mirror of
https://github.com/langchain-ai/langgraph.git
synced 2026-08-29 19:29:43 +02:00
235 KiB
235 KiB
Adaptive RAG
Adaptive RAG is a strategy for RAG that unites (1) query analysis with (2) active / self-corrective RAG.
In the paper, they report query analysis to route across:
- No Retrieval
- Single-shot RAG
- Iterative RAG
Let's build on this using LangGraph.
In our implementation, we will route between:
- Web search: for questions related to recent events
- Self-corrective RAG: for questions related to our index
In [4]:
! pip install -U langchain_community tiktoken langchain-openai langchain-cohere langchainhub chromadb langchain langgraph tavily-pythonRequirement already satisfied: langchain_community in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.0.27) Collecting langchain_community Downloading langchain_community-0.0.31-py3-none-any.whl.metadata (8.4 kB) Requirement already satisfied: tiktoken in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.5.2) Collecting tiktoken Downloading tiktoken-0.6.0-cp311-cp311-macosx_11_0_arm64.whl.metadata (6.6 kB) Requirement already satisfied: langchain-openai in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.0.2.post1) Collecting langchain-openai Downloading langchain_openai-0.1.1-py3-none-any.whl.metadata (2.5 kB) Requirement already satisfied: langchain-cohere in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.1.0) Requirement already satisfied: langchainhub in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.1.15) Requirement already satisfied: chromadb in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.4.24) Requirement already satisfied: langchain in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.1.11) Collecting langchain Downloading langchain-0.1.14-py3-none-any.whl.metadata (13 kB) Requirement already satisfied: langgraph in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.0.30) Collecting langgraph Downloading langgraph-0.0.31-py3-none-any.whl.metadata (44 kB) [2K [38;2;114;156;31m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m44.8/44.8 kB[0m [31m782.7 kB/s[0m eta [36m0:00:00[0m31m1.0 MB/s[0m eta [36m0:00:01[0m [?25hRequirement already satisfied: tavily-python in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (0.3.1) Collecting tavily-python Downloading tavily_python-0.3.3-py3-none-any.whl.metadata (4.4 kB) Requirement already satisfied: PyYAML>=5.3 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (6.0.1) Requirement already satisfied: SQLAlchemy<3,>=1.4 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (2.0.28) Requirement already satisfied: aiohttp<4.0.0,>=3.8.3 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (3.9.3) Requirement already satisfied: dataclasses-json<0.7,>=0.5.7 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (0.6.4) Requirement already satisfied: langchain-core<0.2.0,>=0.1.37 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (0.1.38) Requirement already satisfied: langsmith<0.2.0,>=0.1.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (0.1.23) Requirement already satisfied: numpy<2,>=1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (1.26.4) Requirement already satisfied: requests<3,>=2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (2.31.0) Requirement already satisfied: tenacity<9.0.0,>=8.1.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain_community) (8.2.3) Requirement already satisfied: regex>=2022.1.18 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from tiktoken) (2023.12.25) Requirement already satisfied: openai<2.0.0,>=1.10.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain-openai) (1.13.3) Requirement already satisfied: cohere<6.0.0,>=5.1.4 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain-cohere) (5.1.7) Requirement already satisfied: types-requests<3.0.0.0,>=2.31.0.2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchainhub) (2.31.0.20240311) Requirement already satisfied: build>=1.0.3 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (1.2.1) Requirement already satisfied: pydantic>=1.9 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (2.6.4) Requirement already satisfied: chroma-hnswlib==0.7.3 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (0.7.3) Requirement already satisfied: fastapi>=0.95.2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (0.109.0) Requirement already satisfied: uvicorn>=0.18.3 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from uvicorn[standard]>=0.18.3->chromadb) (0.25.0) Requirement already satisfied: posthog>=2.4.