0.0.43 Patches, bug fixes, and new tooling

Release 0.0.43: Applies several security and bug fixes, adds and restructures existing CLI tools

Version 0.0.42: OpenAI API improvements and bug fixes

v0.0.42

Version 0.0.42: OpenAI API improvements and bug fixes

Release 0.0.41: Enhanced CLI and Hugging Face Hub Interactions

Description

Features

• Introduced comprehensive and user-friendly CLI documentation focusing on Hugging Face Hub interactions.
• Standardized terminology to centralize on the Hugging Face Hub, enhancing its broader applicability.
• Extended Options section for detailed contexts in which each CLI option is required.

Improvements

• Updated the Prerequisites section with a relative link for installation documentation.
• Included a Table of Contents for better navigation within the documentation.
• Added a "What's Required and When" sub-section to better clarify the necessity of each CLI option.
• Removed download.py and upload.py CLI tools in favor of the unified huggingface_hub.py tool, streamlining interactions with the Hugging Face Hub.

Maintenance

• Bumped version to 0.0.41 for package and module metadata.

Prototype-v0.0.38

Prototype v0.0.38

Highlights

• JSON Handling Overhaul: A dedicated package for all JSON-related functionalities, offering more robust serialization and deserialization.

Details

Code Structure

• Isolated JSON Modules: All JSON related components are now part of their own package.
• Refactored JSON Components: Code components have been restructured for enhanced maintainability and functionality.

Unit Tests

• Renamed and Refactored Tests: The unit tests are now more focused and have been renamed to reflect their true purpose.

Additional

• File Renames: Changed the name of certain test files for better clarity.
• Version Bump: Increased from 0.0.37 to 0.0.38.

Prototype-v0.0.37

Prototype-v0.0.37: Fixes for Augmented Memory Management, addressing issues from v0.0.36

Version 0.0.37 Summary of Key Changes:

1. PEP 655 Compliance: The source code for custom types handling messages now adheres to PEP 655 guidelines. This ensures compatibility and proper handling of messages in the application.

2. FunctionCall Object: Introduced the FunctionCall object to represent function calls, including their name and optional JSON Schema representing arguments.

3. OpenAI Model Integration: In the OpenAIModel class, updates were made to correctly process and format function calls and their results, ensuring compatibility with the OpenAI API.

4. Token Counting: The TokenManager now correctly counts tokens in a message, including function calls, ensuring accurate token management.

5. Session Management: The SessionManager class has been updated to handle and enqueue function calls and their results, maintaining proper message order.

6. Function Factory: The FunctionFactory class now handles function arguments and execution more robustly, considering function calls within messages.

7. CLI Improvements: Enhancements were made to the command-line interface (chat.py) to provide better interactive chat functionality, including input prompts and handling function calls.

8. Augmented Memory Management: Augmented Memory Management is integrated into the CLI, allowing users to enable or disable this feature.

9. Version Bump: The application version has been bumped to 0.0.37 to reflect the latest changes and fixes.

These changes collectively improve the functionality and usability of the application, especially in handling function calls and managing messages in an interactive chat environment.

Prototype 0.0.36

Prototype 0.0.36: Augmented Memory Management integration in CLI.

Summary:

• Module Addition: Added a new module, pygptprompt/function/memory.py, to support Augmented Memory Management for Large Language Models.
• Function Definitions: Included JSON Schemas describing functions for managing Augmented Memories.
• Class Addition: Introduced the AugmentedMemoryManager class in the pygptprompt/function/memory.py module. This class registers functions with the FunctionFactory and associates schemas with the ConfigurationManager.
• CLI Script Update: Updated the chat.py CLI script to accommodate augmented episodic memory with a new --memory (or -m) flag, replacing the previous --embed flag. The script now uses the AugmentedMemoryManager to register functions and manage memory.

This release enhances the functionality of the project by adding support for augmented episodic memory, making it easier for the model to manage its own memories and interactions.

Prototype 0.0.35

Prototype 0.0.35: Comprehensive updates to memory management, vector database integration, and overall refinements.

PyGPTPrompt v0.0.34

Release Title: PyGPTPrompt v0.0.34 - Enhanced Functionality and Clarity

Release Body:
I am excited to announce the release of PyGPTPrompt version 0.0.34, featuring several improvements to enhance functionality and code clarity. Here's a summary of the key changes in this release:

• Version Update (v0.0.34): I've updated the version number in 'init.py' and 'pyproject.toml' to reflect version 0.0.34.

• Chat Script Refactoring: The chat script has undergone significant refactoring. I've introduced the FunctionManager and SessionManager to improve organization and clarity. This change simplifies context and function handling, making it easier to manage complex chat interactions.

• Logging Enhancement: The 'quantize.py' script now uses an improved logging mechanism, resulting in better script clarity and more informative logs.

• Package Version Updates: I've updated various package versions, including 'selenium,' 'huggingface_hub,' 'chromadb,' 'llama-cpp-python,' 'openai,' and 'gguf.' These updates ensure compatibility with the latest libraries and dependencies.

• Parameter Management and TypeDict Handling: Parameter management has been enhanced with explicit parameter naming for improved readability. I've also addressed message handling for TypeDict objects, improving code readability and message processing in sequence managers.

