llm-blocks 0.2.0

Simple interface for creating and managing LLM chains

LLM Blocks

LLM Blocks is a Python module that helps users interact with Language Learning Model (LLM) chains. It provides a simple and flexible way to create and manage LLM chains, ensuring efficient interactions with models such as OpenAI's GPT-3.5 Turbo.

Installation

First, make sure to have Python installed. Then, to install the required dependencies for this module, run the following commands:

pip install -r requirements.txt
pip install -r requirements.dev.txt

Configuration

To run the LLM Blocks, you'll need an OpenAI API key. Store your API key in a .env file or export it as an environment variable:

For a .env file:

OPENAI_API_KEY=your_openai_api_key

For environment variables:

export OPENAI_API_KEY=your_openai_api_key

Usage

The llm_blocks folder contains the main ChatUtils class, which can be utilized to create and manage your LLM chains.

Here's a simple example of using the GenericChain class:

from llm_blocks.chat_utils import GenericChain

# Create a chain with a given template
template = "The meaning of {word} is:"
my_chain = GenericChain(template)

# Call the chain with any input you desire
response = my_chain("friendship")
print(response)

Module Structure

• exclude.toml: Configuration file to specify files or directories to exclude.
• requirements.dev.txt: Development dependencies for this module.
• requirements.txt: Main dependencies for this module.
• llm_blocks
  • chat_utils.py: Python file containing the definition of the GenericChain class and utility functions for working with LLM chains.

Logging Responses

The GenericChain class keeps a log of all interactions with the given LLM. The log is stored as a list of dictionaries and can be accessed with my_chain.logs.

Example:

for log in my_chain.logs:
    print(f'Inputs: {log["inputs"]}')
    print(f'Callback: {log["callback"]}')
    print(f'Response: {log["response"]}')
    print(f'Response Time: {log["response_time"]}')
    print('---')

Contributing

Feel free to submit pull requests, report bugs, or suggest new features through the GitHub repository. We appreciate your contributions and feedback!

License

This project is licensed under the MIT License.