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (3.5.0) Requirement already satisfied: typing-extensions>=4.5.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (4.10.0) Requirement already satisfied: pulsar-client>=3.1.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (3.4.0) Requirement already satisfied: onnxruntime>=1.14.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (1.17.1) Requirement already satisfied: opentelemetry-api>=1.2.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (1.24.0) Requirement already satisfied: opentelemetry-exporter-otlp-proto-grpc>=1.2.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (1.24.0) Requirement already satisfied: opentelemetry-instrumentation-fastapi>=0.41b0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (0.45b0) Requirement already satisfied: opentelemetry-sdk>=1.2.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (1.24.0) Requirement already satisfied: tokenizers>=0.13.2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (0.15.2) Requirement already satisfied: pypika>=0.48.9 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (0.48.9) Requirement already satisfied: tqdm>=4.65.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (4.66.2) Requirement already satisfied: overrides>=7.3.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (7.7.0) Requirement already satisfied: importlib-resources in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (6.4.0) Requirement already satisfied: grpcio>=1.58.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (1.62.1) Requirement already satisfied: bcrypt>=4.0.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (4.1.2) Requirement already satisfied: typer>=0.9.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (0.12.0) Requirement already satisfied: kubernetes>=28.1.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (29.0.0) Requirement already satisfied: mmh3>=4.0.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (4.1.0) Requirement already satisfied: orjson>=3.9.12 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from chromadb) (3.9.15) Requirement already satisfied: jsonpatch<2.0,>=1.33 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain) (1.33) Requirement already satisfied: langchain-text-splitters<0.1,>=0.0.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from langchain) (0.0.1) Requirement already satisfied: aiosignal>=1.1.2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain_community) (1.3.1) Requirement already satisfied: attrs>=17.3.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain_community) (23.2.0) Requirement already satisfied: frozenlist>=1.1.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain_community) (1.4.1) Requirement already satisfied: multidict<7.0,>=4.5 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain_community) (6.0.5) Requirement already satisfied: yarl<2.0,>=1.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from aiohttp<4.0.0,>=3.8.3->langchain_community) (1.9.4) Requirement already satisfied: packaging>=19.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from build>=1.0.3->chromadb) (23.2) Requirement already satisfied: pyproject_hooks in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from build>=1.0.3->chromadb) (1.0.0) Requirement already satisfied: fastavro<2.0.0,>=1.9.4 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from cohere<6.0.0,>=5.1.4->langchain-cohere) (1.9.4) Requirement already satisfied: httpx>=0.21.2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from cohere<6.0.0,>=5.1.4->langchain-cohere) (0.26.0) Requirement already satisfied: marshmallow<4.0.0,>=3.18.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from dataclasses-json<0.7,>=0.5.7->langchain_community) (3.21.1) Requirement already satisfied: typing-inspect<1,>=0.4.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from dataclasses-json<0.7,>=0.5.7->langchain_community) (0.9.0) Requirement already satisfied: starlette<0.36.0,>=0.35.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from fastapi>=0.95.2->chromadb) (0.35.1) Requirement already satisfied: jsonpointer>=1.9 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from jsonpatch<2.0,>=1.33->langchain) (2.4) Requirement already satisfied: certifi>=14.05.14 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from kubernetes>=28.1.0->chromadb) (2024.2.2) Requirement already satisfied: six>=1.9.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from kubernetes>=28.1.0->chromadb) (1.16.0) Requirement already satisfied: python-dateutil>=2.5.3 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from kubernetes>=28.1.0->chromadb) (2.9.0.post0) Requirement already satisfied: google-auth>=1.0.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from kubernetes>=28.1.0->chromadb) (2.29.0) Requirement already satisfied: websocket-client!=0.40.0,!=0.41.*,!=0.42.*,>=0.32.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from kubernetes>=28.1.0->chromadb) (1.7.0) Requirement already satisfied: requests-oauthlib in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from kubernetes>=28.1.0->chromadb) (2.0.0) Requirement already satisfied: oauthlib>=3.2.2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from kubernetes>=28.1.0->chromadb) (3.2.2) Requirement already satisfied: urllib3>=1.24.2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from kubernetes>=28.1.0->chromadb) (2.2.1) Requirement already satisfied: coloredlogs in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from onnxruntime>=1.14.1->chromadb) (15.0.1) Requirement already satisfied: flatbuffers in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from onnxruntime>=1.14.1->chromadb) (24.3.25) Requirement already satisfied: protobuf in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from onnxruntime>=1.14.1->chromadb) (4.25.3) Requirement already satisfied: sympy in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from onnxruntime>=1.14.1->chromadb) (1.12) Requirement already satisfied: anyio<5,>=3.5.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from openai<2.0.0,>=1.10.0->langchain-openai) (4.3.0) Requirement already satisfied: distro<2,>=1.7.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from openai<2.0.0,>=1.10.0->langchain-openai) (1.9.0) Requirement already satisfied: sniffio in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from openai<2.0.0,>=1.10.0->langchain-openai) (1.3.1) Requirement already satisfied: deprecated>=1.2.6 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-api>=1.2.0->chromadb) (1.2.14) Requirement already satisfied: importlib-metadata<=7.0,>=6.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-api>=1.2.0->chromadb) (7.0.0) Requirement already satisfied: googleapis-common-protos~=1.52 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-exporter-otlp-proto-grpc>=1.2.0->chromadb) (1.63.0) Requirement already satisfied: opentelemetry-exporter-otlp-proto-common==1.24.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-exporter-otlp-proto-grpc>=1.2.0->chromadb) (1.24.0) Requirement already satisfied: opentelemetry-proto==1.24.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-exporter-otlp-proto-grpc>=1.2.0->chromadb) (1.24.0) Requirement already satisfied: opentelemetry-instrumentation-asgi==0.45b0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-instrumentation-fastapi>=0.41b0->chromadb) (0.45b0) Requirement already satisfied: opentelemetry-instrumentation==0.45b0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-instrumentation-fastapi>=0.41b0->chromadb) (0.45b0) Requirement already satisfied: opentelemetry-semantic-conventions==0.45b0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-instrumentation-fastapi>=0.41b0->chromadb) (0.45b0) Requirement already satisfied: opentelemetry-util-http==0.45b0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-instrumentation-fastapi>=0.41b0->chromadb) (0.45b0) Requirement already satisfied: setuptools>=16.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-instrumentation==0.45b0->opentelemetry-instrumentation-fastapi>=0.41b0->chromadb) (69.0.2) Requirement already satisfied: wrapt<2.0.0,>=1.0.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-instrumentation==0.45b0->opentelemetry-instrumentation-fastapi>=0.41b0->chromadb) (1.16.0) Requirement already satisfied: asgiref~=3.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from opentelemetry-instrumentation-asgi==0.45b0->opentelemetry-instrumentation-fastapi>=0.41b0->chromadb) (3.8.1) Requirement already satisfied: monotonic>=1.5 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from posthog>=2.4.0->chromadb) (1.6) Requirement already satisfied: backoff>=1.10.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from posthog>=2.4.0->chromadb) (2.2.1) Requirement already satisfied: annotated-types>=0.4.