• FunctionManager Introduction: I've added the FunctionManager class, which is responsible for processing function responses. This manager facilitates the execution of functions, generation of function response messages, and their integration into the session manager. It introduces support for handling and processing function output in the chat model.

• Improved Code Organization: Several refactoring efforts have improved code organization, system message handling, and message output flexibility. These changes enhance code maintainability and readability.

• SessionManager Implementation: I've introduced the SessionManager class, streamlining the management of chat sessions by handling both ContextWindow and Transcript managers. This addition simplifies session handling logic and provides a cleaner API for managing chat sessions.

• PDFProcessor Refactoring: The PDFProcessor class has been refactored for chat model-based text chunking, improving its functionality and efficiency.

• Documentation Improvements: I've updated documentation across various modules, including logger configuration, class docstrings, UML diagrams, and type definitions, to enhance code readability and maintainability.

• Version Bump (v0.0.33): A minor version bump to 0.0.33 was also included in this release to reflect the latest changes and updates.

I encourage all users to update to PyGPTPrompt v0.0.34 to take advantage of these enhancements. Thank you for your continued support and contributions to the PyGPTPrompt project. If you encounter any issues or have suggestions for further improvements, please don't hesitate to reach out and share your feedback.

Prototype-v0.0.31

Version 0.0.31: Core Enhancements and Bug Fixes

I'm pleased to introduce PyGPTPrompt version 0.0.31, which focuses on core enhancements and bug fixes to elevate your AI-assisted experience. While the previous release, 0.0.30, brought expanded functionality, this update dives deep into improving the core structure and addressing issues.

Key Highlights:

1. Bug Fixes: Various bugs and issues have been addressed, enhancing the overall stability and reliability of PyGPTPrompt. These fixes ensure a smoother and error-free user experience.

2. Vector Database Abstraction: The vector database has been abstracted into a class for chat loops. This abstraction simplifies and streamlines message sequence management, making it more efficient and user-friendly.

3. Preparation for Manual Embeddings: The groundwork has been laid for manually adding embeddings to the vector store. This feature opens up possibilities for customizing AI interactions and extends the versatility of PyGPTPrompt.

Upgrade Notes:

• Tag Release: To mark these core updates and improvements, a new tag, v0.0.31, has been created.

• Continued Test Coverage: Commitment to robust performance continues with ongoing updates and improvements to test cases for OpenAI and Llama.Cpp.

• Code Cleanliness: Redundant code and comments have been removed, ensuring a clean and maintainable codebase.

This release reinforces your dedication to providing a reliable and efficient AI-assisted tool. While the previous version expanded functionality, 0.0.31 focuses on enhancing the core for a smoother user experience.

Please refer to the documentation for detailed information on these core enhancements and how they can benefit your projects.

Thank you for your continued efforts, and we look forward to your creative use of PyGPTPrompt 0.0.31!

Prototype-v0.0.30

Version 0.0.30: Empowering Your AI Journey

I am excited to introduce PyGPTPrompt version 0.0.30, which brings a suite of powerful tools and enhancements to your AI-assisted tasks. This release marks a significant step forward in making PyGPTPrompt a versatile and indispensable tool for developers and users alike.

Note:

Some features are temporarily disabled because they are still in development. They will be reintegrated and re-enabled in future releases.

Key Highlights:

1. Expanded Toolset: PyGPTPrompt now offers a suite of tools that streamline various tasks, including downloading, converting, quantizing, OCR (Optical Character Recognition), PDF-to-Text conversion, and chat interactions. These tools empower you to accomplish a wider range of tasks with ease.

2. Function Calling Support: I've introduced Function Calling support for both GPT and Llama.Cpp models. This feature enhances the capabilities of PyGPTPrompt, making it even more adaptable to your needs.

3. Hugging Face Model Integration: PyGPTPrompt now seamlessly integrates with Hugging Face, allowing you to effortlessly download models from their extensive library. You have the flexibility to choose between manual and automatic model downloads, depending on your preferences.

4. Llama.Cpp Model Support: PyGPTPrompt now supports llama.cpp model conversions and quantization, giving you greater flexibility in using different models for your projects.

5. OCR and PDF Support: With the addition of OCR and PDF-to-Text conversion support, PyGPTPrompt can now assist you in extracting text from images and PDF documents, expanding its utility in document processing tasks.

6. Refactored Core: I've undertaken significant code refactoring to improve the underlying structure of PyGPTPrompt. These changes lay the foundation for future enhancements, optimizations, and maintenance.

7. Bug Fixes: I've addressed various bugs and issues to enhance the overall stability and reliability of PyGPTPrompt.

Upgrade Notes:

• Bumped Version: I've incremented the version number from 0.0.8 to 0.0.30, reflecting the substantial improvements and additions in this release.

• Test Coverage: Our test cases for OpenAI and Llama.Cpp have been updated and improved, ensuring robust performance and functionality.

• File I/O and Configuration: I've refined file I/O operations and updated the configuration to provide a smoother user experience.

This release is a testament to my commitment to providing a versatile and reliable AI-assisted tool. I hope you find these new features and enhancements valuable in your projects.

Please refer to the documentation for detailed information on how to leverage these new capabilities.

I look forward to your creative use of PyGPTPrompt 0.0.30!