rilerary 0.2.0

A Python library for managing Riley AI personas

Rilerary 🤖

A Python library for managing Riley AI personas - a collection of specialized AI assistant personalities.

Features

• 🎭 Multiple Personas: Chef, Teacher, Finance, Tech, Motivation, Sustainability, Knowledge, and more
• 📦 Easy to Use: Simple API for loading and working with personas
• 🔧 Flexible: Use built-in personas or add your own
• 📝 Well-Structured: JSON-based persona definitions with rich metadata

Installation

pip install rilerary

Or install from source:

git clone https://github.com/NamaiBest/rilerary.git
cd rilerary
pip install -e .

Quick Start

from rilerary import load_persona, get_available_personas

# Get list of available personas
personas = get_available_personas()
print(f"Available personas: {personas}")

# Load a specific persona
chef = load_persona("chef_rile")
print(f"Persona: {chef['persona_name']}")
print(f"Description: {chef['description']}")
print(f"Specialties: {chef['specialties']}")

# Use the prompt template in your AI system
system_prompt = chef['prompt_template']

Available Personas

• 👨‍🍳 Chef Rile: Cooking, recipes, nutrition, and food safety
• 👨‍🏫 Teacher Rile: Education, learning strategies, and academic support
• 💰 Finance Rile: Personal finance, investing, and money management
• 💻 Tech Rile: Technology, programming, and software development
• 💪 Motivation Rile: Personal development and motivation
• 🌱 Sustainability Rile: Environmental awareness and sustainable living
• 📚 Knowledge Rile: General knowledge and research
• 👁️ Locket Visual Assistant: Quick visual analysis for locket mode

Advanced Usage

Using PersonaLoader Class

from rilerary import PersonaLoader

# Create loader instance
loader = PersonaLoader()

# Get all personas at once
all_personas = loader.get_all()

# Get basic info without loading full data
info = loader.get_persona_info("tech_rile")
print(f"{info['emoji']} {info['name']}: {info['description']}")

Building System Prompts

from rilerary import load_persona
from rilerary.utils import build_system_prompt

persona = load_persona("motivation_rile")

# Build complete system prompt with additional context
system_prompt = build_system_prompt(
    persona,
    additional_context="Focus on helping with morning routines."
)

Using Custom Persona Directory

from rilerary import PersonaLoader

# Load from custom directory
loader = PersonaLoader(custom_persona_dir="/path/to/your/personas")
custom_persona = loader.load("my_custom_persona")

Persona Structure

Each persona is defined in a JSON file with the following structure:

{
  "persona_name": "Chef Rile",
  "description": "Expert cooking assistant",
  "tone": "Enthusiastic and knowledgeable",
  "greeting": "Ready to cook!",
  "emoji": "👨‍🍳",
  "specialties": ["Cooking", "Recipes", "Nutrition"],
  "prompt_template": "System instructions...",
  "response_style": "Be friendly and practical...",
  "rubric": {
    "core_principles": {},
    "response_guidelines": {}
  }
}

Creating Custom Personas

1. Create a JSON file with your persona definition
2. Place it in a directory
3. Load it using PersonaLoader with custom directory

from rilerary import PersonaLoader

loader = PersonaLoader(custom_persona_dir="./my_personas")
my_persona = loader.load("my_persona_name")

API Reference

Functions

• load_persona(persona_name: str) - Load a persona by name
• get_available_personas() - Get list of all available personas
• get_persona_info(persona_name: str) - Get basic persona information

Classes

• PersonaLoader - Main class for persona management
  • load(persona_name) - Load a persona
  • get_available() - List available personas
  • get_all() - Load all personas
  • get_persona_info(persona_name) - Get basic info

Utility Functions

• build_system_prompt(persona_data, additional_context) - Build complete system prompt
• get_prompt_template(persona_data) - Extract prompt template
• get_specialties(persona_data) - Get persona specialties
• validate_persona(persona_data) - Validate persona structure

License

MIT License - see LICENSE file for details

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.