HandsOnOllama

HandsOnOllama is a Retrieval-Augmented Generation (RAG) document assistant application built using LangChain and Ollama. It leverages a vector database for efficient document retrieval and a language model for generating context-aware responses.

Features

• Document Ingestion: Ingests PDF documents and splits them into manageable chunks for processing.
• Vector Database: Uses Chroma for storing and retrieving document embeddings.
• Language Model Integration: Employs Ollama for generating embeddings and responses.
• Streamlit UI: Provides an interactive user interface for querying the document assistant.
• Robust Error Handling: All modules include improved error handling and input validation.
• Coding Standards: The codebase follows PEP 8, Google Python Style Guide, and project-specific guidelines for maintainability and clarity.
• Modular Python Package: All core logic is organized under the doc_assist package for maintainability and reusability.

Installation

1. Clone the repository:

   git clone https://github.com/mascarenhasneil/HandsOnOllama.git

2. Navigate to the project directory:

   cd HandsOnOllama

3. Create and activate a virtual environment:

   python -m venv .venv
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate

4. Install the required dependencies:

   pip install -r requirements.txt

Usage

1. Start the Streamlit application:

   streamlit run main.py

   or

   streamlit run doc_assist/rag_streamlit.py

2. Upload a PDF document and interact with the assistant through the UI.

   • The sidebar allows you to upload a PDF file (stored in the eggs directory).
   • The app will create or load a vector database for the uploaded file.
   • Ask questions about the document using the input box.
   • Use the sidebar buttons to stop or close the app with user feedback.

Project Structure

• doc_assist/chain.py: Handles the creation of the RAG processing chain. Includes input validation and Google-style docstrings.
• doc_assist/retriever.py: Creates a multi-query retriever for document retrieval. Includes input validation and usage examples.
• doc_assist/vector_db.py: Manages the vector database for storing and retrieving embeddings. Features robust error handling and clear docstrings.
• doc_assist/utils.py: Utility functions for document ingestion and processing. Includes error handling and usage examples.
• doc_assist/rag_streamlit.py: Streamlit application for the document assistant. Provides user feedback and robust error handling in the UI. Main entry point is rag_streamlit().
• main.py: Entry point for the app, calls doc_assist.rag_streamlit.rag_streamlit().
• model_test_1.py: Demonstrates how to send a JSON POST request to a local API endpoint to generate text using a specified language model.
• model_test_2.py: Demonstrates the use of the Ollama API to interact with language models, including listing models, chatting, generating text, and creating custom models.

Coding Guidelines

• All code must pass Pylance and mypy checks (static type checking and linting).
• Use absolute imports, group imports by standard library, third-party, and local modules.
• Avoid wildcard imports and place all imports at the top of the file after the module docstring.
• Use Google-style docstrings for all public modules, classes, functions, and methods, including usage examples.
• Document all arguments, return values, and exceptions.
• Use type hints for all public APIs.
• Follow PEP 8 and Google Python Style Guide for naming, formatting, and structure.

Contributing

Contributions are welcome! Please follow the coding guidelines and ensure all changes are well-documented. Run static analysis tools and tests before submitting a pull request.

License

This project is licensed under the MIT License. See the LICENSE file for details.