This folder contains auxiliary scripts and modules that enhance the functionality of the LLM (Large Language Model) pipeline. It provides specialized tools that require optional dependencies, which are not included by default in the base environment.
- The tools in this folder rely on external Python libraries like
numpyandpandas, which may not be installed by default. - To ensure these tools function correctly, a
requirements.txtfile is included in this directory. - This mechanism allows tools to be added dynamically without modifying the core environment and re-building container
- During container startup, an
entrypoint.shscript will check forrequirements.txtinside/app/tools/. - The application in container environment dynamically installs missing dependencies using:
This ensures that the required libraries are installed only when the
pip install --no-cache-dir -r /app/tools/requirements.txt
tools/folder is mounted and utilized.
| File | Description |
|---|---|
toy_tools.py |
A module containing helper functions for data manipulation using numpy and pandas. |
requirements.txt |
A list of dependencies required for the tools module (currently includes numpy and pandas). |
Have a look and give a try to tools_agent present in hte package by default. It will:
- Generate a random
numpymatrix. - Summarize a sample
pandasDataFrame. Under the hood it auto-imports the mountedtoy_tools.pymodule and installs the dependencies fromrequirements.txt
- Create python files containing tools for your agents in
tools/folder, they will be available during runtime - Don't forget to add your imports to
requirements.txtso that no missing imports occur
- This directory is designed to be mounted at runtime to allow flexible tool additions without modifying the core system.
- If no dependencies are installed and
requirements.txtis missing, the tools in this folder will not function due to missing imports.