How to do RAG on a already created weaviate persisted data

[hboen1990]

Hi All,
I am exploring weaviate with langchain. I have loaded a bunch of pdf’s and do the standard splitting and create a weaviate class as
class_obj = {
"class": "WMOInfo",
"properties": [{"dataType": ["text"], "name": "content"}],
"vectorizer": "text2vec-huggingface",
}
And use the BGE embedding
model_name = "BAAI/bge-base-en-v1.5"
encode_kwargs = {'normalize_embeddings': True} # set True to compute cosine similarity

model_norm = HuggingFaceBgeEmbeddings(
model_name=model_name,
model_kwargs={'device': 'cpu'},
encode_kwargs=encode_kwargs
embedding = model_norm
and create a vector store as
weaviate_vectorstore = Weaviate.from_documents(documents=texts,
embedding=embedding,
client=client,
by_text=False)
with the restriever:
weaviate_retriever = weaviate_vectorstore.as_retriever(search_kwargs={"k":3})
response = weaviate_retriever.get_relevant_documents(query="some query”)
print(response) -> I get the relevant documents.
The above code for digesting the pdf’s works.

Now I want to use the vectorstore to do RAG. I cannot find any code to access the persisted data to create a vectorstore without doing the ingestion of the documents again. In chroma it's easy you just do vectorstore = Chroma(persist_directory=persist_directory, embedding_function=embedding).
How can I do this in Weaviate similar as Chroma?

I am using a local version of weaviate using Docker.

Thanks in advance.

Schema output:

{
"classes": [
{
"class": "WMOInfo",
"invertedIndexConfig": {
"bm25": {
"b": 0.75,
"k1": 1.2
},
"cleanupIntervalSeconds": 60,
"stopwords": {
"additions": null,
"preset": "en",
"removals": null
}
},
"moduleConfig": {
"text2vec-huggingface": {
"vectorizeClassName": true
}
},
"multiTenancyConfig": {
"enabled": false
},
"properties": [
{
"dataType": [
"text"
],
"indexFilterable": true,
"indexSearchable": true,
"moduleConfig": {
"text2vec-huggingface": {
"skip": false,
"vectorizePropertyName": false
}
},
"name": "content",
"tokenization": "word"
}
],
"replicationConfig": {
"factor": 1
},
"shardingConfig": {
"virtualPerPhysical": 128,
"desiredCount": 1,
"actualCount": 1,
"desiredVirtualCount": 128,
"actualVirtualCount": 128,
"key": "_id",
"strategy": "hash",
"function": "murmur3"
},
"vectorIndexConfig": {
"skip": false,
"cleanupIntervalSeconds": 300,
"maxConnections": 64,
"efConstruction": 128,
"ef": -1,
"dynamicEfMin": 100,
"dynamicEfMax": 500,
"dynamicEfFactor": 8,
"vectorCacheMaxObjects": 1000000000000,
"flatSearchCutoff": 40000,
"distance": "cosine",
"pq": {
"enabled": false,
"bitCompression": false,
"segments": 0,
"centroids": 256,
"trainingLimit": 100000,
"encoder": {
"type": "kmeans",
"distribution": "log-normal"
}
}
},
"vectorIndexType": "hnsw",
"vectorizer": "text2vec-huggingface"
},
{
"class": "LangChain_71db963cf5114df8b31ef1edde728ef3",
"invertedIndexConfig": {
"bm25": {
"b": 0.75,
"k1": 1.2
},
"cleanupIntervalSeconds": 60,
"stopwords": {
"additions": null,
"preset": "en",
"removals": null
}
},
"multiTenancyConfig": {
"enabled": false
},
"properties": [
{
"dataType": [
"text"
],
"indexFilterable": true,
"indexSearchable": true,
"name": "text",
"tokenization": "word"
},
{
"dataType": [
"text"
],
"description": "This property was generated by Weaviate's auto-schema feature on Fri Dec 29 01:00:07 2023",
"indexFilterable": true,
"indexSearchable": true,
"name": "source",
"tokenization": "word"
},
{
"dataType": [
"number"
],
"description": "This property was generated by Weaviate's auto-schema feature on Fri Dec 29 01:00:07 2023",
"indexFilterable": true,
"indexSearchable": false,
"name": "page"
}
],
"replicationConfig": {
"factor": 1
},
"shardingConfig": {
"virtualPerPhysical": 128,
"desiredCount": 1,
"actualCount": 1,
"desiredVirtualCount": 128,
"actualVirtualCount": 128,
"key": "_id",
"strategy": "hash",
"function": "murmur3"
},
"vectorIndexConfig": {
"skip": false,
"cleanupIntervalSeconds": 300,
"maxConnections": 64,
"efConstruction": 128,
"ef": -1,
"dynamicEfMin": 100,
"dynamicEfMax": 500,
"dynamicEfFactor": 8,
"vectorCacheMaxObjects": 1000000000000,
"flatSearchCutoff": 40000,
"distance": "cosine",
"pq": {
"enabled": false,
"bitCompression": false,
"segments": 0,
"centroids": 256,
"trainingLimit": 100000,
"encoder": {
"type": "kmeans",
"distribution": "log-normal"
}
}
},
"vectorIndexType": "hnsw",
"vectorizer": "none"
}
]
}

[dosubot[bot]]

🤖

Hello @hboen1990!

I'm Dosu, a bot designed to help you with your questions and issues related to the LangChain repository. I'm here to assist you while waiting for a human maintainer. Don't worry, I'm here to help you uncover the answers to your questions and navigate through any bugs you might encounter. Let's get started, shall we?

To use the persisted data in Weaviate to create a vectorstore for RAG without re-ingesting the documents, you can use the Weaviate class provided in the weaviate.py file.

Here is an example of how you can do this:

import weaviate
from langchain.vectorstores import Weaviate

client = weaviate.Client(url=os.environ["WEAVIATE_URL"], ...)
weaviate_vectorstore = Weaviate(client, index_name, text_key)

In this example, index_name is the name of the index where your persisted data is stored in Weaviate, and text_key is the key used for uploading/retrieving text to/from the vectorstore.

Once you have created the Weaviate instance, you can use its methods to perform operations on the persisted data. For example, you can use the similarity_search_by_text method to return documents most similar to a given query, or the similarity_search_by_vector method to look up similar documents by embedding vector in Weaviate. You can find more details in the weaviate.py file.

Regarding your question about RAG with Weaviate vectorstore, LangChain does support this. The provided context shows that Weaviate is used as a vectorstore to retrieve context for the RAG model.

Here is the relevant code:

# Add to vectorDB
vectorstore = Weaviate.from_existing_index(WEAVIATE_INDEX_NAME, OpenAIEmbeddings())
retriever = vectorstore.as_retriever()

# RAG prompt
template = """Answer the question based only on the following context:
{context}
Question: {question}
"""
prompt = ChatPromptTemplate.from_template(template)

# RAG
model = ChatOpenAI()
chain = (
    RunnableParallel({"context": retriever, "question": RunnablePassthrough()})
    | prompt
    | model
    | StrOutputParser()
)

You can find more details in the chain.py file.

I hope this helps! If you have any other questions, feel free to ask.

Sources

• libs/langchain/langchain/vectorstores/weaviate.py
• templates/rag-weaviate/rag_weaviate/chain.py

---

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

[hboen1990]

"Weaviate.from_existing_index" i.e. "from_existing_index"does not exists
I tried:
weaviate_vectorstore = Weaviate(client=client, index_name="WMOInfo", text_key="content",embedding=embedding)
retriever = weaviate_vectorstore.as_retriever()
response = retriever.get_relevant_documents(query="some qury")
response = [] empty...but it should not be empty, because the first time when running the load en embedding process it gives results...
Is my definition of the weaviate_vectorstore not correct?

After further testing I notice that langchain created 2 classes i.e. the class I created "WMOInfo" and additionally the class "LangChain_uuid". When I used theLangChain_uuid" class I see results after replacing the above vectorstore definition with: weaviate_vectorstore = Weaviate(client=client, index_name='LangChain_71db963cf5114df8b31ef1edde728ef3', text_key='text',embedding=embedding). It is weird that langchain completely ignore my created class.