Skip to content

Latest commit

 

History

History
64 lines (44 loc) · 2.42 KB

README.md

File metadata and controls

64 lines (44 loc) · 2.42 KB

document-qna-chatbot

Chatbot for asking questions about documents using open-source tools. Uses Langchain with RAG.

Background

Many tutorials on LangChain chatbots rely on using paid tokens from 3rd party providers. Our aim is to construct a simple document question answering bot using open-source tools.

The model architecture is based on a ConversationalRetrievalChain with memory.

We tested with the following data:

  • csv - health record data available from Kaggle
    Where the source column of text is 'Transcription'
  • pdf Several academic papers on using machine learning for automatic code generation.

How to run

  1. Codebase uses pipenv as environment manager. Install pipenv
    pip install pipenv

  2. Navigate to the project directory.

  3. Run command
    pipenv install --ignore-pipfile

You can also of course install the required packages with your preferred environment manager. The package list is found in Pipfile.

  1. Make sure you have a directory with the target documents that you want to ask questions about. At present, pdf and csv formats supported.

  2. Download an open-source language model. In this code, we used GPT4All.
    If you use other models,
    you must change the llm and embeddings imports to the appropriate one:
    from langchain.llms import GPT4All
    from langchain.embeddings import GPT4AllEmbeddings

Refer here for available LLMs.

  1. Create a local_config.py file with a MODEL_PATH variable pointing to your model.

  2. Navigate to your src directory and run:
    python main.py --directory '<document_path>' --file_type 'csv'

Improvements

Here are some ideas for improvements. We list them in order of priority:

  1. Performance improvement. Currently, a relatively involved question takes around 10 minutes.
    Yes, totally impractical but this is a first attempt on open-source data running on consumer-grade hardware.

  2. Document reference improvement. In the case of CSVs, the output includes the source rows of the answer,
    but there appears to be a bug with row duplication. Need to test how this works
    for multiple csv documents.

  3. More advanced pdf processing. It is practical to be able to handle PDFs with charts and tables.