James Colvin jamesc127

Hybrid Search Example

Applications can filter users in a geographic area by filtering first on tag information, and then using an ANN query with geospatial vector embeddings

CREATE TABLE users ( 
id UUID PRIMARY KEY, 
state text, 
city int,
tags set<text>,
sem_vec vector<float, 384> 
);

Here is an example of several different kinds of C* tombstones. For reference, they are:

Partition
Row
Cell
Range

Intro

The obvious culprit of tombstone creation is DELETE, but there are other - less obvious - sources of the tombstone. Let’s see exacly what happens on disk when a tombstone is created. It's funny to say that a tombstone is created... aren't we deleting things? Remember, everything in C* is a write! A DELETE operation writes another sstable entry with a newer timestamp than all the other entries... and the most recent timestamp wins!

I’ve created a DataStax Studio notebook that complements this gist. You should be able to download it and run it for yourself :-) nodetool and sstabledump commands need to be run from a terminal on the node(s) you're working with.

	# pip install langserve langchain openai uvicorn fastapi bs4 duckduckgo-search sse_starlette
	from langchain.chat_models import ChatOpenAI
	from langchain.prompts import ChatPromptTemplate
	from langchain.schema.output_parser import StrOutputParser
	import requests
	from bs4 import BeautifulSoup
	from langchain.schema.runnable import RunnablePassthrough, RunnableLambda
	from langchain.utilities import DuckDuckGoSearchAPIWrapper
	import json
	import os

	from cassandra.cluster import Cluster
	from cassandra.auth import PlainTextAuthProvider
	from langchain.memory import CassandraChatMessageHistory, ConversationBufferMemory
	from langchain.llms import OpenAI
	from langchain import LLMChain, PromptTemplate
	import json

	cloud_config= {
	'secure_connect_bundle': 'secure-connect-choose-your-own-adventure.zip'
	}