Couchbase is an award-winning distributed NoSQL cloud database that delivers unmatched versatility, performance, scalability, and financial value for all of your cloud, mobile, AI, and edge computing applications. Couchbase embraces AI with coding assistance for developers and vector search for their applications. Vector Search is a part of the Full Text Search Service (Search Service) in Couchbase. This tutorial explains how to use Vector Search in Couchbase. You can work with either Couchbase Capella and your self-managed Couchbase Server.

​
Setup

To access the CouchbaseSearchVectorStore you first need to install the langchain-couchbase partner package: 
pip install -qU langchain-couchbase

​
Credentials

Head over to the Couchbase website and create a new connection, making sure to save your database username and password: 
import getpass

COUCHBASE_CONNECTION_STRING = getpass.getpass(
    "Enter the connection string for the Couchbase cluster: "
)
DB_USERNAME = getpass.getpass("Enter the username for the Couchbase cluster: ")
DB_PASSWORD = getpass.getpass("Enter the password for the Couchbase cluster: ")

Enter the connection string for the Couchbase cluster:  ········
Enter the username for the Couchbase cluster:  ········
Enter the password for the Couchbase cluster:  ········
If you want to get best in-class automated tracing of your model calls you can also set your LangSmith API key by uncommenting below: 
# os.environ["LANGSMITH_TRACING"] = "true"
# os.environ["LANGSMITH_API_KEY"] = getpass.getpass()

​
Initialization

Before instantiating we need to create a connection.

​
Create Couchbase Connection Object

We create a connection to the Couchbase cluster initially and then pass the cluster object to the Vector Store. Here, we are connecting using the username and password from above. You can also connect using any other supported way to your cluster. For more information on connecting to the Couchbase cluster, please check the documentation. 
from datetime import timedelta

from couchbase.auth import PasswordAuthenticator
from couchbase.cluster import Cluster
from couchbase.options import ClusterOptions

auth = PasswordAuthenticator(DB_USERNAME, DB_PASSWORD)
options = ClusterOptions(auth)
cluster = Cluster(COUCHBASE_CONNECTION_STRING, options)

# Wait until the cluster is ready for use.
cluster.wait_until_ready(timedelta(seconds=5))
We will now set the bucket, scope, and collection names in the Couchbase cluster that we want to use for Vector Search. For this example, we are using the default scope & collections. 
BUCKET_NAME = "langchain_bucket"
SCOPE_NAME = "_default"
COLLECTION_NAME = "_default"
SEARCH_INDEX_NAME = "langchain-test-index"
For details on how to create a Search index with support for Vector fields, please refer to the documentation.

​
Simple Instantiation

Below, we create the vector store object with the cluster information and the search index name. 
# | output: false
# | echo: false
from langchain_openai import OpenAIEmbeddings

embeddings = OpenAIEmbeddings(model="text-embedding-3-large")

from langchain_couchbase.vectorstores import CouchbaseSearchVectorStore

vector_store = CouchbaseSearchVectorStore(
    cluster=cluster,
    bucket_name=BUCKET_NAME,
    scope_name=SCOPE_NAME,
    collection_name=COLLECTION_NAME,
    embedding=embeddings,
    index_name=SEARCH_INDEX_NAME,
)

​
Specify the Text & Embeddings Field

You can optionally specify the text & embeddings field for the document using the text_key and embedding_key fields. 
vector_store_specific = CouchbaseSearchVectorStore(
    cluster=cluster,
    bucket_name=BUCKET_NAME,
    scope_name=SCOPE_NAME,
    collection_name=COLLECTION_NAME,
    embedding=embeddings,
    index_name=SEARCH_INDEX_NAME,
    text_key="text",
    embedding_key="embedding",
)

​
Manage vector store

Once you have created your vector store, we can interact with it by adding and deleting different items.

