kpkpkp

I’m new to langchain and sought it out specifically to deal with large chunks of text and how they can be pieces of an even larger document. So I was pretty excited to see your example. But when I try it (https://langchain-ask-the-doc.streamlit.app/?ref=blog.streamlit.io) I get a SQLite error.

Is there a workaround? Am I missing something important? Thanks!

The SQLite error:
RuntimeError: e[91mYour system has an unsupported version of sqlite3. Chroma requires sqlite3 >= 3.35.0.e[0me[94mPlease visit 🔍 Troubleshooting | Chroma to learn how to upgrade.e[0m

2 replies

dataprofessor Developer Advocate at Streamlit

Hi,

There’s a similar post below along with solution that you can try out:

https://discuss.streamlit.io/t/issues-with-chroma-and-sqlite/47950/12

DataSetReady

You may have already solved this so I will post a fix that worked for me that may benefit anyone else running into this problem. Installing the pysqlite3 package solves the problem with mismatched Sqlite3 versions and Chromadb. See the pysqlite3 repository for further details on the package itself.

To implement:

1. add pysqlite3-binary to requirements.txt
2. add these to your app file:

import pysqlite3
import sys
sys.modules['sqlite3'] = sys.modules.pop('pysqlite3')
import streamlit as st
import sqlite3

3. Reboot your app in Streamlit.

aitech

@dataprofessor Which tool you are using for diagram above?

1 reply

dataprofessor Developer Advocate at Streamlit

Hi @aitech

The diagrams were hand drawn in the Notability app on an iPad.

Basit-Arif

there is a problem, when we run the code again with the same pdf its append its embedding in chromadb where as it should not store that , to check use vectorestore._collection.count()