S3 Workspaces

[Josh-XT]

The goal in this PR is to switch from locally persisted agent workspaces to putting them in S3 storage. Anyone deploying locally will be able to use a minio container, which will be added to the docker-compose files for anyone running locally. This PR will effectively enable deploying AGiXT on cloud services hosts such as AWS or Digital Ocean without the need to persist files.

This pull request introduces several significant updates to the agixt project, focusing on Docker configuration, agent management, and application initialization. The changes include modifications to Docker-related files, enhancements to agent handling, and improvements to the application startup process.

Docker Configuration Updates:

• Dockerfile: Added a step to run Transcription.py during the build process and updated the entry point to use DB.py.
• docker-compose-dev.yml and docker-compose-nostreamlit-dev.yml: Updated volumes and environment variables for better configuration management, including the addition of MinIO service for storage. [1] [2] [3] [4]
• docker-compose-nostreamlit.yml: Added MinIO service configuration for non-streamlit environments.

Agent Management Enhancements:

• agixt/Agent.py: Improved agent name handling to avoid duplicates by appending a number if the agent already exists.

Application Initialization Improvements:

• agixt/app.py: Refactored the application startup to include workspace management, task monitoring, and ngrok tunnel setup. Added signal handlers for graceful shutdown and a new route for serving files from the workspace. [1] [2] [3] [4] [5]

Miscellaneous Changes:

• agixt/Globals.py: Enhanced the getenv function to accept a default value parameter. [1] [2]
• agixt/Transcription.py and agixt/providers/default.py: Updated the transcription model initialization to use int8 compute type for improved performance. [1] [2]
• Removed the obsolete launch-backend.sh script as its functionality has been integrated into the main application.

These changes collectively improve the robustness, configurability, and maintainability of the agixt project.