PrivateGPT in SQL with EvaDB

This project was inspired by the original privateGPT and localGPT projects.Most of the description is derived from these projects.

In this project, we use EvaDB to replicate the privateGPT workflow using SQL-like queries. Here are the key SQL queries:

    --- Generate feature embeddings of document chunks
    CREATE TABLE IF NOT EXISTS 
    embedding_table AS SELECT embedding(data), data FROM data_table;

    --- Run a similarity search query based on the feature embeddings
    --- Uses a feature index for faster retrieval of relevant chunks
    SELECT data 
    FROM embedding_table 
    ORDER BY Similarity(embedding("Why was NATO created?"),features) 
    ASC LIMIT 3

We plan to add these two features in the coming week:

1. Besides PDF documents, EvaDB also supports search over text documents, images, and videos.
2. Improving the quality of answers by building a hierarchical index over documents at different granularities (sentences, paragraphs, and entire documents).

You can ask questions on your documents without an internet connection using the power of LLMs. 100% private, no data leaves your execution environment at any point.

Built with EvaDB, LangChain, and GPT4ALL.

Environment Setup

To use this software, you must have Python 3.8 or later installed.

To set up your environment, install all the required dependencies by running the following command:

pip install -r requirements.txt

Ask questions to your documents, locally!

Ingest all PDF documents by executing the following command:

python ingest.py

This command will load the PDF files located in the source_documents folder into the EvaDB and build an index on it.

To ask questions to your documents locally, use the following command:

python privateGPT.py

Once you run this command, the script will prompt you for input.

> Enter your question: Why was NATO created?

> Answer:
The purpose of NATO was to secure peace and stability in Europe after World War 2. To accomplish this, American ground forces, air squadrons, and ship deployments were mobilized to protect NATO countries, including Poland, Romania, Latvia, Lithuania, and Estonia. Additionally, a coalition of other freedom-loving nations from Europe, Asia, and Africa was formed to confront Putin.

To exit the script, simply type exit.

If you face any issues, please create an issue on Github or ping us on Slack.