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from pydantic>=1.9->chromadb) (0.6.0) Requirement already satisfied: pydantic-core==2.16.3 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from pydantic>=1.9->chromadb) (2.16.3) Requirement already satisfied: charset-normalizer<4,>=2 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from requests<3,>=2->langchain_community) (3.3.2) Requirement already satisfied: idna<4,>=2.5 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from requests<3,>=2->langchain_community) (3.6) Requirement already satisfied: huggingface_hub<1.0,>=0.16.4 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from tokenizers>=0.13.2->chromadb) (0.22.1) Requirement already satisfied: typer-slim==0.12.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from typer-slim[standard]==0.12.0->typer>=0.9.0->chromadb) (0.12.0) Requirement already satisfied: typer-cli==0.12.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from typer>=0.9.0->chromadb) (0.12.0) Requirement already satisfied: click>=8.0.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from typer-slim==0.12.0->typer-slim[standard]==0.12.0->typer>=0.9.0->chromadb) (8.1.7) Requirement already satisfied: shellingham>=1.3.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from typer-slim[standard]==0.12.0->typer>=0.9.0->chromadb) (1.5.4) Requirement already satisfied: rich>=10.11.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from typer-slim[standard]==0.12.0->typer>=0.9.0->chromadb) (13.7.1) Requirement already satisfied: h11>=0.8 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from uvicorn>=0.18.3->uvicorn[standard]>=0.18.3->chromadb) (0.14.0) Requirement already satisfied: httptools>=0.5.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from uvicorn[standard]>=0.18.3->chromadb) (0.6.1) Requirement already satisfied: python-dotenv>=0.13 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from uvicorn[standard]>=0.18.3->chromadb) (1.0.1) Requirement already satisfied: uvloop!=0.15.0,!=0.15.1,>=0.14.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from uvicorn[standard]>=0.18.3->chromadb) (0.19.0) Requirement already satisfied: watchfiles>=0.13 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from uvicorn[standard]>=0.18.3->chromadb) (0.21.0) Requirement already satisfied: websockets>=10.4 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from uvicorn[standard]>=0.18.3->chromadb) (12.0) Requirement already satisfied: cachetools<6.0,>=2.0.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from google-auth>=1.0.1->kubernetes>=28.1.0->chromadb) (5.3.3) Requirement already satisfied: pyasn1-modules>=0.2.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from google-auth>=1.0.1->kubernetes>=28.1.0->chromadb) (0.4.0) Requirement already satisfied: rsa<5,>=3.1.4 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from google-auth>=1.0.1->kubernetes>=28.1.0->chromadb) (4.9) Requirement already satisfied: httpcore==1.* in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from httpx>=0.21.2->cohere<6.0.0,>=5.1.4->langchain-cohere) (1.0.4) Requirement already satisfied: filelock in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from huggingface_hub<1.0,>=0.16.4->tokenizers>=0.13.2->chromadb) (3.13.3) Requirement already satisfied: fsspec>=2023.5.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from huggingface_hub<1.0,>=0.16.4->tokenizers>=0.13.2->chromadb) (2024.3.1) Requirement already satisfied: zipp>=0.5 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from importlib-metadata<=7.0,>=6.0->opentelemetry-api>=1.2.0->chromadb) (3.18.1) Requirement already satisfied: mypy-extensions>=0.3.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from typing-inspect<1,>=0.4.0->dataclasses-json<0.7,>=0.5.7->langchain_community) (1.0.0) Requirement already satisfied: humanfriendly>=9.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from coloredlogs->onnxruntime>=1.14.1->chromadb) (10.0) Requirement already satisfied: mpmath>=0.19 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from sympy->onnxruntime>=1.14.1->chromadb) (1.3.0) Requirement already satisfied: pyasn1<0.7.0,>=0.4.6 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from pyasn1-modules>=0.2.1->google-auth>=1.0.1->kubernetes>=28.1.0->chromadb) (0.6.0) Requirement already satisfied: markdown-it-py>=2.2.