RAG Langchain V2

A Retrieval-Augmented Generation (RAG) system built with LangChain that combines local document processing with both local and cloud-based LLM support.

Architecture

graph TD
    %% Input Documents
    A[PDF Documents] --> B[Document Loader]
    
    %% Document Processing
    B --> C[Text Splitter]
    C --> D[Vector Store<br>ChromaDB]
    
    %% Query Flow
    E[User Query] --> F[Query Processing]
    D --> F
    F --> G[Embedding Model<br>nomic-embed-text]
    G --> H[Similar Documents]
    
    %% LLM Processing
    H --> I[LLM Processing]
    I --> J[Generated Response]
    
    %% LLM Options with styling
    subgraph LLM_Models[Available LLM Models]
        direction LR
        K[Local Model:<br>deepseek-r1:1.5b]
        L[Cloud Model:<br>Gemini-1.5-flash]
    end
    
    I -.-> LLM_Models

    %% Styling
    classDef primary fill:#e1f5fe,stroke:#01579b,stroke-width:2px
    classDef secondary fill:#f3e5f5,stroke:#4a148c,stroke-width:2px
    classDef storage fill:#e8f5e9,stroke:#1b5e20,stroke-width:2px
    
    class A,E primary
    class B,C,F,G,H,I,J secondary
    class D storage
    class K,L primary

Loading

Project Structure

• chroma/ - Vector database storage
• data/ - PDF documents directory
• create_database.py - Database creation script
• query_data.py - Handles RAG queries and LLM interactions
• get_embedding_functions.py - Embedding model configuration
• constants.py - Import GEMINI_API_KEY from '.env' file
• requirements.txt - Project dependencies

Requirements

• Python 3.8+
• LangChain
• LLMs
  • Local Model: deepseek-r1:1.5b

    ollama pull deepseek-r1:1.5b

  • Cloud Model: Google's Gemini-1.5-flash
• Embedding
  • Local Embedding Model: nomic-embed-text

    ollama pull nomic-embed-text

• ChromaDB

Installation

1. Clone the git repo and change directory to the project-name using the commands below:

   git clone https://github.com/usman619/RAG_langchain_v2.git
   cd RAG_langchain_v2
   code .

2. Create of virtual environment in python using the commands below:

   python -m venv venv
   source venv/bin/activate

3. Installing the required packages using the command below:

   pip install -r requirements.txt

4. If you are using Gemini API, you can set up your environment variable:

   echo "GEMINI_API_KEY=your_api_key_here" > .env

5. Create the vector database:

   python create_database.py
   # reset your chroma db
   python create_database.py --reset

6. Run the code using the command below:

   python query_text.py "Your question here?"