■ RunnableBinding 클래스의 astream 메소드를 사용해 대화형 RAG 체인에서 검색기 전달전 사용자 질문을 재구성하는 방법을 보여준다.
※ 재구성된 질문은 최종 출력의 일부로 반환되지 않는다.
※ 중간 출력을 스트리밍하는 경우 astream_events 메소드를 사용하는 것이 좋다.
※ OPENAI_API_KEY 환경 변수 값은 .env 파일에 정의한다.
▶ main.py
|
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 |
import asyncio import bs4 from dotenv import load_dotenv from langchain_community.document_loaders import WebBaseLoader from langchain_text_splitters import RecursiveCharacterTextSplitter from langchain_chroma import Chroma from langchain_openai import OpenAIEmbeddings from langchain_openai import ChatOpenAI from langchain_core.prompts import ChatPromptTemplate from langchain_core.prompts import MessagesPlaceholder from langchain.chains.combine_documents import create_stuff_documents_chain from langchain.chains import create_retrieval_chain from langchain.chains import create_history_aware_retriever load_dotenv() webBaseLoader = 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"))) ) documentList = webBaseLoader.load() recursiveCharacterTextSplitter = RecursiveCharacterTextSplitter(chunk_size = 1000, chunk_overlap = 200) splitDocumentList = recursiveCharacterTextSplitter.split_documents(documentList) chroma = Chroma.from_documents(documents = splitDocumentList, embedding = OpenAIEmbeddings()) vectorStoreRetriever = chroma.as_retriever() chatOpenAI = ChatOpenAI(model = "gpt-4o-mini") # 질문을 맥락화한다. (질문을 상황에 맞게 이해하고 설명한다.) contextualizeQuestionChatPromptTemplate = ChatPromptTemplate.from_messages( [ ("system", "Given a chat history and the latest user question which might reference context in the chat history, formulate a standalone question which can be understood without the chat history. Do NOT answer the question, just reformulate it if needed and otherwise return it as is."), MessagesPlaceholder("chat_history"), ("human", "{input}"), ] ) contextualizeQuestionChatOpenAI = chatOpenAI.with_config(tags = ["contextualize_question"]) historyAwareRetrieverRunnableBinding = create_history_aware_retriever(contextualizeQuestionChatOpenAI, vectorStoreRetriever, contextualizeQuestionChatPromptTemplate) # 질문에 답변한다. questionAnswerChatPromptTemplate = ChatPromptTemplate.from_messages( [ ("system", "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, say that you don't know. Use three sentences maximum and keep the answer concise.\n\n{context}"), MessagesPlaceholder("chat_history"), ("human", "{input}"), ] ) questionAnswerRunnableBinding = create_stuff_documents_chain(chatOpenAI, questionAnswerChatPromptTemplate) # create_retrieval_chain 함수로 구성된 체인은 "input", "context", "answer" 키를 가진 dict 객체를 반환한다. # 여기서 "answer" 키만 토큰 단위로 스트리밍되고, 검색과 같은 다른 구성 요소는 토큰 수준 스트리밍을 지원하지 않는다. retrievalChainRunnableBinding = create_retrieval_chain(historyAwareRetrieverRunnableBinding, questionAnswerRunnableBinding) async def main(): async for addableDict in retrievalChainRunnableBinding.astream( { "input" : "What are some common ways of doing it?", "chat_history" : [ ("human", "What is task decomposition?"), ("ai" , "Task decomposition involves breaking up a complex task