f-88be-44c88aedc2e6']
Go deeperβ
Embeddings
: Wrapper around a text embedding model, used for converting
text to embeddings.
- Docs: Detailed documentation on how to use embeddings.
- Integrations: 30+ integrations to choose from.
- Interface: API reference for the base interface.
VectorStore
: Wrapper around a vector database, used for storing and
querying embeddings.
- Docs: Detailed documentation on how to use vector stores.
- Integrations: 40+ integrations to choose from.
- Interface: API reference for the base interface.
This completes the Indexing portion of the pipeline. At this point we have a query-able vector store containing the chunked contents of our blog post. Given a user question, we should ideally be able to return the snippets of the blog post that answer the question.
2. Retrieval and Generationβ
Now letβs write the actual application logic. We want to create a simple application that takes a user question, searches for documents relevant to that question, passes the retrieved documents and initial question to a model, and returns an answer.
For generation, we will use the chat model selected at the start of the tutorial.
Weβll use a prompt for RAG that is checked into the LangChain prompt hub (here).
from langchain import hub
prompt = hub.pull("rlm/rag-prompt")
example_messages = prompt.invoke(
{"context": "(context goes here)", "question": "(question goes here)"}
).to_messages()
assert len(example_messages) == 1
print(example_messages[0].content)
You are an assistant for question-answering tasks. Use the following pieces of retrieved context to answer the question. If you don't know the answer, just say that you don't know. Use three sentences maximum and keep the answer concise.
Question: (question goes here)
Context: (context goes here)
Answer:
We'll use LangGraph to tie together the retrieval and generation steps into a single application. This will bring a number of benefits:
- We can define our application logic once and automatically support multiple invocation modes, including streaming, async, and batched calls.
- We get streamlined deployments via LangGraph Platform.
- LangSmith will automatically trace the steps of our application together.
- We can easily add key features to our application, including persistence and human-in-the-loop approval, with minimal code changes.
To use LangGraph, we need to define three things:
- The state of our application;
- The nodes of our application (i.e., application steps);
- The "control flow" of our application (e.g., the ordering of the steps).
State:β
The state of our application controls what data is input to the application, transferred between steps, and output by the application. It is typically a TypedDict
, but can also be a Pydantic BaseModel.
For a simple RAG application, we can just keep track of the input question, retrieved context, and generated answer:
from langchain_core.documents import Document
from typing_extensions import List, TypedDict
class State(TypedDict):
question: str
context: List[Document]
answer: str
Nodes (application steps)β
Let's start with a simple sequence of two steps: retrieval and generation.
def retrieve(state: State):
retrieved_docs = vector_store.similarity_search(state["question"])
return {"context": retrieved_docs}
def generate(state: State):
docs_content = "\n\n".join(doc.page_content for doc in state["context"])
messages = prompt.invoke({"question": state["question"], "context": docs_content})
response = llm.invoke(messages)
return {"answer": response.content}
Our retrieval step simply runs a similarity search using the input question, and the generation step formats the retrieved context and original question into a prompt for the chat model.
Control flowβ
Finally, we compile our application into a single graph
object. In this case, we are just connecting the retrieval and generation steps into a single sequence.
from langgraph.graph import START, StateGraph
graph_builder = StateGraph(State).add_sequence([retrieve, generate])
graph_builder.add_edge(START, "retrieve")
graph = graph_builder.compile()
LangGraph also comes with built-in utilities for visualizing the control flow of your application:
from IPython.display import Image, display
display(Image(graph.get_graph().draw_mermaid_png()))
Do I need to use LangGraph?
LangGraph is not required to build a RAG application. Indeed, we can implement the same application logic through invocations of the individual components:
question = "..."
retrieved_docs = vector_store.similarity_search(question)
docs_content = "\n\n".join(doc.page_content for doc in retrieved_docs)
prompt = prompt.invoke({"question": question, "context": docs_content})
answer = llm.invoke(prompt)
The benefits of LangGraph include:
- Support for multiple invocation modes: this logic would need to be rewritten if we wanted to stream output tokens, or stream the results of individual steps;
- Automatic support for tracing via LangSmith and deployments via LangGraph Platform;
- Support for persistence, human-in-the-loop, and other features.
