|
| 1 | +# rag-self-query |
| 2 | + |
| 3 | +This template performs RAG using the self-query retrieval technique. The main idea is to let an LLM convert unstructured queries into structured queries. See the [docs for more on how this works](https://python.langchain.com/docs/modules/data_connection/retrievers/self_query). |
| 4 | + |
| 5 | +## Environment Setup |
| 6 | + |
| 7 | +In this template we'll use OpenAI models and an Elasticsearch vector store, but the approach generalizes to all LLMs/ChatModels and [a number of vector stores](https://python.langchain.com/docs/integrations/retrievers/self_query/). |
| 8 | + |
| 9 | +Set the `OPENAI_API_KEY` environment variable to access the OpenAI models. |
| 10 | + |
| 11 | +To connect to your Elasticsearch instance, use the following environment variables: |
| 12 | + |
| 13 | +```bash |
| 14 | +export ELASTIC_CLOUD_ID = <ClOUD_ID> |
| 15 | +export ELASTIC_USERNAME = <ClOUD_USERNAME> |
| 16 | +export ELASTIC_PASSWORD = <ClOUD_PASSWORD> |
| 17 | +``` |
| 18 | +For local development with Docker, use: |
| 19 | + |
| 20 | +```bash |
| 21 | +export ES_URL = "http://localhost:9200" |
| 22 | +docker run -p 9200:9200 -e "discovery.type=single-node" -e "xpack.security.enabled=false" -e "xpack.security.http.ssl.enabled=false" docker.elastic.co/elasticsearch/elasticsearch:8.9.0 |
| 23 | +``` |
| 24 | + |
| 25 | +## Usage |
| 26 | + |
| 27 | +To use this package, you should first have the LangChain CLI installed: |
| 28 | + |
| 29 | +```shell |
| 30 | +pip install -U "langchain-cli[serve]" |
| 31 | +``` |
| 32 | + |
| 33 | +To create a new LangChain project and install this as the only package, you can do: |
| 34 | + |
| 35 | +```shell |
| 36 | +langchain app new my-app --package rag-self-query |
| 37 | +``` |
| 38 | + |
| 39 | +If you want to add this to an existing project, you can just run: |
| 40 | + |
| 41 | +```shell |
| 42 | +langchain app add rag-self-query |
| 43 | +``` |
| 44 | + |
| 45 | +And add the following code to your `server.py` file: |
| 46 | +```python |
| 47 | +from rag_self_query import chain |
| 48 | + |
| 49 | +add_routes(app, chain, path="/rag-elasticsearch") |
| 50 | +``` |
| 51 | + |
| 52 | +To populate the vector store with the sample data, from the root of the directory run: |
| 53 | +```bash |
| 54 | +python ingest.py |
| 55 | +``` |
| 56 | + |
| 57 | +(Optional) Let's now configure LangSmith. |
| 58 | +LangSmith will help us trace, monitor and debug LangChain applications. |
| 59 | +LangSmith is currently in private beta, you can sign up [here](https://smith.langchain.com/). |
| 60 | +If you don't have access, you can skip this section |
| 61 | + |
| 62 | +```shell |
| 63 | +export LANGCHAIN_TRACING_V2=true |
| 64 | +export LANGCHAIN_API_KEY=<your-api-key> |
| 65 | +export LANGCHAIN_PROJECT=<your-project> # if not specified, defaults to "default" |
| 66 | +``` |
| 67 | + |
| 68 | +If you are inside this directory, then you can spin up a LangServe instance directly by: |
| 69 | + |
| 70 | +```shell |
| 71 | +langchain serve |
| 72 | +``` |
| 73 | + |
| 74 | +This will start the FastAPI app with a server is running locally at |
| 75 | +[http://localhost:8000](http://localhost:8000) |
| 76 | + |
| 77 | +We can see all templates at [http://127.0.0.1:8000/docs](http://127.0.0.1:8000/docs) |
| 78 | +We can access the playground at [http://127.0.0.1:8000/rag-elasticsearch/playground](http://127.0.0.1:8000/rag-elasticsearch/playground) |
| 79 | + |
| 80 | +We can access the template from code with: |
| 81 | + |
| 82 | +```python |
| 83 | +from langserve.client import RemoteRunnable |
| 84 | + |
| 85 | +runnable = RemoteRunnable("http://localhost:8000/rag-self-query") |
| 86 | +``` |
0 commit comments