emotion-aware-narrative 0.1.0

An AI-powered emotion detection system for game narrative enhancement

Emotion-Aware Narrative Engine

A sophisticated emotion detection system for games and interactive narratives that combines scene context with emotion detection to provide accurate emotional analysis and mood suggestions.

Features

• Emotion Detection: Uses advanced NLP to detect emotions in text
• Scene Context: Understands different scene types and their emotional implications
• Context Overrides: Allows for specific emotional contexts to modify detected emotions
• Mood Suggestions: Provides game mood cues based on detected emotions
• Flexible Integration: Works with various dialogue systems and game engines

Supported Emotions

Basic Emotions

• joy
• anger
• sadness
• fear
• surprise
• disgust
• neutral

Additional Emotions

• hope
• excitement
• gratitude
• anxiety
• disappointment
• guilt
• jealousy
• confusion
• sympathy

Scene Types

The system supports various scene types that can influence emotion detection:

Social Scenes

• formal_meeting - Business meetings, conferences
• party - Celebrations, festivals, gatherings
• family_gathering - Family reunions, home gatherings
• interview - Job interviews, auditions
• negotiation - Business deals, bargaining

Action Scenes

• chase_scene - Pursuits, escapes
• stealth_mission - Infiltration, sneaking
• training_session - Practice, learning
• competition - Contests, matches, tournaments
• rescue_mission - Saving, recovery operations

Emotional Scenes

• funeral - Memorials, burials
• wedding - Marriage ceremonies
• graduation - Commencement ceremonies
• reunion - Meetings, gatherings
• farewell - Goodbyes, departures

Mystery/Thriller Scenes

• investigation - Detective work, searches
• interrogation - Questioning, examinations
• discovery - Findings, revelations
• revelation - Disclosures, exposures
• confrontation - Conflicts, showdowns

Fantasy/Sci-fi Scenes

• magic_ritual - Spellcasting, ceremonies
• space_battle - Space combat, wars
• time_travel - Time jumps, temporal events
• dimension_hop - Portal travel, rifts
• magical_duel - Magical battles, confrontations

Original Scenes

• romantic_date - Romantic encounters
• battle_scene - Combat situations
• casual_conversation - Informal chats

Each scene type has:

• A specific mood
• Intensity level (low/medium/high)
• Default emotions
• Scene aliases for flexible matching

Installation

pip install emotion-aware-narrative

Quick Start

from emotion_engine import EmotionEngine

# Initialize the engine
engine = EmotionEngine().initialize()

# Process dialogue with scene and context
result = engine.process_dialogue_file({
    "text": "quit it",
    "scene": "romantic_date",
    "context": "flirty"
})

print(f"Detected Emotion: {result['emotion']}")
print(f"Mood Suggestion: {result['mood_suggestion']}")

Usage Examples

Basic Emotion Detection

result = engine.detect_emotion("quit it")

Scene-Based Detection

result = engine.detect_emotion("quit it", scene="romantic_date")

Context Override

result = engine.detect_emotion("quit it", context="flirty")

Combined Scene and Context

result = engine.detect_emotion("quit it", scene="romantic_date", context="flirty")

How It Works

1. Base Emotion Detection: Uses a pre-trained model to detect the initial emotion
2. Scene Context: Applies scene-specific emotional adjustments
3. Context Override: Applies any specified context overrides
4. Mood Suggestion: Generates appropriate mood cues for the final emotion

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Uses the emotion detection model from Hugging Face
• Inspired by the need for better emotional understanding in games