into smaller and simpler steps.") ] } ): if answer := addableDict.get("answer"): print(f"{answer}", end = "", flush = True) asyncio.run(main()) """ Common ways of task decomposition include using simple prompting techniques like "Steps for XYZ," task-specific instructions such as "Write a story outline," or incorporating human inputs. Additionally, methods like Chain of Thought (CoT) and Tree of Thoughts can be employed to enhance reasoning and planning. """ |
▶ requirements.txt
|
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 |
aiohappyeyeballs==2.4.4 aiohttp==3.11.9 aiosignal==1.3.1 annotated-types==0.7.0 anyio==4.6.2.post1 asgiref==3.8.1 attrs==24.2.0 backoff==2.2.1 bcrypt==4.2.1 beautifulsoup4==4.12.3 bs4==0.0.2 build==1.2.2.post1 cachetools==5.5.0 certifi==2024.8.30 charset-normalizer==3.4.0 chroma-hnswlib==0.7.6 chromadb==0.5.20 click==8.1.7 colorama==0.4.6 coloredlogs==15.0.1 dataclasses-json==0.6.7 Deprecated==1.2.15 distro==1.9.0 durationpy==0.9 fastapi==0.115.5 filelock==3.16.1 flatbuffers==24.3.25 frozenlist==1.5.0 fsspec==2024.10.0 google-auth==2.36.0 googleapis-common-protos==1.66.0 greenlet==3.1.1 grpcio==1.68.1 h11==0.14.0 httpcore==1.0.7 httptools==0.6.4 httpx==0.28.0 httpx-sse==0.4.0 huggingface-hub==0.26.3 humanfriendly==10.0 idna==3.10 importlib_metadata==8.5.0 importlib_resources==6.4.5 jiter==0.8.0 jsonpatch==1.33 jsonpointer==3.0.0 kubernetes==31.0.0 langchain==0.3.9 langchain-chroma==0.1.4 langchain-community==0.3.8 langchain-core==0.3.21 langchain-openai==0.2.10 langchain-text-splitters==0.3.2 langsmith==0.1.147 markdown-it-py==3.0.0 marshmallow==3.23.1 mdurl==0.1.2 mmh3==5.0.1 monotonic==1.6 mpmath==1.3.0 multidict==6.1.0 mypy-extensions==1.0.0 numpy==1.26.4 oauthlib==3.2.2 onnxruntime==1.20.1 openai==1.55.3 opentelemetry-api==1.28.2 opentelemetry-exporter-otlp-proto-common==1.28.2 opentelemetry-exporter-otlp-proto-grpc==1.28.2 opentelemetry-instrumentation==0.49b2 opentelemetry-instrumentation-asgi==0.49b2 opentelemetry-instrumentation-fastapi==0.49b2 opentelemetry-proto==1.28.2 opentelemetry-sdk==1.28.2 opentelemetry-semantic-conventions==0.49b2 opentelemetry-util-http==0.49b2 orjson==3.10.12 overrides==7.7.0 packaging==24.2 posthog==3.7.4 propcache==0.2.1 protobuf==5.29.0 pyasn1==0.6.1 pyasn1_modules==0.4.1 pydantic==2.10.2 pydantic-settings==2.6.1 pydantic_core==2.27.1 Pygments==2.18.0 PyPika==0.48.9 pyproject_hooks==1.2.0 pyreadline3==3.5.4 python-dateutil==2.9.0.post0 python-dotenv==1.0.1 PyYAML==6.0.2 regex==2024.11.6 requests==2.32.3 requests-oauthlib==2.0.0 requests-toolbelt==1.0.0 rich==13.9.4 rsa==4.9 shellingham==1.5.4 six==1.16.0 sniffio==1.3.1 soupsieve==2.6 SQLAlchemy==2.0.35 starlette==0.41.3 sympy==1.13.3 tenacity==9.0.0 tiktoken==0.8.0 tokenizers==0.21.0 tqdm==4.67.1 typer==0.14.0 typing-inspect==0.9.0 typing_extensions==4.12.2 urllib3==2.2.3 uvicorn==0.32.1 watchfiles==1.0.0 websocket-client==1.8.0 websockets==14.1 wrapt==1.17.0 yarl==1.18.3 zipp==3.21.0 |
※ pip install python-dotenv langchain langchain-community langchain-chroma langchain-openai bs4 명령을 실행했다.