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from rich>=10.11.0->typer-slim[standard]==0.12.0->typer>=0.9.0->chromadb) (3.0.0) Requirement already satisfied: pygments<3.0.0,>=2.13.0 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from rich>=10.11.0->typer-slim[standard]==0.12.0->typer>=0.9.0->chromadb) (2.17.2) Requirement already satisfied: mdurl~=0.1 in /Users/nuno/dev/langgraph/.venv/lib/python3.11/site-packages (from markdown-it-py>=2.2.0->rich>=10.11.0->typer-slim[standard]==0.12.0->typer>=0.9.0->chromadb) (0.1.2) Downloading langchain_community-0.0.31-py3-none-any.whl (1.9 MB) [2K [38;2;114;156;31m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m1.9/1.9 MB[0m [31m9.9 MB/s[0m eta [36m0:00:00[0m[31m10.3 MB/s[0m eta [36m0:00:01[0m [?25hDownloading tiktoken-0.6.0-cp311-cp311-macosx_11_0_arm64.whl (949 kB) [2K [38;2;114;156;31m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m949.8/949.8 kB[0m [31m15.0 MB/s[0m eta [36m0:00:00[0m31m20.3 MB/s[0m eta [36m0:00:01[0m [?25hDownloading langchain_openai-0.1.1-py3-none-any.whl (32 kB) Downloading langchain-0.1.14-py3-none-any.whl (812 kB) [2K [38;2;114;156;31m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m812.8/812.8 kB[0m [31m14.2 MB/s[0m eta [36m0:00:00[0m MB/s[0m eta [36m0:00:01[0m [?25hDownloading langgraph-0.0.31-py3-none-any.whl (55 kB) [2K [38;2;114;156;31m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━[0m [32m55.5/55.5 kB[0m [31m4.5 MB/s[0m eta [36m0:00:00[0m [?25hDownloading tavily_python-0.3.3-py3-none-any.whl (5.4 kB) Installing collected packages: tiktoken, tavily-python, langgraph, langchain-openai, langchain_community, langchain Attempting uninstall: tiktoken Found existing installation: tiktoken 0.5.2 Uninstalling tiktoken-0.5.2: Successfully uninstalled tiktoken-0.5.2 Attempting uninstall: tavily-python Found existing installation: tavily-python 0.3.1 Uninstalling tavily-python-0.3.1: Successfully uninstalled tavily-python-0.3.1 Attempting uninstall: langgraph Found existing installation: langgraph 0.0.30 Uninstalling langgraph-0.0.30: Successfully uninstalled langgraph-0.0.30 Attempting uninstall: langchain-openai Found existing installation: langchain-openai 0.0.2.post1 Uninstalling langchain-openai-0.0.2.post1: Successfully uninstalled langchain-openai-0.0.2.post1 Attempting uninstall: langchain_community Found existing installation: langchain-community 0.0.27 Uninstalling langchain-community-0.0.27: Successfully uninstalled langchain-community-0.0.27 Attempting uninstall: langchain Found existing installation: langchain 0.1.11 Uninstalling langchain-0.1.11: Successfully uninstalled langchain-0.1.11 Successfully installed langchain-0.1.14 langchain-openai-0.1.1 langchain_community-0.0.31 langgraph-0.0.31 tavily-python-0.3.3 tiktoken-0.6.0 [1m[[0m[34;49mnotice[0m[1;39;49m][0m[39;49m A new release of pip is available: [0m[31;49m23.3.1[0m[39;49m -> [0m[32;49m24.0[0m [1m[[0m[34;49mnotice[0m[1;39;49m][0m[39;49m To update, run: [0m[32;49mpip install --upgrade pip[0m
In [ ]:
### LLMs
import os
os.environ['OPENAI_API_KEY'] = <your-api-key>
os.environ['COHERE_API_KEY'] = <your-api-key>
os.environ['TAVILY_API_KEY'] = <your-api-key>In [ ]:
### Tracing (optional)
os.environ['LANGCHAIN_TRACING_V2'] = 'true'
os.environ['LANGCHAIN_ENDPOINT'] = 'https://api.smith.langchain.com'
os.environ['LANGCHAIN_API_KEY'] = <your-api-key>In [1]:
### Build Index
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.document_loaders import WebBaseLoader
from langchain_community.vectorstores import Chroma
from langchain_openai import OpenAIEmbeddings
### from langchain_cohere import CohereEmbeddings
# Set embeddings
embd = OpenAIEmbeddings()
# Docs to index
urls = [
"https://lilianweng.github.io/posts/2023-06-23-agent/",
"https://lilianweng.github.io/posts/2023-03-15-prompt-engineering/",
"https://lilianweng.github.io/posts/2023-10-25-adv-attack-llm/",
]
# Load
docs = [WebBaseLoader(url).load() for url in urls]
docs_list = [item for sublist in docs for item in sublist]
# Split
text_splitter = RecursiveCharacterTextSplitter.from_tiktoken_encoder(
chunk_size=500, chunk_overlap=0
)
doc_splits = text_splitter.split_documents(docs_list)
# Add to vectorstore
vectorstore = Chroma.from_documents(
documents=doc_splits,
collection_name="rag-chroma",
embedding=embd,
)
retriever = vectorstore.as_retriever()In [3]:
### Router
from typing import Literal
from langchain_core.prompts import ChatPromptTemplate
from langchain_core.pydantic_v1 import BaseModel, Field
from langchain_openai import ChatOpenAI
# Data model
class RouteQuery(BaseModel):
"""Route a user query to the most relevant datasource."""
datasource: Literal["vectorstore", "web_search"] = Field(
...,
description="Given a user question choose to route it to web search or a vectorstore.",
)
# LLM with function call
llm = ChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm_router = llm.with_structured_output(RouteQuery)
# Prompt
system = """You are an expert at routing a user question to a vectorstore or web search.
The vectorstore contains documents related to agents, prompt engineering, and adversarial attacks.
Use the vectorstore for questions on these topics. Otherwise, use web-search."""
route_prompt = ChatPromptTemplate.from_messages(
[
("system", system),
("human", "{question}"),
]
)
question_router = route_prompt | structured_llm_router
print(question_router.invoke({"question": "Who will the Bears draft first in the NFL draft?"}))
print(question_router.invoke({"question": "What are the types of agent memory?"}))datasource='web_search' datasource='vectorstore'
In [4]:
### Retrieval Grader
# Data model
class GradeDocuments(BaseModel):
"""Binary score for relevance check on retrieved documents."""
binary_score: str = Field(description="Documents are relevant to the question, 'yes' or 'no'")
# LLM with function call
llm = ChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm_grader = llm.with_structured_output(GradeDocuments)
# Prompt
system = """You are a grader assessing relevance of a retrieved document to a user question. \n
If the document contains keyword(s) or semantic meaning related to the user question, grade it as relevant. \n
It does not need to be a stringent test. The goal is to filter out erroneous retrievals. \n
Give a binary score 'yes' or 'no' score to indicate whether the document is relevant to the question."""
grade_prompt = ChatPromptTemplate.from_messages(
[
("system", system),
("human", "Retrieved document: \n\n {document} \n\n User question: {question}"),
]
)
retrieval_grader = grade_prompt | structured_llm_grader
question = "agent memory"
docs = retriever.get_relevant_documents(question)
doc_txt = docs[1].page_content
print(retrieval_grader.invoke({"question": question, "document": doc_txt}))binary_score='no'
In [5]:
### Generate
from langchain import hub
from langchain_core.output_parsers import StrOutputParser
# Prompt
prompt = hub.pull("rlm/rag-prompt")
# LLM
llm = ChatOpenAI(model_name="gpt-3.5-turbo", temperature=0)
# Post-processing
def format_docs(docs):
return "\n\n".join(doc.page_content for doc in docs)
# Chain
rag_chain = prompt | llm | StrOutputParser()
# Run
generation = rag_chain.invoke({"context": docs, "question": question})
print(generation)The design of generative agents combines LLM with memory, planning, and reflection mechanisms to enable agents to behave based on past experience and interact with other agents. Memory stream is a long-term memory module that records agents' experiences in natural language. The retrieval model surfaces context to inform the agent's behavior based on relevance, recency, and importance.
In [6]:
### Hallucination Grader
# Data model
class GradeHallucinations(BaseModel):
"""Binary score for hallucination present in generation answer."""
binary_score: str = Field(description="Answer is grounded in the facts, 'yes' or 'no'")
# LLM with function call
llm = ChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm_grader = llm.with_structured_output(GradeHallucinations)
# Prompt
system = """You are a grader assessing whether an LLM generation is grounded in / supported by a set of retrieved facts. \n
Give a binary score 'yes' or 'no'. 'Yes' means that the answer is grounded in / supported by the set of facts."""
hallucination_prompt = ChatPromptTemplate.from_messages(
[
("system", system),
("human", "Set of facts: \n\n {documents} \n\n LLM generation: {generation}"),
]
)
hallucination_grader = hallucination_prompt | structured_llm_grader
hallucination_grader.invoke({"documents": docs, "generation": generation})Out [6]:
GradeHallucinations(binary_score='yes')
In [7]:
### Answer Grader
# Data model
class GradeAnswer(BaseModel):
"""Binary score to assess answer addresses question."""
binary_score: str = Field(description="Answer addresses the question, 'yes' or 'no'")
# LLM with function call
llm = ChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
structured_llm_grader = llm.with_structured_output(GradeAnswer)
# Prompt
system = """You are a grader assessing whether an answer addresses / resolves a question \n
Give a binary score 'yes' or 'no'. Yes' means that the answer resolves the question."""
answer_prompt = ChatPromptTemplate.from_messages(
[
("system", system),
("human", "User question: \n\n {question} \n\n LLM generation: {generation}"),
]
)
answer_grader = answer_prompt | structured_llm_grader
answer_grader.invoke({"question": question,"generation": generation})Out [7]:
GradeAnswer(binary_score='yes')
In [8]:
### Question Re-writer
# LLM
llm = ChatOpenAI(model="gpt-3.5-turbo-0125", temperature=0)
# Prompt
system = """You a question re-writer that converts an input question to a better version that is optimized \n
for vectorstore retrieval. Look at the input and try to reason about the underlying sematic intent / meaning."""
re_write_prompt = ChatPromptTemplate.from_messages(
[
("system", system),
("human", "Here is the initial question: \n\n {question} \n Formulate an improved question."),
]
)
question_rewriter = re_write_prompt | llm | StrOutputParser()
question_rewriter.invoke({"question": question})Out [8]:
"What is the role of memory in an agent's functioning?"