​
Add items to vector store

We can add items to our vector store by using the add_documents function. 
from uuid import uuid4

from langchain_core.documents import Document

document_1 = Document(
    page_content="I had chocolate chip pancakes and scrambled eggs for breakfast this morning.",
    metadata={"source": "tweet"},
)

document_2 = Document(
    page_content="The weather forecast for tomorrow is cloudy and overcast, with a high of 62 degrees.",
    metadata={"source": "news"},
)

document_3 = Document(
    page_content="Building an exciting new project with LangChain - come check it out!",
    metadata={"source": "tweet"},
)

document_4 = Document(
    page_content="Robbers broke into the city bank and stole $1 million in cash.",
    metadata={"source": "news"},
)

document_5 = Document(
    page_content="Wow! That was an amazing movie. I can't wait to see it again.",
    metadata={"source": "tweet"},
)

document_6 = Document(
    page_content="Is the new iPhone worth the price? Read this review to find out.",
    metadata={"source": "website"},
)

document_7 = Document(
    page_content="The top 10 soccer players in the world right now.",
    metadata={"source": "website"},
)

document_8 = Document(
    page_content="LangGraph is the best framework for building stateful, agentic applications!",
    metadata={"source": "tweet"},
)

document_9 = Document(
    page_content="The stock market is down 500 points today due to fears of a recession.",
    metadata={"source": "news"},
)

document_10 = Document(
    page_content="I have a bad feeling I am going to get deleted :(",
    metadata={"source": "tweet"},
)

documents = [
    document_1,
    document_2,
    document_3,
    document_4,
    document_5,
    document_6,
    document_7,
    document_8,
    document_9,
    document_10,
]
uuids = [str(uuid4()) for _ in range(len(documents))]

vector_store.add_documents(documents=documents, ids=uuids)

['f125b836-f555-4449-98dc-cbda4e77ae3f',
 'a28fccde-fd32-4775-9ca8-6cdb22ca7031',
 'b1037c4b-947f-497f-84db-63a4def5080b',
 'c7082b74-b385-4c4b-bbe5-0740909c01db',
 'a7e31f62-13a5-4109-b881-8631aff7d46c',
 '9fcc2894-fdb1-41bd-9a93-8547747650f4',
 'a5b0632d-abaf-4802-99b3-df6b6c99be29',
 '0475592e-4b7f-425d-91fd-ac2459d48a36',
 '94c6db4e-ba07-43ff-aa96-3a5d577db43a',
 'd21c7feb-ad47-4e7d-84c5-785afb189160']

​
Delete items from vector store

vector_store.delete(ids=[uuids[-1]])

True

​
Query vector store

Once your vector store has been created and the relevant documents have been added you will most likely wish to query it during the running of your chain or agent.

​
Query directly

Performing a simple similarity search can be done as follows: 
results = vector_store.similarity_search(
    "LangChain provides abstractions to make working with LLMs easy",
    k=2,
)
for res in results:
    print(f"* {res.page_content} [{res.metadata}]")

* Building an exciting new project with LangChain - come check it out! [{'source': 'tweet'}]
* LangGraph is the best framework for building stateful, agentic applications! [{'source': 'tweet'}]

​
Similarity search with Score

You can also fetch the scores for the results by calling the similarity_search_with_score method. 
results = vector_store.similarity_search_with_score("Will it be hot tomorrow?", k=1)
for res, score in results:
    print(f"* [SIM={score:3f}] {res.page_content} [{res.metadata}]")

* [SIM=0.553112] The weather forecast for tomorrow is cloudy and overcast, with a high of 62 degrees. [{'source': 'news'}]

​
Filtering Results

You can filter the search results by specifying any filter on the text or metadata in the document that is supported by the Couchbase Search service. The filter can be any valid SearchQuery supported by the Couchbase Python SDK. These filters are applied before the Vector Search is performed. If you want to filter on one of the fields in the metadata, you need to specify it using . For example, to fetch the source field in the metadata, you need to specify metadata.source. Note that the filter needs to be supported by the Search Index. 
from couchbase import search

query = "Are there any concerning financial news?"
filter_on_source = search.MatchQuery("news", field="metadata.source")
results = vector_store.similarity_search_with_score(
    query, fields=["metadata.source"], filter=filter_on_source, k=5
)
for res, score in results:
    print(f"* {res.page_content} [{res.metadata}] {score}")

* The stock market is down 500 points today due to fears of a recession. [{'source': 'news'}] 0.3873019218444824
* Robbers broke into the city bank and stole $1 million in cash. [{'source': 'news'}] 0.20637212693691254
* The weather forecast for tomorrow is cloudy and overcast, with a high of 62 degrees. [{'source': 'news'}] 0.10404900461435318

​
Specifying Fields to Return

You can specify the fields to return from the document using fields parameter in the searches. These fields are returned as part of the metadata object in the returned Document. You can fetch any field that is stored in the Search index. The text_key of the document is returned as part of the document’s page_content. If you do not specify any fields to be fetched, all the fields stored in the index are returned. If you want to fetch one of the fields in the metadata, you need to specify it using . For example, to fetch the source field in the metadata, you need to specify metadata.source. 
query = "What did I eat for breakfast today?"
results = vector_store.similarity_search(query, fields=["metadata.source"])
print(results[0])

page_content='I had chocolate chip pancakes and scrambled eggs for breakfast this morning.' metadata={'source': 'tweet'}

​
Query by turning into retriever

You can also transform the vector store into a retriever for easier usage in your chains. Here is how to transform your vector store into a retriever and then invoke the retreiever with a simple query and filter. 
retriever = vector_store.as_retriever(
    search_type="similarity",
    search_kwargs={"k": 1, "score_threshold": 0.5},
)
filter_on_source = search.MatchQuery("news", field="metadata.source")
retriever.invoke("Stealing from the bank is a crime", filter=filter_on_source)

[Document(id='c7082b74-b385-4c4b-bbe5-0740909c01db', metadata={'source': 'news'}, page_content='Robbers broke into the city bank and stole $1 million in cash.')]

​
Hybrid Queries

Couchbase allows you to do hybrid searches by combining Vector Search results with searches on non-vector fields of the document like the metadata object. The results will be based on the combination of the results from both Vector Search and the searches supported by Search Service. The scores of each of the component searches are added up to get the total score of the result. To perform hybrid searches, there is an optional parameter, search_options that can be passed to all the similarity searches.
The different search/query possibilities for the search_options can be found here. In order to simulate hybrid search, let us create some random metadata from the existing documents. We uniformly add three fields to the metadata, date between 2010 & 2020, rating between 1 & 5 and author set to either John Doe or Jane Doe. 
from langchain_community.document_loaders import TextLoader
from langchain_text_splitters import CharacterTextSplitter

loader = TextLoader("../../how_to/state_of_the_union.txt")
documents = loader.load()
text_splitter = CharacterTextSplitter(chunk_size=500, chunk_overlap=0)
docs = text_splitter.split_documents(documents)

# Adding metadata to documents
for i, doc in enumerate(docs):
    doc.metadata["date"] = f"{range(2010, 2020)[i % 10]}-01-01"
    doc.metadata["rating"] = range(1, 6)[i % 5]
    doc.metadata["author"] = ["John Doe", "Jane Doe"][i % 2]

vector_store.add_documents(docs)

query = "What did the president say about Ketanji Brown Jackson"
results = vector_store.similarity_search(query)
print(results[0].metadata)

{'author': 'John Doe', 'date': '2016-01-01', 'rating': 2, 'source': '../../how_to/state_of_the_union.txt'}

​
Query by Exact Value

We can search for exact matches on a textual field like the author in the metadata object. 
query = "What did the president say about Ketanji Brown Jackson"
results = vector_store.similarity_search(
    query,
    search_options={"query": {"field": "metadata.author", "match": "John Doe"}},
    fields=["metadata.author"],
)
print(results[0])

page_content='One of the most serious constitutional responsibilities a President has is nominating someone to serve on the United States Supreme Court. 

And I did that 4 days ago, when I nominated Circuit Court of Appeals Judge Ketanji Brown Jackson. One of our nation’s top legal minds, who will continue Justice Breyer’s legacy of excellence.' metadata={'author': 'John Doe'}

​
Query by Partial Match

We can search for partial matches by specifying a fuzziness for the search. This is useful when you want to search for slight variations or misspellings of a search query. Here, “Jae” is close (fuzziness of 1) to “Jane”. 
query = "What did the president say about Ketanji Brown Jackson"
results = vector_store.similarity_search(
    query,
    search_options={
        "query": {"field": "metadata.author", "match": "Jae", "fuzziness": 1}
    },
    fields=["metadata.author"],
)
print(results[0])

page_content='A former top litigator in private practice. A former federal public defender. And from a family of public school educators and police officers. A consensus builder. Since she’s been nominated, she’s received a broad range of support—from the Fraternal Order of Police to former judges appointed by Democrats and Republicans. 

And if we are to advance liberty and justice, we need to secure the Border and fix the immigration system.' metadata={'author': 'Jane Doe'}

​
Query by Date Range Query

We can search for documents that are within a date range query on a date field like metadata.date. 
query = "Any mention about independence?"
results = vector_store.similarity_search(
    query,
    search_options={
        "query": {
            "start": "2016-12-31",
            "end": "2017-01-02",
            "inclusive_start": True,
            "inclusive_end": False,
            "field": "metadata.date",
        }
    },
)
print(results[0])

page_content='And with 75% of adult Americans fully vaccinated and hospitalizations down by 77%, most Americans can remove their masks, return to work, stay in the classroom, and move forward safely. 

We achieved this because we provided free vaccines, treatments, tests, and masks. 

Of course, continuing this costs money. 

I will soon send Congress a request. 

The vast majority of Americans have used these tools and may want to again, so I expect Congress to pass it quickly.' metadata={'author': 'Jane Doe', 'date': '2017-01-01', 'rating': 3, 'source': '../../how_to/state_of_the_union.txt'}

​
Query by Numeric Range Query

We can search for documents that are within a range for a numeric field like metadata.rating. 
query = "Any mention about independence?"
results = vector_store.similarity_search_with_score(
    query,
    search_options={
        "query": {
            "min": 3,
            "max": 5,
            "inclusive_min": True,
            "inclusive_max": True,
            "field": "metadata.rating",
        }
    },
)
print(results[0])

(Document(id='3a90405c0f5b4c09a6646259678f1f61', metadata={'author': 'John Doe', 'date': '2014-01-01', 'rating': 5, 'source': '../../how_to/state_of_the_union.txt'}, page_content='In this Capitol, generation after generation, Americans have debated great questions amid great strife, and have done great things. \n\nWe have fought for freedom, expanded liberty, defeated totalitarianism and terror. \n\nAnd built the strongest, freest, and most prosperous nation the world has ever known. \n\nNow is the hour. \n\nOur moment of responsibility. \n\nOur test of resolve and conscience, of history itself.'), 0.3573387440020518)

​
Combining Multiple Search Queries

Different search queries can be combined using AND (conjuncts) or OR (disjuncts) operators. In this example, we are checking for documents with a rating between 3 & 4 and dated between 2015 & 2018. 
query = "Any mention about independence?"
results = vector_store.similarity_search_with_score(
    query,
    search_options={
        "query": {
            "conjuncts": [
                {"min": 3, "max": 4, "inclusive_max": True, "field": "metadata.rating"},
                {"start": "2016-12-31", "end": "2017-01-02", "field": "metadata.date"},
            ]
        }
    },
)
print(results[0])

(Document(id='7115a704877a46ad94d661dd9c81cbc3', metadata={'author': 'Jane Doe', 'date': '2017-01-01', 'rating': 3, 'source': '../../how_to/state_of_the_union.txt'}, page_content='And with 75% of adult Americans fully vaccinated and hospitalizations down by 77%, most Americans can remove their masks, return to work, stay in the classroom, and move forward safely. \n\nWe achieved this because we provided free vaccines, treatments, tests, and masks. \n\nOf course, continuing this costs money. \n\nI will soon send Congress a request. \n\nThe vast majority of Americans have used these tools and may want to again, so I expect Congress to pass it quickly.'), 0.6898253780130769)
Note The hybrid search results might contain documents that do not satisfy all the search parameters. This is due to the way the scoring is calculated. The score is a sum of both the vector search score and the queries in the hybrid search. If the Vector Search score is high, the combined score will be more than the results that match all the queries in the hybrid search. To avoid such results, please use the filter parameter instead of hybrid search.

​
Combining Hybrid Search Query with Filters

Hybrid Search can be combined with filters to get the best of both hybrid search and the filters for results matching the requirements. In this example, we are checking for documents with a rating between 3 & 5 and matching the string “independence” in the text field. 
filter_text = search.MatchQuery("independence", field="text")

query = "Any mention about independence?"
results = vector_store.similarity_search_with_score(
    query,
    search_options={
        "query": {
            "min": 3,
            "max": 5,
            "inclusive_min": True,
            "inclusive_max": True,
            "field": "metadata.rating",
        }
    },
    filter=filter_text,
)

print(results[0])

(Document(id='23bb51b4e4d54a94ab0a95e72be8428c', metadata={'author': 'John Doe', 'date': '2012-01-01', 'rating': 3, 'source': '../../how_to/state_of_the_union.txt'}, page_content='And we remain clear-eyed. The Ukrainians are fighting back with pure courage. But the next few days weeks, months, will be hard on them.  \n\nPutin has unleashed violence and chaos.  But while he may make gains on the battlefield – he will pay a continuing high price over the long run. \n\nAnd a proud Ukrainian people, who have known 30 years  of independence, have repeatedly shown that they will not tolerate anyone who tries to take their country backwards.'), 0.30549919644400614)

​
Other Queries

Similarly, you can use any of the supported Query methods like Geo Distance, Polygon Search, Wildcard, Regular Expressions, etc in the search_options parameter. Please refer to the documentation for more details on the available query methods and their syntax.

​
Usage for retrieval-augmented generation

For guides on how to use this vector store for retrieval-augmented generation (RAG), see the following sections:

​
Frequently Asked Questions

​
Question: Should I create the Search index before creating the CouchbaseSearchVectorStore object?

Yes, currently you need to create the Search index before creating the CouchbaseSearchVectoreStore object.

​
Question: I am not seeing all the fields that I specified in my search results.

In Couchbase, we can only return the fields stored in the Search index. Please ensure that the field that you are trying to access in the search results is part of the Search index. One way to handle this is to index and store a document’s fields dynamically in the index.
  • In Capella, you need to go to “Advanced Mode” then under the chevron “General Settings” you can check “[X] Store Dynamic Fields” or “[X] Index Dynamic Fields”
  • In Couchbase Server, in the Index Editor (not Quick Editor) under the chevron “Advanced” you can check “[X] Store Dynamic Fields” or “[X] Index Dynamic Fields”
Note that these options will increase the size of the index. For more details on dynamic mappings, please refer to the documentation.

​
Question: I am unable to see the metadata object in my search results.

This is most likely due to the metadata field in the document not being indexed and/or stored by the Couchbase Search index. In order to index the metadata field in the document, you need to add it to the index as a child mapping. If you select to map all the fields in the mapping, you will be able to search by all metadata fields. Alternatively, to optimize the index, you can select the specific fields inside metadata object to be indexed. You can refer to the docs to learn more about indexing child mappings. Creating Child Mappings

​
Question: What is the difference between filter and search_options / hybrid queries?

Filters are pre-filters that are used to restrict the documents searched in a Search index. It is available in Couchbase Server 7.6.4 & higher. Hybrid Queries are additional search queries that can be used to tune the results being returned from the search index. Both filters and hybrid search queries have the same capabilites with slightly different syntax. Filters are SearchQuery objects while the hybrid search queries are dictionaries.

​
API reference

For detailed documentation of all CouchbaseSearchVectorStore features and configurations head to the API reference