ragbinary

How to run with your prompt:

python query_data.py '<your prompt>', it will give your k (configurable) relevant items from DB and answer your prompts

Usage:

usage: query_data.py [-h] [--query_bulk QUERY_BULK] [--query_bulk_prefix QUERY_BULK_PREFIX] [--query_bulk_suffix QUERY_BULK_SUFFIX] [--query_text QUERY_TEXT] [--output OUTPUT]

Queries the Chrome Database to Check for similarity

options:
  -h, --help            show this help message and exit
  --query_bulk QUERY_BULK
                        Provide the CSV filename to query in Bulk
  --query_bulk_prefix QUERY_BULK_PREFIX
                        The prefix to add to each query from the csv
  --query_bulk_suffix QUERY_BULK_SUFFIX
                        The suffix to add to each query from the csv
  --query_text QUERY_TEXT
                        The query text.
  --output OUTPUT       The output filename to store the results to.

NOTE: Query Bulk requires the field with in the CSV to be query_text

Others:

create_database.py is for building our VectorDB using Chroma for RAG

updated with contextual RAG

You can check langchain API in langchain document.

You need a export OPENAI_KEY_API=<look at messages in our slack> for getting access to GPT