basic-chatbot-langchain-ia

Chatbot with LLM Local (LM Studio, LangChain, Python)

📝 Project Description

This project presents a simple chatbot built in Python using the LangChain library for conversation orchestration. The unique and significant advantage of this chatbot is that it communicates with a Large Language Model (LLM) running locally on your machine, without the need for an internet connection for inference (after the initial model download).

The magic of running LLM locally is possible thanks to LM Studio, an intuitive tool that lets you download and serve open-source language models (like Gemma-2-2B-IT, which we're using here) directly on your computer, emulating an OpenAI-compatible API.

The main objective of this project is to demonstrate the ability to:

• Set up a local AI development environment.
• Integrate tools like LM Studio and LangChain.
• Build an interactive chatbot application via terminal.
• Explore the power of LLMs that can run on commodity hardware.

✨ Features

• Chatbot interaction via terminal.
• Using an LLM running locally ( Gemma-2-2B-IT , configured via LM Studio).
• Python virtual environment setup for dependency isolation.
• Test function to check communication with the LLM server from LM Studio.

🛠️ Technologies Used

• Python 3.9+
• LM Studio: Tool to download and run LLMs locally.
• LangChain: Framework for developing applications with LLMs.
• langchain-openai: LangChain integration with OpenAI-compatible APIs (used to connect to LM Studio).
• python-dotenv: To manage environment variables (LLM server URL).

🚀 Environment Configuration and Execution

Follow the steps below to get the chatbot running on your Windows 11 machine.

1. Installing LM Studio

1. Visit the official LM Studio website: https://lmstudio.ai/
2. Download and install the Windows version.

2. Download and Configure the Model in LM Studio

After installing LM Studio:

1. Open LM Studio.
2. In the left sidebar, click on the "Search" tab (magnifying glass icon).
3. In the search field, type gemma-2-2b-it and press Enter.
4. Look for a GGUF version of the model (e.g., gemma-2-2b-it-q4_k_m.gguf). Models with Q4_K_M or Q5_K_M offer a good balance between performance and RAM consumption.
5. Click the "Download" button next to the version you've chosen. Wait for the download to complete (it may take a while depending on your connection).

3. Starting the LLM Local Server in LM Studio

With the downloaded model:

1. In LM Studio, go to the "Local Inference Server" tab (two arrow icon, one up and one down).
2. In the left pane, make sure the template gemma-2-2b-it is selected in the dropdown.
3. Make sure the default port (1234) is set. Note the port if it is different.
4. Click the "Start Server" button.

• You will see a message like "Serving on http://localhost:1234" if the server starts correctly.
• Leave LM Studio running in the background while you configure Python.

4. Python Environment Setup

1. Clone this repository to your machine:

git clone https://github.com/SeuUsuario/NomeDoSeuRepositorio.git
cd NomeDoSeuRepositorio

(Replace SeuUsuario/NomeDoSeuRepositorio.git with the actual path to your repository).

2. Create a virtual environment (in the project root directory):

python -m venv .venv

3. Activate the virtual environment (on Windows):

.venv\Scripts\activate

You will see (.venv) at the beginning of the command line, indicating that the environment is activated.

5. Creating the file requirements.txt
   Create a file named requirements.txt in the root of your project with the following content:

langchain
langchain-community
langchain-openai
python-dotenv

6. Installing Python Dependencies

With the virtual environment activated, install the necessary libraries:

pip install -r requirements.txt

7. Creating the file .env

1. Create a file called .env in the root of your project (at the same level as main.py).
2. Add the following line to the file .env, adjusting the port if it is different from the default 1234:

LM_STUDIO_BASE_URL=http://localhost:1234/v1

8. Running the Chatbot

With LM Studio serving the model and the Python environment configured:

1. In your terminal, make sure you are in the project root folder and that the virtual environment is activated.
2. Run the main script:

python main.py

Interaction

• The chatbot will first attempt to connect to the LLM. If the connection is successful, it will begin the conversation loop.
• Type your questions and press Enter.
• To exit, type exit and press Enter.

📁 Project Structure

.
├── .env                  # Variáveis de ambiente (URL do servidor LLM)
├── main.py               # Script principal do chatbot
├── requirements.txt      # Lista de dependências Python
└── README.md             # Este arquivo