README.md

GAAS (Generative AI Agent Services) Python Tools

This directory contains documentation for the Python tools used by the Generative AI Agent Services (GAAS) within the SEMOSS platform. These tools provide various functionalities that enable AI agents to interact with data, execute code, and access knowledge.

Overview

The GAAS tools are designed to be modular components that can be invoked by an AI agent or orchestration layer to perform specific tasks. They often leverage the genai_client package for interactions with large language models and other AI services.

Available Tools and Components

The following documents detail the specific GAAS tools and server components:

• Database Interaction (gaas_gpt_database.py): Describes the DatabaseEngine class, a Python proxy for executing queries and data operations on SEMOSS database engines.
• Model Interaction (gaas_gpt_model.py): Details the core component for interacting with generative models for various tasks.
• Function Execution (gaas_gpt_function.py): Details the FunctionEngine class, a Python proxy for executing pre-defined SEMOSS FUNCTION engines.
• Storage Access (gaas_gpt_storage.py): Details the StorageEngine class, a Python proxy for performing operations on SEMOSS STORAGE engines.
• Vector Database (gaas_gpt_vector.py): Details the VectorEngine class, a Python proxy for adding documents to and performing similarity searches on SEMOSS VECTOR engines.

Additional GAAS components include:

• Server Proxy (gaas_server_proxy.py): Enables Python GAAS components to call back to the SEMOSS Java backend.
• TCP Servers (gaas_tcp_server_handler.py, gaas_tcp_socket_server.py): Components for TCP-based communication.

Integration

These tools are typically orchestrated by a higher-level agent or service that determines which tool to use based on the user's request or a predefined workflow. Understanding these tools is key to extending the capabilities of SEMOSS's AI agents.