Redis VSS Documentation    Redis VL Documentation   

With this demo application, we use a pre-labelled set of transactions train.csv to predict the categories of a seperate transaction log test.csv. This usecase is specially useful if your fintech application needs to build a expense tracker or categorizer. This approach also avoids the task of finetuning the AI models and allows you to use off the shelf pre-trained model. This demo uses the Open AI text embedding model, you need a functioning Open AI API key to test this application

In this particular demo application, we use the library redisvl which is a python library helping you to use the redis vector database functionality in a hassle free manner.

The easiest way to is to use a docker image using the below command

docker run -d -p 6379:6379 -p 8001:8001 redis/redis-stack:latest

If you do not want to use a docker image, you can sign up for a free Redis Cloud subscription here.

Download the repository

git clone https://github.com/mar1boroman/ExpenseFlow.git && cd ExpenseFlow

Prepare and activate the virtual environment

python3 -m venv venv && source venv/bin/activate

Install necessary libraries and dependencies

pip install -r requirements.txt

Configure your OPEN AI Key in the .env file

vi .env

streamlit run ui/0_📎_Upload_Pre_Labelled_Data.py

In the first screen upload the train.csv file to load the embeddings of pre-labelled data

In the 🔮_Predict_Categories screen load the test.csv file to predict the category of every transaction and show an aggregated view of the expense

If you want to check a sample step by step exection (behind the scenes view), the third screen allows you to enter a single transaction description manually and see how the category of the transaction is predicted.