Many use-cases demand RAG in a conversational experience, such that a user can receive context-informed answers via a stateful conversation. As we will see in Part 2 of the tutorial, LangGraph's management and persistence of state simplifies these applications enormously.
Usageβ
Let's test our application! LangGraph supports multiple invocation modes, including sync, async, and streaming.
Invoke:
result = graph.invoke({"question": "What is Task Decomposition?"})
print(f'Context: {result["context"]}\n\n')
print(f'Answer: {result["answer"]}')
Context: [Document(id='a42dc78b-8f76-472a-9e25-180508af74f3', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 1585}, 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 βthink step by stepβ 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 modelβs thinking process.'), Document(id='c0e45887-d0b0-483d-821a-bb5d8316d51d', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 2192}, page_content='Tree 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.'), Document(id='4cc7f318-35f5-440f-a4a4-145b5f0b918d', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 29630}, page_content='Resources:\n1. Internet access for searches and information gathering.\n2. Long Term memory management.\n3. GPT-3.5 powered Agents for delegation of simple tasks.\n4. File output.\n\nPerformance Evaluation:\n1. Continuously review and analyze your actions to ensure you are performing to the best of your abilities.\n2. Constructively self-criticize your big-picture behavior constantly.\n3. Reflect on past decisions and strategies to refine your approach.\n4. Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.'), Document(id='f621ade4-9b0d-471f-a522-44eb5feeba0c', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 19373}, page_content="(3) Task execution: Expert models execute on the specific tasks and log results.\nInstruction:\n\nWith the input and the inference results, the AI assistant needs to describe the process and results. The previous stages can be formed as - User Input: {{ User Input }}, Task Planning: {{ Tasks }}, Model Selection: {{ Model Assignment }}, Task Execution: {{ Predictions }}. You must first answer the user's request in a straightforward manner. Then describe the task process and show your analysis and model inference results to the user in the first person. If inference results contain a file path, must tell the user the complete file path.")]
Answer: Task decomposition is a technique used to break down complex tasks into smaller, manageable steps, allowing for more efficient problem-solving. This can be achieved through methods like chain of thought prompting or the tree of thoughts approach, which explores multiple reasoning possibilities at each step. It can be initiated through simple prompts, task-specific instructions, or human inputs.
Stream steps:
for step in graph.stream(
{"question": "What is Task Decomposition?"}, stream_mode="updates"
):
print(f"{step}\n\n----------------\n")
{'retrieve': {'context': [Document(id='a42dc78b-8f76-472a-9e25-180508af74f3', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 1585}, 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 βthink step by stepβ 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 modelβs thinking process.'), Document(id='c0e45887-d0b0-483d-821a-bb5d8316d51d', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 2192}, page_content='Tree 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.'), Document(id='4cc7f318-35f5-440f-a4a4-145b5f0b918d', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 29630}, page_content='Resources:\n1. Internet access for searches and information gathering.\n2. Long Term memory management.\n3. GPT-3.5 powered Agents for delegation of simple tasks.\n4. File output.\n\nPerformance Evaluation:\n1. Continuously review and analyze your actions to ensure you are performing to the best of your abilities.\n2. Constructively self-criticize your big-picture behavior constantly.\n3. Reflect on past decisions and strategies to refine your approach.\n4. Every command has a cost, so be smart and efficient. Aim to complete tasks in the least number of steps.'), Document(id='f621ade4-9b0d-471f-a522-44eb5feeba0c', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 19373}, page_content="(3) Task execution: Expert models execute on the specific tasks and log results.\nInstruction:\n\nWith the input and the inference results, the AI assistant needs to describe the process and results. The previous stages can be formed as - User Input: {{ User Input }}, Task Planning: {{ Tasks }}, Model Selection: {{ Model Assignment }}, Task Execution: {{ Predictions }}. You must first answer the user's request in a straightforward manner. Then describe the task process and show your analysis and model inference results to the user in the first person. If inference results contain a file path, must tell the user the complete file path.")]}}
----------------
{'generate': {'answer': 'Task decomposition is the process of breaking down a complex task into smaller, more manageable steps. This technique, often enhanced by methods like Chain of Thought (CoT) or Tree of Thoughts, allows models to reason through tasks systematically and improves performance by clarifying the thought process. It can be achieved through simple prompts, task-specific instructions, or human inputs.'}}
----------------
Stream tokens:
for message, metadata in graph.stream(
{"question": "What is Task Decomposition?"}, stream_mode="messages"
):
print(message.content, end="|")
|Task| decomposition| is| the| process| of| breaking| down| complex| tasks| into| smaller|,| more| manageable| steps|.| It| can| be| achieved| through| techniques| like| Chain| of| Thought| (|Co|T|)| prompting|,| which| encourages| the| model| to| think| step| by| step|,| or| through| more| structured| methods| like| the| Tree| of| Thoughts|.| This| approach| not| only| simplifies| task| execution| but| also| provides| insights| into| the| model|'s| reasoning| process|.||
For async invocations, use:
result = await graph.ainvoke(...)
and
async for step in graph.astream(...):
Returning sourcesβ
Note that by storing the retrieved context in the state of the graph, we recover sources for the model's generated answer in the "context"
field of the state. See this guide on returning sources for more detail.
Go deeperβ
Chat models take in a sequence of messages and return a message.
- Docs
- Integrations: 25+ integrations to choose from.
- Interface: API reference for the base interface.
Customizing the prompt
As shown above, we can load prompts (e.g., this RAG prompt) from the prompt hub. The prompt can also be easily customized. For example:
from langchain_core.prompts import PromptTemplate
template = """Use the following pieces of context to answer the question at the end.
If you don't know the answer, just say that you don't know, don't try to make up an answer.
Use three sentences maximum and keep the answer as concise as possible.
Always say "thanks for asking!" at the end of the answer.
{context}
Question: {question}
Helpful Answer:"""
custom_rag_prompt = PromptTemplate.from_template(template)
Query analysisβ
So far, we are executing the retrieval using the raw input query. However, there are some advantages to allowing a model to generate the query for retrieval purposes. For example:
- In addition to semantic search, we can build in structured filters (e.g., "Find documents since the year 2020.");
- The model can rewrite user queries, which may be multifaceted or include irrelevant language, into more effective search queries.
Query analysis employs models to transform or construct optimized search queries from raw user input. We can easily incorporate a query analysis step into our application. For illustrative purposes, let's add some metadata to the documents in our vector store. We will add some (contrived) sections to the document which we can filter on later.
total_documents = len(all_splits)
third = total_documents // 3
for i, document in enumerate(all_splits):
if i < third:
document.metadata["section"] = "beginning"
elif i < 2 * third:
document.metadata["section"] = "middle"
else:
document.metadata["section"] = "end"
all_splits[0].metadata
{'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/',
'start_index': 8,
'section': 'beginning'}
We will need to update the documents in our vector store. We will use a simple InMemoryVectorStore for this, as we will use some of its specific features (i.e., metadata filtering). Refer to the vector store integration documentation for relevant features of your chosen vector store.
from langchain_core.vectorstores import InMemoryVectorStore
vector_store = InMemoryVectorStore(embeddings)
_ = vector_store.add_documents(all_splits)
Let's next define a schema for our search query. We will use structured output for this purpose. Here we define a query as containing a string query and a document section (either "beginning", "middle", or "end"), but this can be defined however you like.
from typing import Literal
from typing_extensions import Annotated
class Search(TypedDict):
"""Search query."""
query: Annotated[str, ..., "Search query to run."]
section: Annotated[
Literal["beginning", "middle", "end"],
...,
"Section to query.",
]
Finally, we add a step to our LangGraph application to generate a query from the user's raw input:
class State(TypedDict):
question: str
query: Search
context: List[Document]
answer: str
def analyze_query(state: State):
structured_llm = llm.with_structured_output(Search)
query = structured_llm.invoke(state["question"])
return {"query": query}
def retrieve(state: State):
query = state["query"]
retrieved_docs = vector_store.similarity_search(
query["query"],
filter=lambda doc: doc.metadata.get("section") == query["section"],
)
return {"context": retrieved_docs}
def generate(state: State):
docs_content = "\n\n".join(doc.page_content for doc in state["context"])
messages = prompt.invoke({"question": state["question"], "context": docs_content})
response = llm.invoke(messages)
return {"answer": response.content}
graph_builder = StateGraph(State).add_sequence([analyze_query, retrieve, generate])
graph_builder.add_edge(START, "analyze_query")
graph = graph_builder.compile()
display(Image(graph.get_graph().draw_mermaid_png()))
We can test our implementation by specifically asking for context from the end of the post. Note that the model includes different information in its answer.