In [9]:
### Search
from langchain_community.tools.tavily_search import TavilySearchResults
web_search_tool = TavilySearchResults(k=3)In [10]:
from typing_extensions import TypedDict
from typing import List
class GraphState(TypedDict):
"""
Represents the state of our graph.
Attributes:
question: question
generation: LLM generation
documents: list of documents
"""
question : str
generation : str
documents : List[str]In [15]:
from langchain.schema import Document
def retrieve(state):
"""
Retrieve documents
Args:
state (dict): The current graph state
Returns:
state (dict): New key added to state, documents, that contains retrieved documents
"""
print("---RETRIEVE---")
question = state["question"]
# Retrieval
documents = retriever.invoke(question)
return {"documents": documents, "question": question}
def generate(state):
"""
Generate answer
Args:
state (dict): The current graph state
Returns:
state (dict): New key added to state, generation, that contains LLM generation
"""
print("---GENERATE---")
question = state["question"]
documents = state["documents"]
# RAG generation
generation = rag_chain.invoke({"context": documents, "question": question})
return {"documents": documents, "question": question, "generation": generation}
def grade_documents(state):
"""
Determines whether the retrieved documents are relevant to the question.
Args:
state (dict): The current graph state
Returns:
state (dict): Updates documents key with only filtered relevant documents
"""
print("---CHECK DOCUMENT RELEVANCE TO QUESTION---")
question = state["question"]
documents = state["documents"]
# Score each doc
filtered_docs = []
for d in documents:
score = retrieval_grader.invoke({"question": question, "document": d.page_content})
grade = score.binary_score
if grade == "yes":
print("---GRADE: DOCUMENT RELEVANT---")
filtered_docs.append(d)
else:
print("---GRADE: DOCUMENT NOT RELEVANT---")
continue
return {"documents": filtered_docs, "question": question}
def transform_query(state):
"""
Transform the query to produce a better question.
Args:
state (dict): The current graph state
Returns:
state (dict): Updates question key with a re-phrased question
"""
print("---TRANSFORM QUERY---")
question = state["question"]
documents = state["documents"]
# Re-write question
better_question = question_rewriter.invoke({"question": question})
return {"documents": documents, "question": better_question}
def web_search(state):
"""
Web search based on the re-phrased question.
Args:
state (dict): The current graph state
Returns:
state (dict): Updates documents key with appended web results
"""
print("---WEB SEARCH---")
question = state["question"]
# Web search
docs = web_search_tool.invoke({"query": question})
web_results = "\n".join([d["content"] for d in docs])
web_results = Document(page_content=web_results)
return {"documents": web_results, "question": question}
### Edges ###
def route_question(state):
"""
Route question to web search or RAG.
Args:
state (dict): The current graph state
Returns:
str: Next node to call
"""
print("---ROUTE QUESTION---")
question = state["question"]
source = question_router.invoke({"question": question})
if source.datasource == 'web_search':
print("---ROUTE QUESTION TO WEB SEARCH---")
return "web_search"
elif source.datasource == 'vectorstore':
print("---ROUTE QUESTION TO RAG---")
return "vectorstore"
def decide_to_generate(state):
"""
Determines whether to generate an answer, or re-generate a question.
Args:
state (dict): The current graph state
Returns:
str: Binary decision for next node to call
"""
print("---ASSESS GRADED DOCUMENTS---")
question = state["question"]
filtered_documents = state["documents"]
if not filtered_documents:
# All documents have been filtered check_relevance
# We will re-generate a new query
print("---DECISION: ALL DOCUMENTS ARE NOT RELEVANT TO QUESTION, TRANSFORM QUERY---")
return "transform_query"
else:
# We have relevant documents, so generate answer
print("---DECISION: GENERATE---")
return "generate"
def grade_generation_v_documents_and_question(state):
"""
Determines whether the generation is grounded in the document and answers question.
