v0.0.4

Includes multiple fixes and feature additions.

v0.0.3

Includes multiple fixes and feature additions.

v0.0.2

Phenomenal AI v0.0.2

Phenomenal AI is a TypeScript package for building a turn-based
conversational prompt generator for large language models. This package provides
the building blocks for simulating conversational interactions between multiple
entities, referred to as actors. Actors can share information, engage in
dialogue, and dynamically adjust the course of conversation based on predefined
rules.

• 🎭 Multiple Actors: This package supports interaction between multiple
  actors, allowing for complex conversational scenarios. Each actor can have
  their own personality and preferences.
• 📚 Conversation History: Every conversation has a history which tracks the
  course of dialogue over time. The history is used to generate responses to new
  lines of dialogue.
• 💾 Shared Context: Actors in a conversation share a context that allows
  them to store and retrieve information relevant to the ongoing discussion. The
  context can be used to store information such as the topic of conversation,
  the location, the time of day, etc.
• 🔄 Scheduling: The package provides a mechanism to decide the order of
  turns between the actors using schedulers. The package comes with a default
  scheduler that schedules turns in a round-robin fashion.
• 🎬 Conversation Turn Management: Controls the turn flow of the
  conversation and provides a method to inject new messages into the
  conversation. The package also provides a method to query the conversation
  with a question and get a response.
• 📜 Context Window Management: The package provides a mechanism to manage
  the context window. The context window is the number of tokens to unmask in
  the history when generating a response. This allows you to control the prompt
  size and the amount of context provided to the model.
• ✨ High-Level Abstraction: Phenomenal AI provides a high level of
  abstraction for AI-based conversation models. This allows you to focus on the
  conversation logic and not the underlying model.
• 📝 TypeScript: The package is written in TypeScript and provides type
  definitions.
• 📦 Lightweight: The package has only one external dependency (mustache),
  making it lightweight and easy to use.
• 🧪 Extensible: The package is designed to be extensible. You can provide
  your own text generation function, scheduler, etc.