for step in graph.stream(
{"question": "What does the end of the post say about Task Decomposition?"},
stream_mode="updates",
):
print(f"{step}\n\n----------------\n")
{'analyze_query': {'query': {'query': 'Task Decomposition', 'section': 'end'}}}
----------------
{'retrieve': {'context': [Document(id='d6cef137-e1e8-4ddc-91dc-b62bd33c6020', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 39221, 'section': 'end'}, page_content='Finite context length: The restricted context capacity limits the inclusion of historical information, detailed instructions, API call context, and responses. The design of the system has to work with this limited communication bandwidth, while mechanisms like self-reflection to learn from past mistakes would benefit a lot from long or infinite context windows. Although vector stores and retrieval can provide access to a larger knowledge pool, their representation power is not as powerful as full attention.\n\n\nChallenges in long-term planning and task decomposition: Planning over a lengthy history and effectively exploring the solution space remain challenging. LLMs struggle to adjust plans when faced with unexpected errors, making them less robust compared to humans who learn from trial and error.'), Document(id='d1834ae1-eb6a-43d7-a023-08dfa5028799', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 39086, 'section': 'end'}, page_content='}\n]\nChallenges#\nAfter going through key ideas and demos of building LLM-centered agents, I start to see a couple common limitations:'), Document(id='ca7f06e4-2c2e-4788-9a81-2418d82213d9', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 32942, 'section': 'end'}, page_content='}\n]\nThen after these clarification, the agent moved into the code writing mode with a different system message.\nSystem message:'), Document(id='1fcc2736-30f4-4ef6-90f2-c64af92118cb', metadata={'source': 'https://lilianweng.github.io/posts/2023-06-23-agent/', 'start_index': 35127, 'section': 'end'}, page_content='"content": "You will get instructions for code to write.\\nYou will write a very long answer. Make sure that every detail of the architecture is, in the end, implemented as code.\\nMake sure that every detail of the architecture is, in the end, implemented as code.\\n\\nThink step by step and reason yourself to the right decisions to make sure we get it right.\\nYou will first lay out the names of the core classes, functions, methods that will be necessary, as well as a quick comment on their purpose.\\n\\nThen you will output the content of each file including ALL code.\\nEach file must strictly follow a markdown code block format, where the following tokens must be replaced such that\\nFILENAME is the lowercase file name including the file extension,\\nLANG is the markup code block language for the code\'s language, and CODE is the code:\\n\\nFILENAME\\n\`\`\`LANG\\nCODE\\n\`\`\`\\n\\nYou will start with the \\"entrypoint\\" file, then go to the ones that are imported by that file, and so on.\\nPlease')]}}
----------------
{'generate': {'answer': 'The end of the post highlights that task decomposition faces challenges in long-term planning and adapting to unexpected errors. LLMs struggle with adjusting their plans, making them less robust compared to humans who learn from trial and error. This indicates a limitation in effectively exploring the solution space and handling complex tasks.'}}
----------------
In both the streamed steps and the LangSmith trace, we can now observe the structured query that was fed into the retrieval step.
Query Analysis is a rich problem with a wide range of approaches. Refer to the how-to guides for more examples.
Next stepsβ
We've covered the steps to build a basic Q&A app over data:
- Loading data with a Document Loader
- Chunking the indexed data with a Text Splitter to make it more easily usable by a model
- Embedding the data and storing the data in a vectorstore
- Retrieving the previously stored chunks in response to incoming questions
- Generating an answer using the retrieved chunks as context.
In Part 2 of the tutorial, we will extend the implementation here to accommodate conversation-style interactions and multi-step retrieval processes.
Further reading:
- Return sources: Learn how to return source documents
- Streaming: Learn how to stream outputs and intermediate steps
- Add chat history: Learn how to add chat history to your app
- Retrieval conceptual guide: A high-level overview of specific retrieval techniques