Args:
state (dict): The current graph state
Returns:
str: Decision for next node to call
"""
print("---CHECK HALLUCINATIONS---")
question = state["question"]
documents = state["documents"]
generation = state["generation"]
score = hallucination_grader.invoke({"documents": documents, "generation": generation})
grade = score.binary_score
# Check hallucination
if grade == "yes":
print("---DECISION: GENERATION IS GROUNDED IN DOCUMENTS---")
# Check question-answering
print("---GRADE GENERATION vs QUESTION---")
score = answer_grader.invoke({"question": question,"generation": generation})
grade = score.binary_score
if grade == "yes":
print("---DECISION: GENERATION ADDRESSES QUESTION---")
return "useful"
else:
print("---DECISION: GENERATION DOES NOT ADDRESS QUESTION---")
return "not useful"
else:
pprint("---DECISION: GENERATION IS NOT GROUNDED IN DOCUMENTS, RE-TRY---")
return "not supported"In [16]:
from langgraph.graph import END, StateGraph
workflow = StateGraph(GraphState)
# Define the nodes
workflow.add_node("web_search", web_search) # web search
workflow.add_node("retrieve", retrieve) # retrieve
workflow.add_node("grade_documents", grade_documents) # grade documents
workflow.add_node("generate", generate) # generatae
workflow.add_node("transform_query", transform_query) # transform_query
# Build graph
workflow.set_conditional_entry_point(
route_question,
{
"web_search": "web_search",
"vectorstore": "retrieve",
},
)
workflow.add_edge("web_search", "generate")
workflow.add_edge("retrieve", "grade_documents")
workflow.add_conditional_edges(
"grade_documents",
decide_to_generate,
{
"transform_query": "transform_query",
"generate": "generate",
},
)
workflow.add_edge("transform_query", "retrieve")
workflow.add_conditional_edges(
"generate",
grade_generation_v_documents_and_question,
{
"not supported": "generate",
"useful": END,
"not useful": "transform_query",
},
)
# Compile
app = workflow.compile()In [17]:
from pprint import pprint
# Run
inputs = {"question": "What player at the Bears expected to draft first in the 2024 NFL draft?"}
for output in app.stream(inputs):
for key, value in output.items():
# Node
pprint(f"Node '{key}':")
# Optional: print full state at each node
# pprint.pprint(value["keys"], indent=2, width=80, depth=None)
pprint("\n---\n")
# Final generation
pprint(value["generation"])---ROUTE QUESTION---
---ROUTE QUESTION TO WEB SEARCH---
---WEB SEARCH---
"Node 'web_search':"
'\n---\n'
---GENERATE---
---CHECK HALLUCINATIONS---
---DECISION: GENERATION IS GROUNDED IN DOCUMENTS---
---GRADE GENERATION vs QUESTION---
---DECISION: GENERATION ADDRESSES QUESTION---
"Node 'generate':"
'\n---\n'
('It is expected that the Chicago Bears could have the opportunity to draft '
'the first defensive player in the 2024 NFL draft. The Bears have the first '
'overall pick in the draft, giving them a prime position to select top '
'talent. The top wide receiver Marvin Harrison Jr. from Ohio State is also '
'mentioned as a potential pick for the Cardinals.')
In [18]:
# Run
inputs = {"question": "What are the types of agent memory?"}
for output in app.stream(inputs):
for key, value in output.items():
# Node
pprint(f"Node '{key}':")
# Optional: print full state at each node
# pprint.pprint(value["keys"], indent=2, width=80, depth=None)
pprint("\n---\n")
# Final generation
pprint(value ["generation"])---ROUTE QUESTION---
---ROUTE QUESTION TO RAG---
---RETRIEVE---
"Node 'retrieve':"
'\n---\n'
---CHECK DOCUMENT RELEVANCE TO QUESTION---
---GRADE: DOCUMENT RELEVANT---
---GRADE: DOCUMENT RELEVANT---
---GRADE: DOCUMENT NOT RELEVANT---
---GRADE: DOCUMENT RELEVANT---
---ASSESS GRADED DOCUMENTS---
---DECISION: GENERATE---
"Node 'grade_documents':"
'\n---\n'
---GENERATE---
---CHECK HALLUCINATIONS---
---DECISION: GENERATION IS GROUNDED IN DOCUMENTS---
---GRADE GENERATION vs QUESTION---
---DECISION: GENERATION ADDRESSES QUESTION---
"Node 'generate':"
'\n---\n'
('The types of agent memory include Sensory Memory, Short-Term Memory (STM) or '
'Working Memory, and Long-Term Memory (LTM) with subtypes of Explicit / '
'declarative memory and Implicit / procedural memory. Sensory memory retains '
'sensory information briefly, STM stores information for cognitive tasks, and '
'LTM stores information for a long time with different types of memories.')
In [ ]: