Skip to main content


Infinispan is an open-source key-value data grid, it can work as single node as well as distributed.

Vector search is supported since release 15.x For more: Infinispan Home

# Ensure that all we need is installed
# You may want to skip this
%pip install sentence-transformers
%pip install langchain
%pip install langchain_core
%pip install langchain_community


To run this demo we need a running Infinispan instance without authentication and a data file. In the next three cells weโ€™re going to: - download the data file - create the configuration - run Infinispan in docker

#get an archive of news
#create infinispan configuration file
echo 'infinispan:
name: default
cluster: cluster
stack: tcp
name: public
default-interface: public
port-offset: 0
name: default
port: 11222
socket-binding: default
' > infinispan-noauth.yaml
!docker rm --force infinispanvs-demo
!docker run -d --name infinispanvs-demo -v $(pwd):/user-config -p 11222:11222 infinispan/server:15.0 -c /user-config/infinispan-noauth.yaml

The Code

Pick up an embedding modelโ€‹

In this demo weโ€™re using a HuggingFace embedding mode.

from langchain.embeddings import HuggingFaceEmbeddings
from langchain_core.embeddings import Embeddings

model_name = "sentence-transformers/all-MiniLM-L12-v2"
hf = HuggingFaceEmbeddings(model_name=model_name)

Setup Infinispan cacheโ€‹

Infinispan is a very flexible key-value store, it can store raw bits as well as complex data type. User has complete freedom in the datagrid configuration, but for simple data type everything is automatically configured by the python layer. We take advantage of this feature so we can focus on our application.

Prepare the dataโ€‹

In this demo we rely on the default configuration, thus texts, metadatas and vectors in the same cache, but other options are possible: i.e.ย content can be store somewhere else and vector store could contain only a reference to the actual content.

import csv
import gzip
import time

# Open the news file and process it as a csv
with"bbc_news.csv.gz", "rt", newline="") as csvfile:
spamreader = csv.reader(csvfile, delimiter=",", quotechar='"')
i = 0
texts = []
metas = []
embeds = []
for row in spamreader:
# first and fifth values are joined to form the content
# to be processed
text = row[0] + "." + row[4]
# Store text and title as metadata
meta = {"text": row[4], "title": row[0]}
i = i + 1
# Change this to change the number of news you want to load
if i >= 5000:

Populate the vector store

# add texts and fill vector db

from langchain_community.vectorstores import InfinispanVS

ispnvs = InfinispanVS.from_texts(texts, hf, metas)

An helper func that prints the result documents

By default InfinispanVS returns the protobuf ลงext field in the Document.page_content and all the remaining protobuf fields (except the vector) in the metadata. This behaviour is configurable via lambda functions at setup.

def print_docs(docs):
for res, i in zip(docs, range(len(docs))):
print("----" + str(i + 1) + "----")
print("TITLE: " + res.metadata["title"])

Try it!!!

Below some sample queries

docs = ispnvs.similarity_search("European nations", 5)
print_docs(ispnvs.similarity_search("Milan fashion week begins", 2))
print_docs(ispnvs.similarity_search("Stock market is rising today", 4))
print_docs(ispnvs.similarity_search("Why cats are so viral?", 2))
print_docs(ispnvs.similarity_search("How to stay young", 5))
!docker rm --force infinispanvs-demo

Help us out by providing feedback on this documentation page: